A system and method for estimating cell source proportions using methylation information.

Abstract

This invention provides a system and method for estimating the proportion of cell sources using methylation information. [Solution] The method obtains, for each training target in a plurality of training targets, the corresponding methylation pattern of each cell-free fragment in a plurality of corresponding training cell-free fragments, and the cancerous signs of the corresponding target. Each cell-free fragment is mapped to a bin in a plurality of bins, each bin representing a portion of the human reference genome. The method also inputs the corresponding methylation pattern of each cell-free fragment into a classifier. As a function of the classifier, the cancerous status of the cell-free fragment is assigned to each cell-free fragment. The method further determines a criterion for the correlation between the cancerous status of the target and the cancerous status of the cell-free fragment for each bin. Multiple features for estimating the cell source fraction of the target are identified as subsets of the plurality of bins.

Description

Cross-reference to Related Applications

[0001] This application claims priority to U.S. Provisional Patent Application No. 62 / 950,071, filed on December 18, 2019, entitled "Systems and Methods for Estimati ng Cell Source Fractions using Methylation Information", the content of which is hereby incorporated by reference in its entirety for all purposes.

Technical Field

[0002] This specification describes estimating the cell source fraction, such as the tumor fraction, in a biological sample obtained from a subject using nucleic acids of the subject, particularly cell-free nucleic acid samples.

Background Art

[0003] With the increasing knowledge of the molecular basis of cancer and the rapid development of next-generation sequencing technologies, research on the initial molecular changes involved in cancer development in body fluids has advanced. Next-generation sequencing (NGS) and other large-scale sequencing technologies provide opportunities to achieve sequencing at a cost of less than $1 per million bases, and in fact, a cost of less than 10 cents has been achieved. Cell-free DNA (cfDNA) in plasma, serum, and urine has been found to have specific genetic and epigenetic changes related to cancer development. Such changes may be potentially useful as diagnostic biomarkers for several classes of cancer (see Salvi et al., 2016, Onco Targets Ther. 9:6549-6559).

[0004] Cell-free DNA (cfDNA) is found in serum, plasma, urine, and other body fluids (Chan et al., 2003, A It was observed in nn Clin Biochem. 40(Pt 2):122-130) and became a "liquid biopsy," and was identified. This shows the circulatory picture of the disease (De Mattos-Arruda and Caldas, 2016, Mol Oncol.10(3):464-474). (See reference). This represents a potential non-invasive method for screening various cancers. It represents

[0005] The existence of cfDNA was demonstrated by Mandel and Metais several decades ago. 1948, CR Seances Soc Biol Fil. 142(3-4):241-243). cfDNA is found in necrotic cells or apo It originates from ptosis cells and is generally released from all types of cells. Stroun et al. further stated that This study demonstrated that specific cancerous changes were observed in the patients' cfDNA (Stroun et al., 1989 Oncology). See 1989 46(5):318-322). In many subsequent papers, cfDNA has been shown to have mutations, methylation, and It has been confirmed that certain tumor-related changes, such as copy number variations (CNV), are included in the circulating tumor. The presence of tumor DNA (ctDNA) was confirmed (Goessl et al., 2000 Cancer Res. 60(21):5941-59). 45 and Frenel et al., 2015, Clin Cancer Res. 21(20):4586-4596).

[0006] While cfDNA in plasma or serum is well-characterized, urinary cfDNA (ucfDNA) has traditionally been... It was not well characterized. However, recent studies have shown that ucfDNA is also a promising biomer. It has been demonstrated that it can be a source of kers (e.g., Casadio et al. 31(8):1744-1750). ).

[0007] In blood, apoptosis is a frequent phenomenon that determines the amount of cfDNA. However, in cancer patients... In some individuals, the amount of cfDNA appears to be affected by necrosis (Hao et al., 201 4, Br J Cancer 111(8):1482-1489 and Zonta et al., 2015 Adv Clin Chem.70:197-246 (See reference). Since apoptosis is thought to be the main release mechanism, circulating cfDNA is apo A short fragment of approximately 167 base pairs, corresponding to a nucleosome produced by ptosis cells. It has a size distribution that reveals it is rich in (Heitzer et al., 2015, Clin Chem. 61(1 ):112-123 and Lo et al., 2010, Sci Transl Med.2(61):61ra91).

[0008] The amount of circulating cfDNA in serum and plasma is higher in tumor patients than in healthy controls, especially in the early stages. It appears to be significantly higher in patients with advanced tumors than in patients with tumors (Sozzi et al., 2003, J Cl in Oncol. 21(21):3902-3908, Kim et al., 2014, Ann Surg Treat Res. 86(3):136-142; (See and Shao et al., 2015, Oncol Lett. 10(6):3478-3482). The variation in the amount of circulating cfDNA is The risk was higher in cancer patients than in healthy individuals (Heitzer et al., 2013, Int J Cancer. 133(2):346-356). Reference), the amount of circulating cfDNA is affected by several physiological and pathological conditions, including inflammatory diseases (see Raptis and Menard, 1980, J Clin Invest. 66(6):1391-1399, and Shapiro et al., 1983, Cancer 51(11):2116-2120).

[0009] Methylation status and other epigenetic modifications are known to correlate with the presence of several disease states, such as cancer (see Jones, 2002, Oncogene 21:5358-5360). Furthermore it has been determined that specific patterns of methylation are associated with specific cancer states ( see Paska and Hudler, 2015, Biochemia Medica 25(2):161-176). Warton and Samimi demonstrated that methylation patterns can also be observed in cell-free DNA (Warton and Samimi, 2015, Front Mol Biosci, 2(13) doi: 10.3389 / fmolb.2015.00013).

[0010] Considering the promise of circulating cfDNA, and other forms of genotypic data, as diagnostic markers, there is a need in the art for methods to evaluate such data to identify epigenetic patterns. SUMMARY OF THE INVENTION

[0011] The present disclosure addresses the identified drawbacks in the background by providing robust techniques for determining cell source fractions, such as tumor fraction, in a biological sample obtained from a subject using cfDNA. Methylation data and whole genome, or targeted genomic sequencing ​​​​​​The combination with single data provides further diagnostic power beyond conventional screening methods.

[0012] Technical solutions (e.g., computer systems, methods, and non - transient computer - readable storage media) for addressing the above - identified problems related to the analysis of datasets are provided in this disclosure.

[0013] The following presents an overview of the invention to provide a basic understanding of some aspects of the invention. This overview is not an extensive overall view of the invention. It is neither identifying the key / determining elements of the invention nor intended to detail the scope of the invention. Its sole purpose is to present, as a prelude to the more detailed description that follows, a simplified summary of some of the concepts of the invention.

[0014] A. Embodiments of estimating the cell source ratio based at least in part on a subset of array groups identified by the ratio of cancer - derived fragments in each array group.

[0015] One aspect of the disclosure provides a method of identifying a plurality of features for estimating the cell source ratio of a subject. The method includes, in a computer system having one or more processors and a memory storing one or more programs for execution by the one or more processors, obtaining a training dataset in electronic form. The training dataset includes, for each of a plurality of training subjects: a) the corresponding methylation pattern of each cell - free fragment among a plurality of cell - free fragments for corresponding training, and b) the signs of the subject's cancer for each training subject. Each cell - free ​​​​​​​​​​​​​ The corresponding methylation patterns of the fragments are (i) the corresponding raw materials obtained from each training subject. Methylation sequencing of one or more nucleic acid samples containing each fragment in a body sample. (ii) Determined by (ii) the methyl group of each CpG site among the corresponding CpG sites in each fragment It includes the state of transformation. The state of the cancer in question is the state of the first cancer and the state of the second cancer. On the other hand, this method involves each cell-free fragment of a plurality of arrangements This further includes mapping to one of the sequence groups within the sequence group. Here, of multiple sequence groups Each sequence group represents a corresponding region of the human reference genome, thereby allowing for the replication of cell-free fragments. You obtain a number of training sets, but each training set of self-free fragments contains multiple The sequence groups are mapped to different sequence groups within the column group. This method maps each cell-free fragment to a different sequence group. The state of the cell-free fragment is a function of the output of the classifier when a turn is input into the classifier. The state of each training of a self-free fragment in multiple training sets of self-free fragments This further includes assigning to each cell-free fragment within the set. The state is either the first cancer state or the second cancer state. This method involves multiple sequences For each sequence group in the group, (a) for each training target among multiple training targets, (b) The state of the elephant's cancer and the tray of corresponding cell-free fragments mapped to each sequence group. Corresponding relationship between the cancer status of each cell-free fragment in the Ningset This further includes determining the scale of. In some embodiments, this association method is This is a correlation calculation. In some embodiments, this correlation method is a mutual information calculation. In some embodiments, this association method is based on a distance metric (e.g., Manhattan). Tann distance, maximum value, normalized Euclidean distance, normalized Manhattan distance, Dice coefficient, Co This method involves calculating the sine distance (or Jacquard coefficient, etc.). Identifying multiple features for estimating cell source proportions as subsets of multiple sequence groups. The process proceeds by and . Each sequence group in a subset of multiple sequence groups is, for each sequence group It satisfies the selection criteria based on the corresponding relevance scale. For example, in several embodiments Therefore, the sequence group that has the top rank in terms of relevance to all other sequence groups is selected based on the selection criteria. It is considered to satisfy the condition.

[0016] In some embodiments, the method involves each of several cell-free fragments for testing. By a procedure including obtaining the corresponding methylation pattern of the rufree fragment in electronic format This further includes estimating the cell source proportion for each cell-free fragment. The corresponding methylation pattern includes (i) each fragment in the biological sample obtained from the test subject. (ii) ) Includes the methylation state of each CpG site among the corresponding multiple CpG sites in each fragment. Each cell-free fragment for the multiple cell-free fragments is one of several sequence groups. It is mapped to, thereby obtaining multiple test sets of cell-free fragments, but the cell Each test set of free fragments is mapped to a different set of sequences from among several sets of sequences. This is a function of the output of a classifier when the methylation pattern of a cell-free fragment is input to the classifier. The cancerous state of the cell-free fragments is self- Assigned to each cell-free fragment within each test set of Lee fragments. Number of cell-free fragments The first representative value is for each test set of self-free fragments across subsets of multiple sequences. The first cancer state is calculated from the assigned test subject. Cell-free fragments The second representative value of the number is each test set of cell-free fragments across subsets of multiple array sets. In this, it is calculated from the test subject. Next, the cell source proportion of the test subject is the first and the second It is estimated using two representative values.

[0017] In some embodiments, the second cancer state is the absence of cancer, and the subject of the test is... The cell source percentage includes the percentage of the cell source being tested.

[0018] In some embodiments, the classification device, formula:

number

[0019] In some embodiments, the relevance scale I is:

number

[0020] In some embodiments, the measure of relevance is correlation. In some embodiments, The correlation is the Pearson correlation coefficient. In some embodiments, the correlation is the adjusted correlation coefficient, weighted correlation coefficient. This is done using the correlation coefficient, reflection correlation coefficient, or scaled correlation coefficient.

[0021] In some embodiments, the number of sequence groups is 1,000 to 100,000. In some embodiments, the number of sequence groups ranges from 15,000 to 80,000. It consists of a group of sequences. In some embodiments, each of the multiple sequence groups is, on average, It has 10 to 1200 residues. In some embodiments, each of the multiple sequence groups is On average, they have 10 to 10,000 residues.

[0022] In some embodiments, the first representative value is a set of sequences across a subset of multiple sequence groups. Multiple tests assigned to each test set of Ruflee fragments indicate the first cancer state. Arithmetic mean, weighted mean, midrange, mid-hinge, three of the number of self-free fragments from the target. This is the trimean, Windsor mean, mean, or mode.

[0023] In some embodiments, the second representative value is a self across a subset of multiple sequence groups. Arithmetic of the number of self-free fragments from multiple test subjects in each test set of Lee fragments. Average, weighted average, midrange, mid-hinge, ternary mean, Windsor mean, average, t is the mode.

[0024] In some embodiments, estimating the cell source proportion involves first representative values ​​and second representative values. This includes division by a certain factor.

[0025] In some embodiments, there are multiple training subjects, ranging from 10 to 100 training subjects. It consists of 0 training targets.

[0026] In some embodiments, the selection criterion is a distribution of one of the top N relevance measures. Specify the selection of a group of columns, where N is a positive integer greater than or equal to 50. In some embodiments, N is 5 The range is 00 to 5000. In some embodiments, N is 800 to 1500.

[0027] In some embodiments, methylation sequencing is paired-end sequencing. In some embodiments, methylation sequencing is performed using single-read sequencing. In some embodiments, the corresponding number of self-free fragments for training are They have an average length of less than 500 nucleotides.

[0028] In some embodiments, the first cancer state is cancer, and the second cancer state is It is non-existent.

[0029] In some aspects, the first cancerous conditions are adrenal cancer, biliary tract cancer, bladder cancer, Bone / bone marrow cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, stomach cancer, head / neck Hmm, hepatobiliary tract cancer, kidney cancer, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, pelvic cancer, pleural cancer, Prostate cancer, kidney cancer, skin cancer, stomach cancer, testicular cancer, thymic cancer, thyroid cancer, uterine cancer, Lymphoma, melanoma, multiple myeloma, or leukemia are the second cancerous conditions. Cancer is not present.

[0030] In some embodiments, the first cancer state is adrenal cancer of a certain stage, a certain stage Bile tract cancer, bladder cancer at a certain stage, bone / bone marrow cancer at a certain stage Brain cancer, breast cancer at a certain stage, cervical cancer at a certain stage, colon cancer at a certain stage, Esophageal cancer at a certain stage, stomach cancer at a certain stage, head / neck cancer at a certain stage, Stage 1 hepatobiliary tract cancer, stage 1 kidney cancer, stage 1 liver cancer, stage 1 lung cancer Cancer, ovarian cancer at a certain stage, pancreatic cancer at a certain stage, pelvic cancer at a certain stage, Pleural cancer at a certain stage, prostate cancer at a certain stage, kidney cancer at a certain stage Skin cancer, stomach cancer at a certain stage, testicular cancer at a certain stage, thymus cancer at a certain stage Hmm, a certain stage of thyroid cancer, a certain stage of uterine cancer, a certain stage of lymphoma, ah Among melanoma of a certain stage, multiple myeloma of a certain stage, or leukemia of a certain stage One of these is the absence of cancer, and the second cancerous state is the absence of cancer.

[0031] In some embodiments, methylation sequencing is performed as whole-genome methylation sequencing. In some embodiments, methylation sequencing involves multiple nucleic acid probes. The target sequencing used involved multiple sequence groups, with each sequence group containing a small number of nucleic acid probes. It is associated with at least one nucleic acid probe.

[0032] In some embodiments, the multiple nucleic acid probes include more than 1,000 nucleic acid probes. Over 2,000 nucleic acid probes, over 3,000 nucleic acid probes, over 5,000 nucleic acid probes, 10,000 or more nucleic acid probes, or 1,000 to 30,000 It contains several nucleic acid probes.

[0033] In some embodiments, each of the multiple sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or that The above CpG sites are included. In some embodiments, each of the multiple sequence groups is 2 , 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 1 It includes 8, 19, 20, or more consecutive CpG sites. Several embodiments So, each of the multiple sequence groups corresponds to 2 to 100 consecutive CpG sites in the human reference genome. Ranaru.

[0034] In some embodiments, the corresponding biological sample is a liquid biological sample. In one embodiment, the corresponding biological sample is a blood sample. So, the corresponding biological samples are blood, whole blood, plasma, serum, urine, and brain spinal cord of the subject being trained. It contains cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites. Several implementation forms In this context, the corresponding biological samples are the blood, whole blood, plasma, serum, urine, and brain of the subject being trained. It consists of cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites.

[0035] In some embodiments, among the corresponding CpG sites in each fragment, each CpG site is The chilled state is determined by methylation sequencing, which checks if each CpG site is methylated. If this occurs, it is in a methylated state, and if each CpG site is not methylated, it is a methylated sequence. If determined by testing, it is in a non-methylated state, and the methylation state of each CpG site is determined by testing. If it cannot be called as either methylated or unmethylated, it will be called as "other". The mark is added.

[0036] In some embodiments, methylation sequencing is performed on each fragment, resulting in one or Multiple 5-methylcytosine (5mC) and / or 5-hydroxymethylcytosine (5hC) ) detects.

[0037] In some embodiments, methylation sequencing involves the sequence reads of each fragment. , one or more unmethylated cytosines or one or more methylated cytosines This includes conversion to one or more corresponding uracils. In some embodiments, 1 One or more uracils as one or more corresponding thymines, methylated sequence Detected during sync. In some embodiments, one or more unmethylated citrosin The conversion of one or more methylated cytosines can be done through chemical conversion, enzymatic conversion, and or a combination of those is included.

[0038] In some embodiments, the first model is a first mixture comprising a first plurality of submodels The first model is a second hybrid model that includes a second set of submodels, and the second model is a second hybrid model that includes a second set of submodels. Each of the submodels of the first and second submodels is self-reactive in the corresponding biological sample. —Represents the independent, corresponding methylation models of the source fragments.

[0039] In some embodiments, the independent corresponding methylation models are each a binary model. It is one of the following: a 2D model, a beta-binomial model, an independent-part model, or a Markov model.

[0040] In some embodiments, two or more submodels of the first set of submodels are independent of each other. This is a site model, and two or more submodels of the second submodel are independent sites. It is a model.

[0041] In some embodiments, the method involves one or more cells in a plurality of cell-free fragments. This further includes applying filter conditions.

[0042] In some embodiments, one or more filter conditions are selected from the filter conditions. One is that among multiple cell-free fragments, the corresponding methylation pattern of each cell-free fragment is p A value threshold may be applied, and the p-value threshold is determined when the methylation pattern is in a cohort of non-cancer subjects. This represents the frequency of observations made in that situation.

[0043] In some embodiments, the p-value threshold is 0.001 to 0.20.

[0044] In some embodiments, the cohort includes at least 20 subjects, and multiple cell-free The fragment contains at least 10,000 different corresponding methylation patterns.

[0045] In some embodiments, the p-value threshold is set for each cell-free fragment among a plurality of cell-free fragments. The corresponding methylation pattern of the fragment is 0.10 or less, 0.05 or less, or 0.01 or less. The p-value is satisfied when it is available, and this is true for the methylation pattern from the target.

[0046] In some embodiments, one or more filter conditions are selected from the filter conditions. One possibility is that each cell-free fragment in a group of cell-free fragments corresponds to each fragment in the corresponding biological sample. The odor of multiple corresponding sequence reads measured from one or more nucleic acid samples containing The requirement is that it be represented by a threshold number of array reads.

[0047] In some embodiments, the number of thresholds is 2, 3, 4, 5, 6, 7, 8, 9, 10, or These are integers between 10 and 100.

[0048] In some embodiments, one or more filter conditions are selected from the filter conditions. One possibility is that each cell-free fragment in a group of cell-free fragments corresponds to each fragment in the corresponding biological sample. In one or more nucleic acid samples containing a threshold number of cell-free nucleic acids The requirement is to apply the condition that it must be met.

[0049] In some embodiments, the number of thresholds is 2, 3, 4, 5, 6, 7, 8, 9, 10, or These are integers between 10 and 100.

[0050] In some embodiments, one or more filter conditions are used as filter strips. One of the findings is that each cell-free fragment has a threshold number of CpG sites. The requirement is to apply the necessary conditions.

[0051] In some embodiments, the threshold number of CpG sites is at least 1, 2, 3, 4, 5, 6. These are 7, 8, 9, or 10 CpG sites.

[0052] In some embodiments, one or more filter conditions are selected from the filter conditions. One is that each cell-free fragment has a length less than a threshold number of base pairs. This is a requirement.

[0053] In some embodiments, the threshold number of base pairs has lengths of 1,000, 2,000, 3,0 It consists of 00 or 4,000 consecutive base pairs.

[0054] In some embodiments, the method involves acquisition, mapping, assignment, and the first and third Calculation of 2 representative values, and detailed analysis of the test subject at each of the multiple time points across the epoch. The cell source proportion is repeatedly estimated, thereby determining the cell source proportion for the test subjects at each time point. Obtain the corresponding cell source ratio from among them, and use the multiple cell source ratios to determine the epoch The state or progression of the disease being tested is determined by an increase or decrease in the proportion of primary cells over an epoch. This includes determining the result as "few".

[0055] In some embodiments, an epoch is a period of several months, and each of several time points is These are different points in time within a period of several months.

[0056] In some embodiments, a period of several months is less than four months.

[0057] In some embodiments, an epoch is a period of several years, and each of several time points is: These are different points in time within a period of several years.

[0058] In some embodiments, the period of several years is 2 to 10 years.

[0059] In some embodiments, an epoch is a period of several hours, and each of several time points is These are different points in time within a period of several hours.

[0060] In some embodiments, the period of several hours is between 1 and 6 hours.

[0061] In some embodiments, the method is used when the proportion of the target first cell source over an epoch. If only a threshold amount is observed to change, further modify the diagnosis being tested. include.

[0062] In some embodiments, the method is used when the proportion of the target first cell source over an epoch. If only a threshold amount is observed to change, it further modifies the prognosis of the subject under test. include.

[0063] In some embodiments, the method is used when the proportion of the target first cell source over an epoch. If only a threshold amount is observed to change, further modifications to the treatment being tested may be considered. include.

[0064] In some embodiments, the thresholds are greater than 10%, greater than 20%, greater than 30%, greater than 40%, and 50%. It is more than twice, more than three times, or more than five times.

[0065] In some embodiments, the proportion of tumors under test is between 0.003 and 1.0.

[0066] In some embodiments, the method involves at least a portion of the value of the cell source percentage being tested. This further includes applying a treatment regimen to the test subject based on the criteria.

[0067] In some embodiments, the treatment regimen involves applying an anticancer drug to the subject of test. This includes.

[0068] In one embodiment, cancer drugs are hormones, immunotherapy, X-rays, or anticancer drugs. be.

[0069] In some embodiments, the cancer drug is lenalidomide, pembrolizumab, tra Sustuzumab, bevacizumab, rituximab, ibrutinib, human papillomavirus quadrivalent (Types 6, 11, 16, and 18) vaccines, pertuzumab, pemetrexed, nilotini Nilotinib, denosumab, abiraterone acetate, promacta, imatinib, everolim S, palbociclib, erlotinib, bortezomib, bortezomib, or their respective drugs It is equivalent to Nelic.

[0070] In some embodiments, the test subject is treated with an anticancer drug, and this method The proportion of cells from the test source is used to evaluate the response to the cancer drug being tested. This further includes the following.

[0071] In one embodiment, cancer drugs are hormones, immunotherapy, X-rays, or anticancer drugs. be.

[0072] In some embodiments, the cancer drug is lenalidomide, pembrolizumab, tra Sustuzumab, bevacizumab, rituximab, ibrutinib, human papillomavirus quadrivalent (Types 6, 11, 16, and 18) vaccines, pertuzumab, pemetrexed, nilotini Nilotinib, denosumab, abiraterone acetate, promacta, imatinib, everolim S, palbociclib, erlotinib, bortezomib, bortezomib, or their respective drugs It is equivalent to Nelic.

[0073] In some embodiments, the test subject is treated with an anti-cancer drug, and this method Using the proportion of the cell source in the test subjects, we can either enhance or discontinue cancer drugs in the test subjects. This further includes deciding whether or not to do so.

[0074] In some embodiments, the test subject undergoes surgical intervention to treat cancer. This method uses the proportion of cell sources of the test subject to evaluate the response of test subjects to surgical intervention. This further includes assessing the elephant's condition.

[0075] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 20 Lists 1-8 (published as No. 20 / 069350) and / or International Patent Application No. P CT / US2020 / 015082 (published as International Publication No. 2020 / 154682) These correspond to one or more genomic regions listed in lists 1-16 of the following: Each of these shall be incorporated herein by reference in its entirety.

[0076] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 202) Lists 1-8 (published as No. 0 / 069350) and / or International Patent Application No. PC T / US2020 / 015082 (published as International Publication No. 2020 / 154682) At least 30% of the genomic regions listed in one or more of the lists 1-16 are affected. It will be topped.

[0077] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 202) Lists 1-8 (published as No. 0 / 069350) and / or International Patent Application No. PC T / US2020 / 015082 (published as International Publication No. 2020 / 154682) At least 50–95 of the genomic regions listed in one or more of lists 1–16. It is mapped to a percentage.

[0078] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 202) Lists 1-8 (published as No. 0 / 069350) and / or International Patent Application No. PC T / US2020 / 015082 (published as International Publication No. 2020 / 154682) In one or more of the lists 1-16, 1-10 unique corresponding genomic regions It will be topped.

[0079] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 202) Lists 1-8 (published as patent number 0 / 069350) and international patent application number PCT / US List of publications in issue 2020 / 015082 (published as international publication no. 2020 / 154682) In one or more of the 1-16 regions, it is mapped to a single, unique corresponding genomic region.

[0080] In some embodiments, for each of the multiple training targets, Multiple cell-free fragments for training, at least 100,000 cell-free fragments Includes.

[0081] In some embodiments, for each of the multiple training targets, Multiple cell-free fragments for training, at least 100,000 cell-free fragments Includes.

[0082] In some embodiments, for each of the multiple training targets, Multiple cell-free fragments for training contain at least 1 million cell-free fragments. .

[0083] In some embodiments, each sequence group of multiple sequence groups comprises fewer than 100 nucleic acid residues, 50 Nucleic acid residues less than 0, less than 1000, less than 2500, 5000 Nucleic acid residues less than 10,000, nucleic acid residues less than 25,000, 5 Nucleic acid residues less than 0,000, less than 100,000, less than 250,000 It consists of nucleic acid residues, or fewer than 500,000 nucleic acid residues.

[0084] Another aspect of this disclosure provides a computer system for estimating the proportion of a target cell source. The computer system comprises one or more processors and one or more Includes memory that stores one or more programs executed by the processor. Hmm. One or more programs give instructions to obtain the training dataset. It is included in electronic format. The training dataset is for each of the multiple training targets. Regarding the training target, a) each of the multiple cell-free fragments for corresponding training b) the corresponding methylation pattern of each training target, and the target cancer. The symptoms include: The corresponding methylation pattern of each cell-free fragment is (i) each training One or more nucleic acid samples containing each fragment from the corresponding biological sample obtained from the subject (ii) Determined by methylation sequencing of the fragment, and the corresponding number of CpGs in each fragment. It includes the methylation status of each CpG site within the site. The state of the target cancer is the first cancer. It is one of the conditions and a second cancerous condition. One or more programs are multiple cells Each free fragment is mapped to one of several sequence groups. Further includes instructions for performing the following: Here, each of the multiple sequence groups is a human reference genome. This represents the corresponding part, thereby obtaining multiple training sets of self-free fragments. However, each training set of cell-free fragments matches different sequences from among multiple sequences. It is pinged. One or more programs determine the methylation pattern of each cell-free fragment. The cancer state of the cell-free fragment, which is a function of the output of the classification device when input into the classification device, Each training set of the self-free fragment is one of several training sets of the self-free fragment. Further includes instructions for assigning to each cell-free fragment within the cell-free fragment. The condition is either a first cancer state or a second cancer state. One or more pro Gram considers each of the multiple sequence groups to be (a) each of the multiple training targets. (b) the state of the target cancer of the training subject, and the corresponding sequence mapped to each sequence group. Corresponding associations between cancer states of each cell-free fragment in a training set of cell-free fragments Further includes instructions for determining the sex scale I. One or more programs are targeted To identify multiple features for estimating the proportion of cell sources as subsets of multiple sequence groups Further instructions include: In a subset of multiple sequence groups, each sequence group is related to each sequence group. It satisfies the selection criteria based on the corresponding relevance scale.

[0085] Another aspect of this disclosure is that one or more programs are used in the manner disclosed herein. The above disclosed instructions further include instructions for performing either one or a combination thereof. We provide computer systems.

[0086] Another aspect of this disclosure is one or more programs for estimating the proportion of a target cell source. Provides a non-temporary computer-readable storage medium that stores one or more programs. A program is configured to be run by a computer. One or more programs It includes instructions in electronic format for obtaining the training dataset. The training dataset is for each of the multiple training targets: a) corresponding The corresponding methylation of each cell-free fragment among multiple cell-free fragments used for training. The pattern and b) the signs of the target cancer for each training subject. Each cell-free fragment The corresponding methylation patterns are (i) the corresponding biological samples obtained from each training subject. Methylation sequencing of one or more nucleic acid samples containing each fragment in the sample (ii) including the methylation state of each CpG at the corresponding multiple CpG sites in each fragment The cancer condition in question is either the first or second cancer condition. One or more programs process each cell-free fragment of each of the multiple cell-free fragments. It contains instructions for mapping to one of several sets of sequences. Here, Each of the multiple sequence groups represents a corresponding region of the human reference genome, thereby forming a cell. Obtain multiple training sets of free fragments, but each training set of a cell free fragment The sequence is mapped to a different sequence group from among multiple sequence groups. One or more programs M is the output of the classification device when the methylation pattern of each cell-free fragment is input to the classification device. The function of the cancer state of a cell-free fragment is expressed using multiple training sets of cell-free fragments. To assign each self-free fragment within each training set of the self-free fragments. The instructions further include: The cancer state of the cell-free fragment is the first cancer state and the second It is one of the states. One or more programs are in each of the multiple array sets. (a) The cancer status of each of the multiple training subjects, (b) Each cell free fragment of the training set mapped to each sequence group Determine the corresponding association Scale I between the cancer status of refractory and refractory fragments. Further instructions include: One or more programs estimate the proportion of the target cell source. Includes instructions for identifying multiple features as subsets of multiple array sets. Each sequence group in the subset of the sequence groups corresponds to a corresponding measure of relevance for each sequence group. It meets the selection criteria.

[0087] Another aspect of this disclosure is that one or more programs are used in the manner disclosed herein. The above disclosed non- Provides a temporary computer-readable storage medium.

[0088] B. Using methylation data obtained from cell-free DNA, the proportion of the cell source of the test subject was determined. An embodiment intended to make a decision.

[0089] Another aspect of this disclosure provides a method for estimating the proportion of a target cell source. This method is one or multiple processors, and one or more processors for execution by one or more processors In a computer system having memory that stores a number of programs, electronic form The formula obtains the corresponding methylation pattern of each cell-free fragment from among multiple cell-free fragments. This includes the following: Here, the corresponding methylation pattern of each cell-free fragment is (i) the target Methylation of one or more nucleic acid samples containing each fragment in a biological sample obtained from (ii) Determined by sequencing, (ii) each C of the multiple corresponding CpG sites in each fragment This method includes the methylation state of the pG site. Map each piece to one of several sequence groups, thereby creating multiple cell-free This includes obtaining a set of fragments. Each cell-free fragment set is a different set of multiple sequences. This method maps to a set of sequences. This method also divides the methylation pattern of each cell-free fragment. The cancer state of cell-free fragments, which is a function of the output of the classification device when input into the classification device, is used for multiple purposes. Assign to each cell-free fragment within each cell-free fragment set in the cell-free fragment set This includes the following. The cancerous state of the cell-free fragment is the first cancerous state and the second cancerous state. This is one of the states. This method applies to each cell-free fragment set across multiple sequence groups. The first cancer state is determined by calculating the first representative value of the number of cell-free fragments from the assigned subject. In each cell-free fragment set spanning multiple sequence groups, the cell-free fragments from the target This method proceeds by calculating the second representative value of a number. This further includes using the values ​​to estimate the proportion of the target cell source.

[0090] In some embodiments, the multiple sequence groups consist of ~1000 sequence groups. In the application configuration, multiple sequence groups consist of 15,000 to 80,000 sequence groups.

[0091] In some embodiments, each of the multiple sequence groups comprises, on average, 10 to 1200 sequences. It has residues. In some embodiments, each of the multiple sequence groups has, on average, 10~ Remove 10,000 residues.

[0092] In some embodiments, the first representative value is each cell-free fragment across multiple sequence groups. In the first cancer state, the arithmetic average of the number of cell-free fragments from the assigned subject. Average, weighted average, midrange, mid-hinge, ternary mean, Windsor mean, average, also This is the mode. In some embodiments, the second representative value is the self across multiple sets of sequences. The arithmetic mean, weighted mean, and minus the number of self-free fragments from the target in each set of free fragments. This refers to the midrange, mid-hinge, ternary mean, Windsor mean, mean, or mode.

[0093] In some embodiments, estimating the cell source proportion involves dividing the first representative value by the second representative value. This includes doing so.

[0094] In some embodiments, methylation sequencing is paired-end sequencing. In some embodiments, methylation sequencing is performed using single-read sequencing. It is.

[0095] In some embodiments, multiple cell-free fragments have an average length of less than 500 nucleotides. It has.

[0096] In some embodiments, the first cancer state is cancer, and the second cancer state is It is non-existent.

[0097] In some aspects, the first cancerous conditions are adrenal cancer, biliary tract cancer, bladder cancer, Bone / bone marrow cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, stomach cancer, head / neck Hmm, hepatobiliary tract cancer, kidney cancer, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, pelvic cancer, pleural cancer, Prostate cancer, kidney cancer, skin cancer, stomach cancer, testicular cancer, thymic cancer, thyroid cancer, uterine cancer, It is one of the following: lymphoma, melanoma, multiple myeloma, or leukemia, and the condition is secondary cancer. It is non-existent.

[0098] In some embodiments, the first cancer state is adrenal cancer of a certain stage, a certain stage Bile tract cancer, bladder cancer at a certain stage, bone / bone marrow cancer at a certain stage Brain cancer, breast cancer at a certain stage, cervical cancer at a certain stage, colon cancer at a certain stage, Esophageal cancer at a certain stage, stomach cancer at a certain stage, head / neck cancer at a certain stage, Stage 1 hepatobiliary tract cancer, stage 1 kidney cancer, stage 1 liver cancer, stage 1 lung cancer Cancer, ovarian cancer at a certain stage, pancreatic cancer at a certain stage, pelvic cancer at a certain stage, Pleural cancer at a certain stage, prostate cancer at a certain stage, kidney cancer at a certain stage Skin cancer, stomach cancer at a certain stage, testicular cancer at a certain stage, thymus cancer at a certain stage Hmm, a certain stage of thyroid cancer, a certain stage of uterine cancer, a certain stage of lymphoma, ah Among melanoma of a certain stage, multiple myeloma of a certain stage, or leukemia of a certain stage One of these is the absence of cancer, and the second cancerous state is the absence of cancer.

[0099] In some embodiments, methylation sequencing is performed on the whole genome. This is a process. In some embodiments, methylation sequencing is performed using multiple nucleic acid probes. This is a targeted sequencing method using multiple sequence groups, where each sequence group is one of multiple nucleic acid probes. It will be associated with at least one of our corresponding nucleic acid probes.

[0100] In some embodiments, the multiple nucleic acid probes include more than 1,000 nucleic acid probes. Over 2,000 nucleic acid probes, over 3,000 nucleic acid probes, over 5,000 nucleic acid probes, 10,000 or more nucleic acid probes, or 1,000 to 30,000 Contains 0 nucleic acid probes.

[0101] In some embodiments, each of the multiple sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or that The above CpG sites are included. In some embodiments, each of the multiple sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 1 It includes 9, 20, or more consecutive CpG sites. In some embodiments, multiple Each sequence group consists of 2 to 100 consecutive CpG sites from the human reference genome.

[0102] In some embodiments, the biological sample is a liquid biological sample. Morphologically, the biological sample is a blood sample. In some embodiments, the biological sample The samples include blood, whole blood, plasma, serum, urine, cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, and pericardial fluid. , or including ascites. In some embodiments, the biological sample is the subject's blood, whole blood, blood It consists of plasma, serum, urine, cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites.

[0103] In some embodiments, among the corresponding CpG sites in each fragment, each CpG site is Chilled state is determined by methylation sequencing, where each CpG site is methylated. If this occurs, it is in a methylated state, and if each CpG site is not methylated, it is a methylated sequence. If determined by testing, it is in a non-methylated state, and the methylation state of each CpG site is determined by testing. If it cannot be called as methylated or unmethylated, it is flagged as "Other". It can be done.

[0104] In some embodiments, methylation sequencing is performed on each fragment, resulting in one or Multiple 5-methylcytosine (5mC) and / or 5-hydroxymethylcytosine (5hC) ) is detected.

[0105] In some embodiments, methylation sequencing involves the sequence reads of each fragment. , one or more unmethylated cytosines or one or more methylated cytosines This includes conversion to one or more corresponding uracils. In some embodiments, 1 One or more uracils, as one or more corresponding thymines, form a methylated sequence. Detected during sync. In some embodiments, one or more unmethylated citrosin The conversion of one or more methylated cytosines can be done through chemical conversion, enzymatic conversion, and or a combination of those is included.

[0106] In some embodiments, the first model is a first mixture comprising a first plurality of submodels The first model is a second hybrid model that includes a second set of submodels, and the second model is a second hybrid model that includes a second set of submodels. Each of the submodels of the first and second submodels is self-reactive in the corresponding biological sample. —Represents an independent, corresponding methylation model for the fragment source.

[0107] In some embodiments, the independent corresponding methylation models are each a binary model. It is one of the following: a 2D model, a beta-binomial model, an independent-part model, or a Markov model.

[0108] In some embodiments, two or more submodels among the first multiple submodels are independent This is a site model in which two or more of the second set of submodels are independent This is a morphological model.

[0109] In some embodiments, the method involves one or more cells in a plurality of cell-free fragments. This further includes applying filter conditions.

[0110] In some embodiments, one or more filter conditions are selected from the filter conditions. One is to apply p to the corresponding methylation pattern of each cell-free fragment among several cell-free fragments. This involves applying a value threshold, where the p-value threshold is determined by the methylation pattern in the non-cancer cohort. This represents the frequency observed in that context.

[0111] In some embodiments, the p-value threshold is 0.001 to 0.20. In some embodiments, the p-value threshold is 0.01 to 0.10. , over 0.001, over 0.005, over 0.010, over 0.020, over 0.030, 0.04 Greater than 0, greater than 0.050, greater than 0.060, greater than 0.070, greater than 0.080, greater than 0.090, and also It is greater than 0.010.

[0112] In some embodiments, the cohort consists of at least 20, at least 30, and fewer 50 people each, at least 100 people, at least 500 people, or at least 1000 people Includes the target. In some embodiments, there are at least 300 cell-free fragments. At least 500, at least 1000, at least 5000, at least 800 It contains 0 or at least 10,000 different corresponding methylation patterns.

[0113] In some embodiments, the p-value threshold is set for each cell-free fragment among a plurality of cell-free fragments. The corresponding methylation pattern of the fragment is 0.10 or less, 0.05 or less, or 0.01 or less. The p-value is satisfied when it is available, and this is true for the methylation pattern from the target.

[0114] In some embodiments, one or more filter conditions are selected from the filter conditions. One possibility is that each cell-free fragment in a group of cell-free fragments corresponds to each fragment in the corresponding biological sample. The odor of multiple corresponding sequence reads measured from one or more nucleic acid samples containing The requirement is that it be represented by a threshold number of array reads. In this embodiment, the number of thresholds is 2, 3, 4, 5, 6, 7, 8, 9, 10, or 10 to 100. It is an integer.

[0115] In some embodiments, one or more filter conditions are selected from the filter conditions. One possibility is that each cell-free fragment in a group of cell-free fragments corresponds to each fragment in the corresponding biological sample. In one or more nucleic acid samples containing a threshold number of cell-free nucleic acids The requirement is to apply that the threshold number is 2, 3, 4. It is an integer between 5, 6, 7, 8, 9, 10, or 10 to 100.

[0116] In some embodiments, the filter conditions in one or more filter conditions and The requirement is that each cell-free fragment of a group of cell-free fragments has a threshold number of CpG sites. The application of this is to apply the threshold number of CpG sites, These are 2, 3, 4, 5, 6, 7, 8, 9, or 10 CpG sites.

[0117] In some embodiments, the filter conditions in one or more filter conditions and This means that each cell-free fragment of a plurality of cell-free fragments has a length less than a threshold number of base pairs. This is the requirement. In some embodiments, the threshold number of base pairs is defined as having a length of 1000, 20 These are consecutive base pairs of 00, 3000, or 4000.

[0118] In some embodiments, a single filter condition is applied. Two filter conditions are applied. In some embodiments, three filter conditions are applied. The following applies. In some embodiments, four filter conditions are applied.

[0119] In some embodiments, this method involves acquisition, mapping, assignment, and first and second generations. Calculation of table values, and the cell source under test at each of the multiple time points across epochs. Estimate the proportion and, for each test subject at each time point, identify the corresponding cell among multiple cell source proportions. This further includes repeating the process of obtaining the source proportion. In some embodiments, this multiple The proportion of the cell source is expressed as an increase or decrease in the proportion of the first cell source over an epoch. It is used to determine the state or progression of the disease condition being tested for during the examination.

[0120] In some embodiments, each epoch is a period of several months, and each of the multiple point in time This refers to different points in time within a period of several months. In some embodiments, the period of several months is: It is less than 4 months. In some embodiments, each epoch is 1 month long. In one embodiment, each epoch is two months long. In some embodiments, each epoch Each epoch is 3 months long. In some embodiments, each epoch is 4 months long. In some embodiments, each epoch is 5, 6, 7, 8, 9, 10, 11, 12, 1 Length of 3, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 or 24 months It is.

[0121] In some embodiments, an epoch is a period of several years, and each of several time points is: These are different points in time within a period of several years. In some embodiments, the period of several years is 1 to 10 years. It is a year. In some embodiments, the period of several years is 1 year, 2 years, 3 years, 4 years, 5 years, 6 years. This can be one year, seven years, eight years, nine years, or ten years. In some embodiments, the epoch is one year. It's been about 30 years.

[0122] In some embodiments, an epoch is a period of several hours, and each of several time points is , different points in time within a period of several hours. In some embodiments, the period of several hours is 1 hour The interval is ~24 hours. In some embodiments, the period of several hours is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 It is 21, 22, 23, or 24 hours.

[0123] In some embodiments, the method is used when the proportion of the first cell source under test reaches an epoch. If only a threshold amount is observed to change, further modifying the diagnosis of the subject may be necessary. This includes, for example, in some embodiments, the diagnosis is changed from cancer present to remission. As another example, in some embodiments, the diagnosis is changed from "no cancer" to "cancer present." As another example, in some embodiments, the diagnosis is from the presence of first-stage cancer to second-stage cancer. The stage is changed to one with cancer. As another example, in some embodiments, the diagnosis is second The stage of cancer is changed from "stage 1 cancer" to "stage 3 cancer." As another example, In some embodiments, the diagnosis progresses from the presence of stage 3 cancer to the presence of stage 4 cancer. It will be changed. As yet another example, in some embodiments, the diagnosis is that there is no metastasis. The diagnosis will be changed from "cancer present" to "cancer with metastasis present."

[0124] In some embodiments, the method is used to determine the proportion of the first cell source under test over an epoch. If a change of only a threshold amount is observed, it further includes modifying the prognosis of the subject. For example, in some embodiments, the prognosis includes life expectancy, and the prognosis ranges from the first life expectancy to the second. The life expectancy was changed to two, with the first and second life expectancies having different durations. In the application method, the change in prognosis increases the average life expectancy of the subject. In some embodiments, Subsequent changes reduce the average life expectancy of the subjects.

[0125] In some embodiments, the method is used to determine the proportion of the first cell source under test over an epoch. If a change of only a threshold amount is observed, the treatment of the subject is further modified. In some embodiments, the change in treatment involves initiating cancer drug therapy, cancer drugs Increasing the dosage of physical therapy, discontinuing cancer drug therapy, or discontinuing cancer drug therapy This includes reducing the dosage. In some embodiments, the modification of treatment involves lenalidomin Pembrolizumab, trastuzumab, bevacizumab, rituximab, ibrutinib, Human papillomavirus quadrivalent (types 6, 11, 16, 18) vaccine, pertuzumab Pemetrexed, nilotinib, nilotinib, denosumab, abiraterone acetate, promacaneous Ta, imatinib, everolimus, palbociclib, erlotinib, bortezomib, bortezomib Initiating or discontinuing treatment for the target using zomib or its generic equivalents In some embodiments, the treatment modifications include lenalidomide, pembrolizumab. Trastuzumab, bevacizumab, rituximab, ibrutinib, human papillomavirus 4-valent (types 6, 11, 16, and 18) vaccine, pertuzumab, pemetrexate Denosumab, nilotinib, nilotinib, denosumab, abiraterone acetate, promacta, imatinib , everolimus, palbociclib, erlotinib, bortezomib, bortezomib, or This includes increasing or decreasing the dosage of those generic equivalents administered to the subject. It is achieved. In some embodiments, the threshold is greater than 10%, greater than 20%, greater than 30%, greater than 40%, greater than 5 0%, greater than 2-fold, greater than 3-fold, or greater than 5-fold.

[0126] In some embodiments, the tumor ratio of the test subject is from 0.003 to 1.0. In some embodiments, the tumor ratio of the test subject is from 0.005 to 0.80. In some embodiments, the tumor ratio of the test subject is from 0.01 to 0.70. In some embodiments [[ID=--]] the tumor ratio of the test subject is from 0.05 to 0.60.

[0127] In some embodiments, the method further comprises applying a treatment regimen to the test subject, at least in part based on the value of the cell source ratio of the test subject. In some embodiments, the treatment regimen comprises applying a cancer agent to the test subject. In some embodiments, the cancer agent is a hormone, immunotherapy, radiation, or an anticancer agent. In some embodiments, the cancer agent is lenalidomide, pembrolizumab, trastuzumab, bevacizumab, rituximab, ibrutinib, human papillomavirus quadrivalent (types 6, 11, 16, and 18) vaccine, pertuzumab, pemetrexed, nilotinib, nilotinib, denosumab, abiraterone acetate, promacta, imatinib, everolimus, palbociclib, erlotinib, bortezomib, bortezomib, or their generic equivalents.

[0128] In some embodiments, the test subject is being treated with a cancer agent, and the method comprises evaluating the response of the test subject to the cancer agent using the cell source ratio of the test subject. Furthermore, it includes: In some embodiments, cancer drugs include hormones, immunotherapy, X-ray examinations, or it is an anticancer agent. In some embodiments, the cancer agent is lenalidomide, pemb Lolizumab, trastuzumab, bevacizumab, rituximab, ibrutinib, human papillomavirus Pertuzumab, a quadrivalent (types 6, 11, 16, and 18) vaccine. Methrexed, nilotinib, nilotinib, denosumab, abiraterone acetate, promacta, Imatinib, everolimus, palbociclib, erlotinib, bortezomib, bortezomib It is either [product name] or its generic equivalent.

[0129] In some embodiments, the test subject is treated with an anticancer drug, and this method is used Use the proportion of the cell source being tested to either enhance or discontinue cancer drugs in the test subjects. This further includes determining whether or not. For example, in some embodiments, at least a threshold (e.g.) For example, greater than 0.05, greater than 0.10, greater than 0.15, greater than 0.20, greater than 0.25, or 0.30 Observing the proportion of cell sources (e.g., superimposed) enhances cancer drug in the test subject (e.g., administration It is used as justification for increasing the amount (raising the radiation level in radiation therapy). In the application method, values ​​below the threshold (e.g., 0.05, 0.10, 0.15, 0.20, 0.25, Observing the cell source percentage (or less than 0.30, etc.) indicates the use of cancer drugs in the test subject. It will be used as a basis for discontinuing it.

[0130] In some embodiments, the subject of test has undergone surgical intervention to address cancer. This method uses the cell source proportion of the test subjects to determine the response of test subjects to surgical intervention. Further includes evaluating the state. In some embodiments, the state is provided in this disclosure. This is a measurement standard based on the cell source ratio calculated using a specific method.

[0131] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 202) Lists 1-8 (published as No. 0 / 069350) and / or International Patent Application No. PC T / US2020 / 015082 (published as International Publication No. 2020 / 154682) This corresponds to one or more single genomic regions listed in lists 1-16, and these Each of these shall be incorporated herein in its entirety by reference.

[0132] In some embodiments, the sequence group in a group of sequences is International Patent Application No. PCT / Table of US2019 / 025358 (published as International Publication No. 2019 / 195268) 1-24, International Patent Application No. PCT / US2019 / 053509 (International Publication No. 202) Lists 1-8 (published as No. 0 / 069350) and / or International Patent Application No. PC T / US2020 / 015082 (published as International Publication No. 2020 / 154682) This corresponds to one or more combinations of genomic regions listed in lists 1-16. Each of these shall be incorporated herein by reference in whole. For example, how many In that embodiment, the sequence group in the plurality of sequence groups is as described in International Publication No. 2019 / 195268 Tables 1-24 of issue, Lists 1-8 of international publication No. 2020 / 069350, and / or One, two, three, four, five, or more than five regions listed in Lists 1 to 16 of International Publication No. WO 2020 / 154682 are included.

[0133] In some embodiments, the array groups in the plurality of array groups are at least 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, 99% or 100% mapped to one or more of the genomic regions listed in Tables 1 to 24 of International Publication No. WO 2019 / 195268, Lists 1 to 8 of International Publication No. WO 2020 / 069350, and / or Lists 1 to 16 of International Publication No. WO 2020 / 154682.

[0134] In some embodiments, the array groups in the plurality of array groups are at least 50 - 95% mapped to one or more of the genomic regions listed in Tables 1 to 24 of International Publication No. WO 2019 / 195268, Lists 1 to 8 of International Publication No. WO 2020 / 069350, and / or Lists 1 to 16 of International Publication No. WO 2020 / 154682.

[0135] In some embodiments, the array groups in the plurality of array groups are mapped to 1 - 10 specific corresponding genomic regions in one or more of Tables 1 to 24 of International Publication No. WO 2019 / 195268, Lists 1 to 8 of International Publication No. WO 2020 / 069350, and / or Lists 1 to 16 of International Publication No. WO 2020 / 154682.

[0136] In some embodiments, the array groups in the plurality of array groups are in one or more of Tables 1 to 24 of International Publication No. WO 2019 / 195268, Lists 1 to 8 of International Publication No. WO 2020 / 069350, and / or Lists 1 to 16 of International Publication No. WO 2020 / 154682. It is mapped to a single, unique, and corresponding genomic region.

[0137] In some embodiments, for each object, there are at least 10 cell-free fragments. ,000, 15,000, 20,000, 25,000, 50,000, 100,000 200,000, 300,000, 500,000, or 1,000,000 cell-free fragments Includes. In some embodiments, for each object, there are at least several cell-free fragments. It also contains 1 million cell-free fragments.

[0138] In some embodiments, each sequence group of multiple sequence groups comprises fewer than 100 nucleic acid residues, 50 Nucleic acid residues less than 0, less than 1000, less than 2500, 5000 Nucleic acid residues less than 10,000, nucleic acid residues less than 25,000, 5 Nucleic acid residues less than 0,000, less than 100,000, less than 250,000 It contains 100 nucleic acid residues, or fewer than 500,000 nucleic acid residues.

[0139] In some embodiments, each sequence group of a plurality of sequences consists of (i) 100 nucleic acid residues. ~(ii) 500 pieces, 1000 pieces, 2500 pieces, 5000 pieces, 10,000 pieces, 25,0 00, 50,000, 100,000, 250,000, or 500,000 It contains [number] nucleic acid residues.

[0140] Another aspect of this disclosure provides a computer system for estimating the proportion of a target cell source. A computer system consists of one or more processors and one or more pro It includes memory that stores one or more programs to be executed by the examiner. One or more programs correspond to each cell-free fragment among a plurality of cell-free fragments. It includes instructions for obtaining the methylation pattern in electronic form. Here, each self - The corresponding methylation pattern of the fragment is (i) each fragment in the biological sample obtained from the subject. Determined by methylation sequencing of one or more nucleic acid samples containing, ( ii) It includes the methylation state of each CpG site among the corresponding multiple CpG sites in each fragment. One or more programs process each cell-free fragment into multiple arrays We map to one of the sequence groups within the group, thereby obtaining multiple sets of cell-free fragments. It further includes instructions for doing so. Each cell-free fragment set is a different sequence from a group of sequences. It is mapped to a group. One or more programs are mapped to multiple sets of self-free fragments. In each cell-free fragment set, assign the cancer status of the cell-free fragment to each cell-free fragment. Further instructions for targeting are included. The cancer status of each cell-free fragment is determined by the The first cancer is a function of the output of the classification device when the methylation pattern is input to the classification device. It is one of the conditions and the second cancer condition. One or more programs are multiple sequences In each cell-free fragment set across the group, the subjects were assigned a first cancer state. Instructions for calculating a first representative value of the number of self-free fragments from a set of multiple arrays For each set of cell-free fragments, calculate a second representative value for the number of cell-free fragments from the target. The instructions further include instructions for doing so. One or more programs include a first representative value and a second representative value. The code further includes instructions for estimating the proportion of the target cell source using the values.

[0141] Another aspect of this disclosure is that one or more programs are used in any of the ways disclosed above. The above disclosed combination further includes instructions for performing either one or in combination. We provide a computer system.

[0142] Another aspect of this disclosure is one or more programs for estimating the proportion of a target cell source. Provides a non-temporary computer-readable storage medium that stores one or more programs. A program is configured to be run by a computer. One or more programs This involves electronically determining the corresponding methylation pattern of each cell-free fragment among multiple cell-free fragments. Includes instructions for obtaining in that form. The corresponding methylation pattern of each cell-free fragment. (i) One or more nucleic acid samples containing each fragment in the biological sample obtained from the subject (ii) Determined by methylation sequencing of the sample, and the corresponding multiple in each fragment It includes the methylation state of each CpG site within the CpG sites. One or more programs are Map each cell-free fragment into one of several sequence groups. Includes instructions for performing a slash and thereby obtaining multiple self-free fragment sets. Here, each A set of cell-free fragments is mapped to a different set of sequences from among several sets of sequences. Alternatively, when multiple programs input the methylation pattern of each cell-free fragment into the classification device... The cancer state of cell-free fragments is a function of the output of the classification device, and multiple cell-free fragments The set further includes instructions for assigning to each cell-free fragment within each cell-free fragment set. Hmm. The cancerous state of the cell-free fragment is one of the first cancerous state and the second cancerous state. Yes. One or more programs apply to each cell-free fragment set across multiple array sets. In this case, the first representative number of cell-free fragments from the subject assigned to the first cancer state. Calculate the value and, in each set of cell-free fragments across multiple array sets, the cells from the target. Further includes instructions for calculating a second representative value of the number of free fragments. One or more programs This includes instructions for estimating the proportion of the target cell source using the first and second representative values. .

[0143] Another aspect of this disclosure is that one or more programs are any of the methods disclosed above. The above disclosed non-transient instructions further include instructions for executing the above, either alone or in combination. To provide a computer-readable storage medium.

[0144] Various embodiments of systems, methods, and apparatus within the scope of the attached claims are, respectively There are several embodiments, but none of them are solely involved in the desirable attributes described herein. There is no need to limit the scope of the attached claims, and several outstanding features The characteristics are described herein. After considering this discussion, in particular, “Modes for Carrying Out the Invention” and After reading the titled section, understand how the features of the various embodiments are used. It will be possible.

[0145] [Integration by reference] All publications, patents, and patent applications referenced herein are subject to the individual publications, patents, and This is as much as when a patent application is specifically and individually indicated so as to be incorporated by reference. Each time, the entirety of that reference is incorporated herein.

[0146] The embodiments disclosed herein are not limiting, but are shown as examples in the accompanying drawings. Similar reference figures refer to the corresponding parts through several diagrams in the drawing. [Brief explanation of the drawing]

[0147] [Figure 1] This is an exemplary block diagram illustrating a computing device according to some embodiments of the present disclosure. [Figure 2] Figure 2A is a diagram illustrating exemplary flowcharts of methods for identifying multiple features for estimating the proportion of target cell sources according to several embodiments of the present disclosure, where dashed boxes represent any step. Figure 2B is a diagram illustrating exemplary flowcharts of methods for identifying multiple features for estimating the proportion of target cell sources according to several embodiments of the disclosure, where dashed boxes represent any step. [Figure 3] Figure 3A is a diagram summarizing exemplary flowcharts of methods for estimating the proportion of target cell sources according to several embodiments of the present disclosure, where dashed boxes represent any step. Figure 3B is a diagram summarizing exemplary flowcharts of methods for estimating the proportion of target cell sources according to several embodiments of the present disclosure, where dashed boxes represent any step. [Figure 4] This figure plots the ctDNA rate of subjects having any of the listed cancers as a function of cancer stage, according to some embodiments of the present disclosure. [Figure 5] This figure shows a flowchart illustrating a method for preparing nucleic acid samples for sequencing according to some embodiments of the present disclosure. [Figure 6] This figure shows a graphical representation of the process for obtaining sequence reads according to some embodiments of the present disclosure. [Figure 7]This figure compares tumor proportion estimates based on whole-genome bisulfite sequencing data from several embodiments of the present disclosure with known tumor proportions obtained from tissue-based whole-genome sequencing data. In particular, the WGBS-estimated tumor proportion includes the ratio of the average number of abnormal fragments to the average total number of fragments (e.g., each fragment is mapped to a specific set of sequences or regions in the reference genome). Figure 7 is based on sequencing information for 495 subjects. For known tissue tumor proportions > 0.01, the Spearman correlation for the WGBS tumor proportion estimate is 0.86. For known tissue tumor proportions > 0.005, the Spearman correlation for the WGBS tumor proportion estimate is 0.90. For known tissue tumor proportions > 0.001, the Spearman correlation for the WGBS tumor proportion estimate is 0.89. For known tissue tumor proportions > 0.0001, the Spearman correlation for the WGBS tumor proportion estimate is 0.74. This indicates that the tumor proportion estimates based on WGBS correlate with the tumor proportions of known tissues. [Figure 8] This figure shows a measure of mutual information used according to some embodiments of the present disclosure for feature identification. [Modes for carrying out the invention]

[0148] Next, the embodiments illustrated in the attached drawings will be described in detail. Detailed description follows. Therefore, to provide a complete understanding of this disclosure, numerous specific details are provided. However, it will be obvious to those skilled in the art that this disclosure can be implemented without these specific details. It is likely that, in other examples, well-known methods, procedures, components, circuits, and networks are, The embodiments are not described in detail in order to avoid unnecessarily obscuring their aspects.

[0149] The embodiments described herein are various technical methods for determining the estimated cell source proportion of a target. A solution is provided. In an exemplary embodiment, nucleic acid fragments are obtained from a biological sample of interest. Body samples contain cell-free nucleic acids. Therefore, nucleic acid fragments are cell-free nucleic acids. The acid fragment is evaluated for its methylation state against a predefined set of methylation sites, and the methylation state is evaluated against the methylation site. Each score is assigned based on the methylation state. Multiple methylation state scores are assigned to multiple counters. It is converted to a set of predefined methylation sites, and the corresponding site for each methylation site in the set of methylation sites is determined. It is compared to the methylation score. The corresponding methylation score is the methylation pattern in the cell source. This is obtained from the analysis of turns. By comparing these, the frequency of methylation of the target can be determined, and this Using this method, the proportion of cell sources can be estimated.

[0150] [Definition] As used herein, the terms “about” or “approximately” are: This means that the tolerance range for a specific value determined by those skilled in the art is such that the value is... It is measured or determined by, for example, partly depending on the limits of the measurement system. In some embodiments, "about" means, according to the convention of those skilled in the art, 1 or more than 1 standard deviation. This means within the difference. In some embodiments, "approximately" means ±20% of a given value. This means a range of ±10%, ±5%, or ±1%. In some embodiments, the term " "Approximately" or "about" means within one decimal place, within five times, or within twice the value of a given value. If a specific value is stated in this application and claims, unless otherwise specified, for a particular value The term "approximately" should be assumed to mean within an acceptable margin of error. It may have meanings that are generally understood by the trade. In some embodiments, the term "Approximately" refers to ±10%. In some embodiments, the term "approximately" refers to ±5%.

[0151] In this specification, the term "assay" means a substance, such as nucleic acids, proteins, cells, or combinations of substances. This refers to techniques for determining the properties of tissues or organs. Assays (e.g., first assay) (The second assay) involves checking the copy number variation of nucleic acids in the sample and the methylation of nucleic acids in the sample. Nucleic acid state, size distribution of nucleic acid fragments in the sample, mutation state of nucleic acids in the sample, or sample This may include techniques for determining the fragmentation pattern of nucleic acids during pulling. Use any assay known to those skilled in the art to detect any of the characteristics. This is possible. Nucleic acid characteristics include sequence, genomic identity, copy number, and one or more nucleos Methylation status at the nucleotide position, nucleic acid size, and position of one or more nucleotides The presence or absence of nucleic acid mutations and the pattern of nucleic acid fragmentation (for example, the nucleic acid fragmentation pattern) The assay or method may include the rheotide position. It possesses this capability, and its relative usefulness as a diagnostic tool can be measured using ROC-AUC statistics. It is possible to do so.

[0152] As used herein, the terms “biological sample,” “patient sample,” and “sun” are used interchangeably. "Pull" is used interchangeably and refers to any sample taken from the subject, and is related to the subject. It may reflect the biological state. In some embodiments, such samples are cell-free. It contains cell-free nucleic acids such as DNA. In some embodiments, such samples are It contains nucleic acids other than cell-free nucleic acids or cell-free nucleic acids in addition to nucleic acids. Examples of biological samples and The samples used are blood, whole blood, plasma, serum, urine, cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, and heart. This includes, but is not limited to, vesicular fluid or ascites. Several implementation forms In this context, biological samples include the subject's blood, whole blood, plasma, serum, urine, cerebrospinal fluid, feces, saliva, It consists of sweat, tears, pleural fluid, pericardial fluid, or ascites. In such embodiments, the biological sample is The samples include blood, whole blood, plasma, serum, urine, cerebrospinal fluid, stool, saliva, sweat, tears, pleural fluid, pericardial fluid, and more. This refers to ascites fluid only, and does not include other components of the subject (e.g., solid tissue). Biological sample This may include any tissue or substance derived from a living or dead subject. A sample may be a cell-free sample. A biological sample is a nucleic acid (e.g., DNA or RNA). ) or fragments thereof may be included. The sample may be a liquid sample or a solid sample (e.g., This may be a cell or tissue sample. Biological samples include blood, plasma, serum, urine, and vaginal fluid. , edema (e.g., scrotal fluid), vaginal douche fluid, pleural fluid, ascites, cerebrospinal fluid, saliva, sweat, tears, sputum, air Bronchoalveolar lavage fluid, drainage from the nipple, and fluid from different parts of the body (e.g., thyroid gland, breast). It may be a bodily fluid such as an aspirated fluid. The biological sample may be a stool sample. Various embodiments In this context, a bio-sample enriched with cell-free DNA (e.g., using a centrifugation protocol) Most of the DNA in plasma samples obtained via the method may be cell-free (for example) , more than 50%, more than 60%, more than 70%, more than 80%, more than 90%, more than 95%, or 99% of DNA (It may be cell-free). Biological samples are physically destroyed to remove tissue or cellular structures. Process the cells (e.g., by centrifugation and / or cell lysis) to extract intracellular components with enzymes, slowly It can be released into a solution that may further contain a suction fluid, salt, detergent, etc., and the said solution is for analysis. It can be used to prepare samples. Biological samples are invasively removed from the subject (for example). (by surgical means) or non-invasive methods (e.g., blood sampling, swab, or discharged sample) (Collection) can be obtained.

[0153] In some embodiments, the biological sample is a single tissue type (e.g., breast, lung, prostate). , large intestine, kidneys, uterus, pancreas, esophagus, lymph nodes, ovaries, cervix, epidermis, thyroid gland, bladder, or It originates from a single organ (such as the stomach). In some embodiments, the biological sample consists of two or more. It originates from the tissue type (for example, a combination of tissues from two or more organs). In the application method, the biological sample is one or more cell types (e.g., a single organ or a predetermined cell type). It originates from cells derived from a set of organs.

[0154] Where disclosed herein, the terms “nucleic acid” and “nucleic acid molecule” are used interchangeably. These terms refer to nucleic acids in any compositional form, for example, deoxyribonucleic acid (DNA, for example, phase Complementary DNA (cDNA), genomic DNA (gDNA), etc.), ribonucleic acid (RNA, for example, message RNA (mR) NA), small interfering RNA (siRNA), ribosomal RNA (rRNA), transfer RNA (tRNA), micro-R NA (such as RNA highly expressed in the fetus or placenta), and / or DNA or RNA analogs. (For example, including base analogs, sugar analogs and / or non-natural skeletons), RNA / DNA This refers to hybrid and polyamide nucleic acids (PNAs), all of which are single-stranded or double-stranded. It can be in the form of a chain. Unless otherwise specified, nucleic acids are known analogs of natural nucleotides. It may contain some of which function in a similar way to naturally occurring nucleotides. Yes, it is possible. Nucleic acids can be in any form useful for carrying out the processes described herein (e.g., linear, It can be circular, supercoiled, single-stranded, double-stranded, etc. In some embodiments... Nucleic acids can originate from a single chromosome or a fragment thereof (for example, nucleic acid samples are diploid). (May originate from one chromosome of a sample obtained from an object). In certain embodiments, nucleic acids are Nucleosomes, or fragments of nucleosomes or nucleosome-like structures Nucleic acids include parts of proteins (e.g., histones, DNA-binding proteins, etc.). In some cases, the nucleic acids analyzed by the processes described herein are substantially isolated. Nucleic acids may not be substantially bound to proteins or other molecules. "Sense" or "Antisense", "Plus" chain or "Minus" chain, "Forward" Reading frame or "reverse" reading frame) and double-stranded polynucleus RNA or DNA derivatives, variants, and analogues synthesized, replicated, or amplified from rheotides are also included. It contains. Deoxyribonucleotides include deoxyadenosine, deoxycytidine, and It contains oxyguanosine and deoxythymidine. In the case of RNA, the base cytosine is ura The syl group is replaced, and the 2' position of the sugar contains a hydroxyl group. The nucleic acid was obtained from the subject. It can be prepared using nucleic acids as a template.

[0155] As used herein, the term “cell-free nucleic acid” refers to the area outside of cells, the blood of the subject, or whole blood. Body fluids such as plasma, serum, urine, cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites. This refers to nucleic acid molecules that may be present within a cell. Cell-free nucleic acids are found in one or more healthy cells and Derived from one or more cancer cells. Cell-free nucleic acids are circulating nucleic acids. They are used interchangeably. Examples of cell-free nucleic acids include RNA, mitochondrial DNA, or ge Examples include, but are not limited to, Nomu DNA. When used herein, the term "Se "Cell-free nucleic acids," "cell-free DNA," and "cfDNA" are used interchangeably. When used in writing, the term “circulating tumor DNA” or “ctDNA” refers to tumor cells or other types of This refers to nucleic acid fragments derived from cancer cells, which are involved in apoptosis or necrosis of dying cells. As a result of biological processes, it is released into the fluid from an individual's body (e.g., blood flow), or Alternatively, they can be actively released by surviving tumor cells. Examples of cell-free nucleic acids include RNA, mi This includes, but is not limited to, tochondria DNA or genomic DNA.

[0156] Where disclosed herein, the terms “circulating tumor DNA” or “ctDNA” refer to tumor or other This refers to nucleic acid fragments derived from abnormal tissues such as cancer cells, and the apoptosis of dying cells. It is released into the target's bloodstream as a result of biological processes such as necrosis or other processes, or It can be actively released by surviving tumor cells.

[0157] Where disclosed herein, the term “reference genome” refers to an identification sequence from a subject. A specific and known sequencer of any organism or virus that can be used for this purpose. Refers to any genome, whether partial or complete, that has been identified or characterized. The exemplary reference genome used for elephants and many other organisms is from the National Center for Biotechnology Information. Online hosted by NCBI or the University of California, Santa Cruz (UCSC) Provided by the genome browser. "Genome" refers to a sequence of nucleic acids that can be represented by a biological or web browser. This refers to the complete genetic information of Russ. When used herein, the reference sequence or reference genome is Often, genome sequences constructed or partially constructed from an individual or multiple individuals. In some embodiments, the reference genome is constructed from one or more human individuals. A reference genome is a completed or partially constructed genome sequence. This can be considered a representative example. In some embodiments, the reference genome is divided into chromosomes. Includes the assigned sequence. An example human reference genome is NCBI Build 34 (UCSC equivalent). :hg16), NCBI build 35 (UCSC equivalent:hg17), NCBI build 36.1 (UCSC equivalent:hg 18) Examples include GRCh37 (UCSC equivalent: hg19) and GRCh38 (UCSC equivalent: hg38). However, it is not limited to these.

[0158] As disclosed herein, the terms “region of the reference genome,” “genome region,” or “stained” may be used. A "chromophore region" refers to any part of the reference genome, continuous or discontinuous. For example, a sequence... It can also be called a group, partition, genome segment, part of a reference genome, part of a chromosome, etc. In some embodiments, the genome section is based on a genome sequence of a specific length. In some embodiments, the method involves multiple nuclei mapped to multiple genomic regions. This may include the analysis of acid fragments. Genomic regions can be of approximately the same length, and The genome sections can be of different lengths. In some embodiments, the genome The genomic regions are approximately the same length. In some embodiments, genomic regions of different lengths are used. It is adjusted or weighted. In some embodiments, the genomic region is approximately 10 kilobases (k b)~500kb, approx. 20kb~approx. 400kb, approx. 30kb~approx. 300kb, approx. 40k b is approximately 200kb, and in some cases approximately 50kb to 100kb. In the embodiment, the genome region is approximately 100kb to approximately 200kb. The genome region is sequence It is not limited to consecutive executions. Therefore, a genomic region is a sequence of consecutive sequences and / or It can consist of discontinuous sequences. Genomic regions are not limited to a single chromosome. In some embodiments, the genomic region is all or part of one chromosome, or two or more chromosomes. It includes all or part of the chromosome. In some embodiments, the genomic region is one, two It may span multiple chromosomes or more. Furthermore, genomic regions may include multiple It may span the common or non-common regions of the chromosome.

[0159] As used herein, the term “fragment” is interchangeable with “nucleic acid fragment” (e.g., DNA fragment). A polynucleotide or This refers to a portion of a polypeptide sequence. In the context of quenching, the terms "fragment" and "nucleic acid fragment" are interchangeable in the context of biological samples. This refers to cell-free nucleic acid molecules found in or in representations of it. , sequencing data (e.g., whole genome sequencing, targeted sequencing, etc.) These sequence reads are one or more copies of all or part of such nucleic acid fragments. These sequence reads are used for derivation. In fact, PCR of the original nucleic acid fragment is used. It may also be obtained from sequencing of the replicas, and therefore the nucleic acid fragment is “representative” or “ To "support" specific nucleic acid fragments in a biological sample (e.g., PCR replicas) to each representative... Alternatively, there may be multiple supporting sequence reads. In some embodiments, nucleic acid fragmentation The fragments can be considered cell-free nucleic acids. In some embodiments, PCR duplicates or The sequence reads may be misleading; for example, the presence of certain cell-free nucleic acid molecules. This is the case when it is necessary to determine the quantity level. In such embodiments, one of the nucleic acid fragments Only one copy is used to represent the original cell-free nucleic acid molecule (for example, a library). During the reprocessing process, duplicates are removed by molecular identifiers attached to the cell-free nucleic acid molecules. (to be). In some embodiments, methylation sequencing data are obtained from these nucleic acids. It can be used to further distinguish fragments. For example, those that share the same or nearly identical sequence. If two nucleic acid fragments each possess different methylation patterns, they are still different. This may correspond to the original cell-free nucleic acid molecule.

[0160] In some embodiments, the two fragments have a sequence where each fragment has fewer than 2 nucleotides and 3 nucleotides. Less than rheotide, less than 4 nucleotides, less than 5 nucleotides, less than 6 nucleotides, 7 nuclei Less than 1 rheotide, less than 8 nucleotides, less than 9 nucleotides, less than 10 nucleotides, 15 Less than nucleotides, less than 20 nucleotides, less than 25 nucleotides, less than 30 nucleotides Full, less than 35 nucleotides, less than 40 nucleotides, less than 45 nucleotides, or 50 nucleotides If they differ from each other below the creotide level, they are considered to share nearly identical nucleic acid sequences. In some embodiments, the two fragments are such that each fragment sequence contains less than 1% of the total nucleotides. Less than 2% of creotides, less than 3% of total nucleotides, less than 4% of total nucleotides, or If less than 5% of all nucleotides differ from each other, they are not considered to share nearly identical sequences. It can be done.

[0161] In some embodiments, the first (e.g., first or second) nucleic acid fragments The fragments are aligned to the first position in the reference genome, and each (e.g., the first or second) The second fragment from multiple nucleic acid fragments is aligned to the second position in the reference genome. In one embodiment, the first and second positions correspond to different regions within the reference genome. In some embodiments, the first and second positions are the same position (for example, the first The first and second positions correspond to the same region in the reference genome. In some embodiments, The first and second positions are at least one residue, at least two residues, and a small number of residues in the reference genome. At least 3 residues, at least 4 residues, at least 5 residues, at least 6 residues, at 7 residues, at least 8 residues, at least 9 residues, at least 10 residues, at least 11 residues Base, at least 12 residues, at least 13 residues, at least 14 residues, at least 15 residues Base, at least 16 residues, at least 17 residues, at least 18 residues, at least 19 residues Base, at least 20 residues, at least 30 residues, at least 40 residues, at least 50 residues Base, at least 60 residues, at least 70 residues, at least 80 residues, at least 90 residues The reference genome consists of a base or at least 100 residues. In some embodiments, the first position The second position and the second position overlap by 1 to 50 residues in the reference genome.

[0162] In some embodiments, each fragment is located at at least a first and second location in the reference genome. It is mapped to (for example, the nucleic acid sequence corresponding to each fragment is at least 2 of the reference genome) (They exist in different locations). In some embodiments, each fragment is less than a reference genome. Three locations, at least four locations, at least five locations, at least six locations, at least seven locations Places, at least 8 locations, at least 9 locations, at least 10 locations, at least 11 locations, At least 12 locations, at least 13 locations, at least 14 locations, at least 15 locations, At least 16 locations, at least 17 locations, at least 18 locations, at least 19 locations It is mapped to at least 20 locations. In some embodiments, a small portion of the reference genome At the very least, the two mapped regions have at least one residue in the reference genome, and at the very least... At least 5 residues, at least 10 residues, at least 25 residues, at least 50 residues, Each has 100 residues, at least 200 residues, at least 300 residues, and at least 400 residues. , at least 500 residues, at least 600 residues, at least 700 residues, at least 8 00 residues, at least 900 residues, or at least 1000 residues are spaced apart from each other. In some embodiments, at least two mapped locations are in the reference genome. It contains different genes. In some embodiments, at least two mappings The affected regions are located on different chromosomes in the reference genome.

[0163] Nucleic acid fragments retain the biological activity and / or some characteristics of the parent polynucleotide. It is possible. For example, nasopharyngeal cancer cells can be caused by Epstein-Barr virus (EBV). V) DNA fragments can be targeted, for example, by accumulating in the patient's bloodstream. These fragments are one or more It can contain multiple BamHI-W sequence fragments, which can detect the level of tumor-derived DNA in plasma. It can be used for this purpose. The BamHI-W sequence fragment is recognized using BamHI restriction enzyme. Corresponds to sequences that can be called / digested. The BamHI-W sequence corresponds to sequence 5'-GGATCC-3 It could refer to '.

[0164] Furthermore, polynucleotides are, for example, cfDNA fragments that can naturally exist in biological samples. Multiple segments can be removed through natural processes or in vitro manipulations. It is possible to divide or fragment nucleic acids into parts. Various methods for fragmenting nucleic acids are available in this technology. These methods are well known in the field of technology. These methods are essentially, for example, chemical or physical It may be either enzymatic or denatured. Enzymatic fragmentation includes partial degradation by DNAse; acid Partial depurination by means of purinase; use of restriction enzymes; endonucleases encoded in introns; The cleavage agent is applied to specific parts of the nucleic acid molecule depending on the specific hybridization of the nucleic acid segment. DNA-based cleavage methods such as triple helix and hybrid formation methods that localize the DNA at a specific location; or already This may include other enzymes or compounds that cleave polynucleotides at known or unknown sites. Physical fragmentation methods may include exposing polynucleotides to high shear rates. High shear rates are, for example, in chambers or channels that have pits or spikes. Moving DNA within a channel, or through a channel of limited size, for example, a micron or a small channel. This forces the DNA sample into an aperture with cross-sectional dimensions in the submicron range. This can be caused by other physical methods, such as ultrasonic treatment and nebulization. These include a combination of physical and chemical fragmentation methods, such as thermal fragmentation. Bi-ion-mediated hydrolysis can also be employed in a similar manner. For example, Sambrook et al. "Molecular Cloning:A Laboratory Manual", 3rd Ed.Cold Spring Harbor Laboratory P See ress, Cold Spring Harbor, NY (2001) ("Sambrook et al.") (This is (and shall be incorporated herein by reference for all purposes). These methods This can be optimized to digest nucleic acids into fragments within a selected size range.

[0165] As used herein, the terms “sequence read” or “read” are defined as described herein. or generated by any sequencing process known in the aforementioned technical field Refers to a nucleotide sequence. A read can be generated from one end of a nucleic acid fragment ("single-ended"). Reads are sometimes generated from both ends of a nucleic acid (e.g., paired-end reads, double-end reads). (Dreed). In some embodiments, array reads (e.g., single-ended reads or Paired-end reads (Paired-end reads) can be generated from one or both strands of the target nucleic acid fragment. Column read length is often related to specific sequencing techniques. For example, high Throughput methods provide sequence reads that can vary in size from tens to hundreds of base pairs (bp). In some embodiments, the sequence reads have an average, median, and length of approximately 15 bp to 900 bp. This is the average length (for example, approximately 20 bp, approximately 25 bp, approximately 30 bp, approximately 35 bp, approximately 40 bp, approximately 45 bp, approximately 50 bp, approximately 55 bp). bp, about 60bp, about 65bp, about 70bp, about 75bp, about 80bp, about 85bp, about 90bp, about 95bp, about 100bp, about 1 10bp, about 120bp, about 130bp, about 140bp, about 150bp, about 200bp, about 250bp, about 300bp, about 350bp, about The sequence reads are 400 bp, approximately 450 bp, or approximately 500 bp. In some embodiments, the sequence reads are approximately 100 0 bp, 2000 bp, 5000 bp, 10,000 bp, or 50,000 bp or more, mean, median or This is the average length. Nanopore sequencing is used for large numbers of base pairs, such as tens, hundreds, or thousands of base pairs. It can provide sequence reads that can vary in size. Illumina parallel sequencing is so It can provide sequence reads that do not change much, for example, most of the sequence reads are 200 bp long. It can be made smaller. Sequence reads (or sequencing reads) are nucleic acid molecules ( For example, it can refer to sequence information corresponding to a sequence of nucleotides. A column read is a sequence of nucleotides from a portion of a nucleic acid fragment (for example, approximately 20 to 150 nucleotides). ) may correspond to a sequence of nucleotides at one or both ends of a nucleic acid fragment. , or it can correspond to nucleotides of the entire nucleic acid fragment. Sequence reads can be obtained in various ways, for example , using sequencing technology, or, for example, in a hybridization array Using probes or capture probes, or amplification techniques, such as polymerase chain Using a reaction (PCR) or linear amplification or isothermal amplification using a single primer , you can obtain it.

[0166] As disclosed herein, “sequencing” and “sequence” as used herein Terms such as "SING" generally refer to determining the order of biomolecules such as nucleic acids or proteins. This refers to any and all biochemical processes that can be used for that purpose. For example, sequencing. Data contains all or part of the nucleotide bases in nucleic acid molecules such as DNA fragments. It is possible.

[0167] Where disclosed herein, the term “single nucleotide variant” or “SNV” means “nucleotide A sequence, for example, a single nucleus at a position (e.g., a region) of a sequence read from an individual. This refers to substitution of a nucleotide with a different nucleotide. Specifically, a substitution from the first nucleic acid base X to the second nucleic acid base Y. This is sometimes written as "X>Y". For example, the SNV from cytosine to thymine is "C>T It is sometimes written as "".

[0168] As used herein, the term “methylation profile” (also known as “methylation state”) The information may contain information related to DNA methylation in a particular region. The information includes the methylation index of the CpG site, the methylation density of the CpG site within the region, and the contiguous region. Distribution of CpG sites across regions, and methylation of individual CpG sites within regions containing multiple CpG sites. This may include patterns or levels of methylation, and non-CpG methylation. A significant portion of the genome The methylation profile can be considered equivalent to that of methylomes in the mammalian genome. DNA methylation is the process of methylation of the cytosine heterocycle at position 5 of a CpG dinucleotide. This can refer to the addition of a group (for example, to produce 5-methylcytosine). Methylation of cytosine is For example, 5'-CHG-3' and 5'-CHH-3' (where H is adenine, cytosine, or thymine). This can happen with cytosine in other array contexts such as (is). Methylation can also occur in the form of 5-hydroxymethylcytosine. DNA methylation is N6- This may also include methylation of non-cytosine nucleotides such as methyladenine.

[0169] As used herein, "methylome" refers to multiple locations or loci in the genome. It can be used as an indicator of the amount of DNA methylation in the genome. Methylomes are found in the entire genome. This can correspond to a significant portion of the tumor, or a relatively small portion of the genome. "Oncological methylome" This could be the methylome of a tumor in the subject (e.g., a human). Tumor methylome is found in tumor tissue. Alternatively, it can be determined using cell-free tumor DNA in plasma. Tumor methylome is, This could be an example of the target methylome. The target methylome is a nucleic acid, such as DNA, that enters the body fluids. Organs that can be supplied (e.g., brain cells, bones, lungs, heart, muscles, kidneys, etc.) It could be a methylome of (m). The organ may be a transplanted organ.

[0170] When used herein, each genomic site (e.g., a CpG site, along its 5' → 3' direction) In the linear sequence of the bases, cytosine nucleotides are followed by guanine nucleotides. The term "methylation index" for a DNA region is the total methylation index of the nucleic acid fragments covering that region. This can refer to the ratio of nucleic acid fragments showing methylation at the aforementioned site to the total number of fragments. "Degree" refers to the number of reads in a region showing methylation, and the number of reads covering the region. It can be calculated by dividing by the total number of leads. The part may have specific characteristics. It is possible (for example, the site may be a CpG site). The "CpG methylation density" of a certain region is Cp The number of reads showing G methylation is determined by the number of CpG sites within that region (e.g., a specific CpG site, within a CpG island). Alternatively, it can be calculated by dividing the total number of reads (or CpG sites within a larger region) by the total number of reads covering that region. Yes, it is possible. For example, the methylation density of each 100kb sequence group in the human genome is such that the mapping in the 100kb region is... As the percentage of all CpG sites covered by the coated nucleic acid fragment, the non-variable CpG sites It can be determined from the total number of converted cytosines (which can correspond to methylated cytosines). How many? In that embodiment, this analysis is performed for other sequence group sizes, such as 50kb or 1MB. This is done. In some embodiments, the region is the entire genome or a chromosome or stained. It is a part of the body (for example, a chromosome arm). The methylation index of a CpG site is the region of that CpG site If it includes , it may be the same as the methylation density of the region. The "percentage of methylated cytosine" is For example, the analyzed cytosine residues within a region, including cytosine outside the CpG context. In relation to the total number, those that are methylated (for example, those that are not converted after bisulfite conversion) (i) This can refer to the number of cytosine sites "C" that have been shown to be such. Methylation index, methylation density, The percentage of methylated cytosine is an example of a "methylation level."

[0171] As used herein, "plasma methylome" refers to the plasma of an animal (e.g., human) or This may be methylome determined from serum. Plasma and serum may contain cell-free DNA. Therefore, plasma methylome may be an example of cell-free methylome. Plasma methylome is It can be a mixture of tumor / patient methylomes, and therefore could be an example of a mixed methylome. "Cellular methylome" refers to the cells of the subject, for example, a patient's cells (e.g., blood cells or tumor cells). It may be determined to be methylome. Methylome of blood cells is blood cell methylome (or blood It can be called methylome.

[0172] As used herein, the terms “unusual methylation pattern” or “unusual methylation” are used. A "pattern" is a methylation state that is expected to be found in the sample at a frequency lower than a threshold. Methylation of a DNA molecule having a vector, methylation pattern, or methylation state vector This refers to the morphological state. In the specific embodiments provided herein, a healthy control includes a healthy individual. The probability of finding a specific methylation state vector within a group is expressed by the p-value. In some embodiments, the p-value of the methylation state vector was released on May 22, 2020. The requested "Systems and Methods for Determining Whether a Subject Has a Cancer Congenital The actual article in PCT / US2020 / 034317, titled "Discourse Using Transfer Learning," is available. Example 5, and filed on March 13, 2019, currently under US2019 / 0287652. The US has published a paper titled "Anomalous fragment detection and classification". National Patent Application No. 16 / 352,602 (each of which is incorporated herein in whole by reference) It is determined as described in (assuming it is the case). A low p-value score is due to Generally, in healthy control groups, other methylation status vectors in samples from healthy individuals Compared to the previous example, this corresponds to a relatively unexpected methylation state vector. High p-value score. This generally refers to other methylation states found in samples from healthy individuals in a healthy control group. Compared to the vector, it corresponds to the relatively more expected methylation state vector. Threshold ( For example, methyl The methylation state vector can be defined as an abnormal methylation pattern. Alternatively, to calculate the p-value or expected value of the methylation state vector, as is known in the art Various methods can be used. The exemplary methods provided herein are CpG A Markov chain assuming that the methylation state of a site depends on the methylation state of adjacent CpG sites. This includes the use of chain probability. Alternative methods provided herein involve multiple mixed components (each, It is assumed that methylation at CpG sites is independent of the methylation status at other CpG sites. By using a mixed model that includes an independent site model, it is possible to obtain odors in healthy individuals. The probability of observing a specific methylation state vector is calculated. The method provided herein This uses a genomic region with an unusual methylation pattern. This refers to cfDNA fragments derived from genomic regions appearing in the reference sample at a frequency below the threshold. If a methylation state vector is present, the genomic region will have an unusual methylation pattern. This can be determined. The reference sample may be a sample from a control group or a healthy group. The frequency of the appearance of the methylation state vector in the reference sample can be expressed as a p-value score. Yes, it is possible. A cfDNA fragment corresponding to or derived from a genomic region is a single, uniform me If a chemilation state vector is not present, the genomic region has multiple methylation state vectors. It may have multiple p-value scores. In this case, the multiple p-value scores are compared to the threshold before being compared. They can be summed or averaged. For comparing p-value scores and thresholds corresponding to genomic regions, arithmetic equality is used. We employ various methods known in this field, such as the mean, geometric mean, harmonic mean, median, and mode. These are possible, but are not the only options.

[0173] As used herein, the term “relative abundance” means a specific characteristic (for example, if one is present). Aligning to a specific region of the genome, ending at multiple specific coordinates / terminal positions, A specific amount of nucleic acid fragment having a specific methylation state and a specific length, Features (for example, ending at one or more specific coordinates / terminal positions, or identifying the genome) This refers to the ratio of a second amount of nucleic acid fragment having a specific length (aligned to a region) to a region. It is possible. For example, the relative abundance is the number of DNA fragments ending at the first set of genomic locations. It can refer to the ratio of the number of DNA fragments ending at two sets of genomic locations. In this context, "relative abundance" refers to the amount of cell-free DNA that ends within a certain window of genomic location. The amount of offspring (a certain value) is the amount of cell-free DNA molecules that end within a different window of genome location ( It can be a type of isolation value associated with other values. The two windows can overlap. However, they can be of different sizes. In other embodiments, the two windows cannot overlap. Furthermore, in some embodiments, the window is 1 nucleotide wide, and therefore Therefore, it is equivalent to a single genome location.

[0174] As used herein, the term “methylation” refers to the methylation of water on the pyrimidine ring of a cytosine base. Deoxyribonucleic acid (DNA) is formed when elementary atoms are converted to methyl groups, creating 5-methylcytosine. This refers to the modification of the cytosine-guanine dinucleotide (as specified herein). In particular, methylation refers to the modification of the cytosine-guanine dinucleotide (as specified herein). This tends to occur at the "CpG site" (which is called the "CpG site"). In other examples, methylation occurs at a part of the CpG site. This can occur with cytosine, or with other nucleotides other than cytosine; however, these occurrences This is rare. In this disclosure, methylation is discussed with reference to the CpG site for clarification. Unusual cfDNA methylation can be identified as either hypermethylation or hypomethylation, and both can occur. , which may indicate a cancerous state. As is well known in this art, (healthy control) DNA methylation abnormalities (compared to others) cause different effects and may contribute to cancer. ru.

[0175] Identifying abnormally methylated cfDNA fragments presents various challenges. First, the target cfDNA is different... Determining that something is always methylated is only important when comparing it to a control group. Furthermore, when the number of control groups is small, the judgment becomes unreliable in small control groups. Within the control group, the methylation status of the subjects varied, and the cfDNA of the subjects was abnormally methylated. In some cases, it can be difficult to consider when making a determination. Also, the cytosine in a certain CpG site Chilling is thought to be causally related to the subsequent methylation of the CpG site.

[0176] Those skilled in the art will know that the principle described herein involves non-CpG methylation, including non-cytosine methylation. You will understand that this is equally applicable to the detection of methylation in text.

[0177] Where disclosed herein, the term “subject” means any living or non-living thing, including humans (e.g., For example, male humans, female humans, fetuses, pregnant women or children, non-human animals, plants This includes, but is not limited to, substances, bacteria, fungi, or protists. Any human or Non-human animals can function as targets, including mammals, reptiles, birds, amphibians, fish, and non-human animals. Undactyls, ruminants, cattle (e.g., cows), horses (e.g., horses), goats and sheep (e.g., Sheep, goats, pigs (e.g., pigs), camels (e.g., camels, llamas, alpacas), monkeys Great apes (e.g., gorillas, chimpanzees), bears (e.g., bears), poultry, dogs, cats, maws This refers to rats, fish, dolphins, whales, and sharks. The terms "subject" and "patient" are used in this document. Used interchangeably in books, for example, to indicate that a person has a medical condition or disorder such as cancer. Refers to a human or non-human animal that is or potentially possesses a human or non-human animal. In some embodiments The subject is a male or female of any stage (e.g., male, female, or child). ru.

[0178] A sample is taken, or by any of the methods or compositions described herein. The subjects treated can be of any age and can be adults, infants, or children. In that case, the subjects, for example, the patients, were 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 , 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 3 7, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50 , 51, 52, 53, 54, 53, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 7 6, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 years old, or within that range. (For example, approximately 2 to 20 years old, approximately 20 to 40 years old, or approximately 40 to 90 years old) Certain classes of subjects who may benefit from the methods disclosed herein, for example, patients, are 40 years of age or older. The object of this hierarchy is, for example, a patient.

[0179] Another specific class of subjects who may benefit from the methods of this disclosure, such as patients, These are pediatric patients who may be at a higher risk of developing cardiac symptoms. Furthermore, samples are taken. or subjects treated by any of the methods or compositions described herein, e.g. For example, the patient could be male or female.

[0180] As used herein, the term "normalization" means normalizing values ​​or sets of values ​​to a common standard for comparative purposes. This means converting to a reference frame. For example, if the diagnostic ctDNA level is baseline ctD When "normalized" at the NA level, the diagnostic ctDNA level is the baseline. In order to determine different amounts from the intrinsic ctDNA level, the baseline ctDNA level and the ratio They are compared.

[0181] As used herein, the terms “cancer” or “tumor” mean an abnormal mass of tissue, and that This refers to a condition where the growth of a mass exceeds the growth of normal tissue and is not coordinated. Cancer or tumors are defined as follows: Characteristics: Depending on the degree of cell differentiation, including morphology and functionality, proliferation rate, local invasion and metastasis They can be defined as "benign" or "malignant." Benign tumors are highly differentiated, Compared to malignant tumors, they are characterized by slower growth and localization to the site of origin. Furthermore, Benign tumors may not have the ability to invade, invade, or metastasize to distant sites. Malignant tumors, on the other hand, may not have the ability to spread to distant sites. They have a low degree of phagocytosis (anaplasia) and rapidly proliferate while invading, invading, and destroying surrounding tissues. It may be characterized by the following. Furthermore, malignant tumors may also have the ability to metastasize to distant sites.

[0182] As used herein, the term "level of cancer" means whether or not cancer is present (for example) (Presence or absence), stage of cancer, size of tumor, presence or absence of metastasis, total tumor burden in the body This refers to the level of cancer and / or other indicators of the severity of the cancer (e.g., cancer recurrence). The "L" can be a number, or other indicator such as a symbol, alphabet letter, or color. The value can be zero. The level of cancer is related to the mutation or the number of mutations in pre-malignant cancers. Alternatively, it can also include precancerous conditions. The cancer level can be used in various ways. It is possible to screen for cancer that has not been previously known to be cancerous. It is possible to determine whether or not a person has cancer. The evaluation involves surveying people who have been diagnosed with cancer. This allows for monitoring the progression of cancer over time, investigating the effectiveness of treatment, or determining the prognosis. In one embodiment, prognosis is the probability that the subject will die from cancer, or within a certain period of time. Alternatively, it can be expressed as the probability that cancer will progress after a certain amount of time, or the probability that cancer will metastasize. Detection may include "screening" or suggestive features of cancer (e.g., symptoms). This includes determining whether a person with a positive test result (or other positive test) has cancer. obtain.

[0183] The terms "cancer load," "tumor load," "cancer volume," and "tumor volume" are used in the test sample. The term is used interchangeably herein to refer to the concentration or presence of tumor-derived nucleic acids. Thus, the terms "cancer load," "tumor load," "cancer volume," and "tumor volume" refer to the biological load. This is a non-limiting example of the cell source percentage or tumor percentage in the sample. In some embodiments, Therefore, tumor percentage is a specific version of cell source percentage.

[0184] As used herein, the term "tissue" refers to a group of cells that function as a single unit. Yes. Multiple types of cells may be found in a single tissue. Different types of tissues, It may also consist of different types of cells (e.g., hepatocytes, alveolar cells, or blood cells), but This can also correspond to tissue from one organism (mother versus fetus) or healthy cells versus tumor cells. The term "tissue" generally refers to any group of cells found in the human body (e.g., heart tissue, lung tissue). It can refer to renal tissue, nasopharyngeal tissue, oropharyngeal tissue. In some embodiments, The terms "tissue" or "tissue type" are used to refer to the tissue from which the cell-free nucleic acid originates. Obtainable. In one example, viral nucleic acid fragments may originate from blood tissue. In another example, viral nucleic acid The fragments may originate from tumor tissue.

[0185] As used herein, the term “untrained classifier” means target data This refers to a classification system that has not been trained in a set. However, an untrained classification system is a primary data Partially train on a dataset (e.g., a small and / or reference dataset) This can happen. The term "untrained classifier" refers to an untrained classifier. In such training, we cannot rule out the possibility that transfer learning techniques may be used. This will be understood. For example, Fernandes et al., 2017, “Transfer Learning with Par tial Observability Applied to Cervical Cancer Screening,” Pattern Recognition a Image Analysis: 8 th Iberian Conference Proceedings, 243-250 (see this document for reference) (To be incorporated into the book) provides an unrestricted example of such transfer learning. When transfer learning is used, the untrained classifier receives the primary training data. Additional data beyond the set of data is provided. Typically, this additional data is separate. The format of coefficients (e.g., regression coefficients) learned from the auxiliary training dataset is Furthermore, although we disclosed a description of a single auxiliary training dataset, this is the final version. In the demonstration, when training an untrained classification device, the primary training data... There is no limit to the number of auxiliary training datasets that can be used to complement the main dataset. This will be understood. For example, in some embodiments, two or more auxiliary training days A dataset, three or more auxiliary training datasets, four or more auxiliary training data A set, or five or more auxiliary training datasets, is used for primary training through transfer learning. Used to complement training datasets, each such auxiliary dataset is a primary dataset. This is different from the training dataset. In such embodiments, any method Transfer learning may be used. For example, in addition to the primary training dataset, the first supplemental dataset may be used. Let's consider the case where an auxiliary training dataset and a second auxiliary training dataset exist. The coefficients learned from the first auxiliary training dataset (first auxiliary training dataset) (By applying classification devices such as regression to a dataset) is a transfer learning technique (for example, the two mentioned above) It was applied to the second auxiliary training dataset using (multiplication of dimensional matrices), and as a result, The number then becomes a trained intermediate classifier that is applied to the primary training dataset. This is then applied to the untrained classifier along with the primary training dataset itself. Alternatively, the first set of coefficients learned from the first auxiliary training dataset (1st (By applying classification mechanisms such as regression to auxiliary training datasets) and a second auxiliary tray The second set of coefficients learned from the training dataset (second auxiliary training dataset) (By applying classification devices such as regression to a specific target) Each of these is individually trained on a primary training data set. It may also be applied to separate instances of the matrix (for example, by multiplication of separate, independent matrices). Furthermore, the primary training dataset itself (or the data learned from the primary training set) (Some form of reduction of the primary training dataset, such as principal components or regression coefficients) Together, the application of coefficients to separate instances of the primary training dataset Both then train the untrained classifier. This may also be applied to a classification device. In any example, first and second auxiliary training data Knowledge about the cell source from which the set originates (e.g., type of cancer) is labeled to the cell source. Along with the deleted primary training dataset, the untrained classifier is trained It is used for rubbing.

[0186] The term "classification" refers to any number or other character associated with a specific characteristic of a sample. It can indicate that the sample has deletions or amplifications. For example, the "+" symbol (or the word "positive") indicates that the sample has deletions or amplifications. It can indicate that something is classified as having something. In another example, the term "classification" can mean that the subject and / or the amount of tumor tissue in the sample, the target and / or the size of the tumor in the sample, The stage of the tumor in the elephant, the tumor burden in the subject and / or sample, and the tumor in the subject. This refers to the presence of tumor metastasis. In some embodiments, the classification is binary (e.g., positive or negative). (Sex) or more classification levels (e.g., scales of 1-10 or 0-1) It has ( ). In some embodiments, the terms "cutoff" and "threshold" are used in the operation. This refers to a predetermined numerical value used. For example, the cutoff size is the size from which fragments are excluded. This refers to the threshold. In some embodiments, a specific classification is applied above or below the threshold. It refers to a value. Any of these terms can be used in any of these contexts.

[0187] As used herein, the terms “cancer-related changes” or “cancer-specific changes” include: Mutations originating from (single nucleotide mutations, nucleotide deletions or Insertions, deletions of genes or chromosomal segments, translocations, inversions, gene amplification, virus Russ-related sequences (e.g., viral episomes, viral inserts, infecting cells and subsequently cells) Viral DNA released from (circulating or cell-free viral DNA), abnormal methylation pro File or tumor-specific methylation signature, abnormal cell-free nucleic acids (e.g., DNA) ) Size profile, abnormal histone modification marks and other epigenetic modifications, This may include the terminal locations of cancer-related or cancer-specific cell-free DNA fragments.

[0188] Where used herein, the terms “comparison,” “comparison sample,” “reference,” and “reference sample” are used with respect to the same context. "Pull," "normal object," and "normal sample" refer to items that do not have a specific condition, i.e., are normal. This represents a sample from a healthy subject. For example, a method such as the one disclosed herein This can be performed on subjects with tumors, where the reference sample is a healthy sample of the subject. These are samples taken from the same tissue. Reference samples can be obtained from the subject or from a database. It is also possible to obtain it from the sample. For example, nucleic acids obtained from a sample of the subject. It may be a reference genome used to map fragments. The reference genome is a living organism Nucleic acid fragments from pull and constitutional samples can be aligned and compared. This can refer to a ploid or diploid genome. An example of a constitutional sample is a leukocyte obtained from a subject. It could be DNA. In the case of a haploid genome, there can be only one nucleotide at each gene locus. In the case of a diploid genome, heterozygous loci can be identified; each heterozygous locus is It can have two alleles, and both alleles can align to the gene locus. It is possible to make them match.

[0189] Several aspects are described below with reference to the application of examples for illustrative purposes. Details, relationships, and methods are described herein to provide a complete understanding of the features described herein. Please understand that it will be published. However, if you are skilled in the art in the relevant technical field The features described herein may not be present without one or more of the specific details, or by any other means. It will be easy to recognize that this can be done using. Some actions can be done in different orders. Features described herein may occur simultaneously with or / or other actions or events. This is not limited by the order in which the actions or events are shown. Furthermore, as described herein In order to implement a methodology that follows these characteristics, not all of the actions or events shown are necessary. There isn't one.

[0190] Exemplary System Embodiment

[0191] Here, we will describe the details of an exemplary system with reference to Figure 1. Figure 1 shows several implementation forms. This is a block diagram illustrating system 100 according to the configuration. Vice 100 consists of one or more processing units CPU 102 (processor or processing unit (Also called A), and one or more network interfaces 104, and user Interface 106, non-persistent memory 111, persistent memory 112, and their configuration It includes one or more communication buses 114 for interconnecting the elements. The communication bus 114 optionally interconnects and controls communication between system components (chips) This includes non-persistent memory 111 (sometimes called a set). DRAM is a typical example. High-speed RAM such as SRAM, DDR RAM, ROM, EEPROM, and flash memory. It includes dumb access memory, while persistent memory 112 is typically a CD-ROM. Digital Multipurpose Discs (DVDs) or other optical storage, magnetic cassettes, magnetic tapes , magnetic disk storage or other magnetic storage device, magnetic disk storage device Optical disc storage devices, flash memory devices, or other non-volatile solid storage A memory device is included. Persistent memory 112 is optionally located remotely from CPU 102. Includes one or more storage devices. Persistent memory 112 and non-persistent memory 11 The non-volatile memory devices within 2 constitute a non-transient computer-readable storage medium. In that implementation, non-persistent memory 111 or alternatively non-transient computer-readable memory is used. The medium contains the following programs, modules, and data structures, or subsets thereof, at time It is stored in combination with persistent memory 112: • Procedures for handling various basic system services and hardware-dependent tasks Optional operating system 116, which includes instructions for running the program; • Options for connecting System 100 to other devices or communication networks. Network communication module (or instruction) 118; • To determine the proportion of test cell sources (158) in 140 biological samples being tested. Cell source proportion estimation module 120; • Each training target is 124 (for example, 124-1, ..., 124-Z (where Z is 1) Each training target has 126 cell-free fragments (for example, each cell-free fragment is a positive integer greater than 126) Then, 126 - 1 - X, ..., 126 - 1 - Y (where X and any positive integer, Y is X) For the largest of the two, at least (i) each CpG site of each cell-free fragment From each methylation state of 130 (for example, 130-1-XA, ..., 130-1-XQ) (ii) Each train Training data set including the corresponding cancer symptoms 136 for each of the 136 target subjects. 122. • Each of the multiple cell-free fragments derived from the biological sample under test 14 2 (For example, 142-G, ..., 142-H (where G and H are positive integers and H is greater than G) (i) of each cell-free fragment, (i) at least each CpG site 148 (e.g., 1 46-GM,..., 146-GN,..., 146-HO,...146H-P (here Each methylation is determined from each methylation state (where M, N, O, and P are positive integers). Pattern 144 (for example, 144-G, ..., 144-H) and (ii) each sequence group PING 148 (e.g., 148-G, ..., 148H) and (iii) each predicted sec The cancerous state of the refractory fragments 150 (e.g., 150-G, ..., 150-H) and the The target dataset 140, wherein the first representative value 152, the second representative value 154, and the estimated This product further contains a constant cell source ratio of 156.

[0192] In accordance with this disclosure, the corresponding sequence group mapping 132 (e.g., 13) for each cell-free fragment is used. 2-1-X) and the cancerous state of each cell-free fragment 134 (e.g., 1 The assignment 34-1-X) is made. For convenience and ease of interpretation, these data The construct is shown as being present in the training dataset. However, typical In a typical embodiment, such a data construct is a cell-free fragment within the training set. It is calculated from the methylation pattern and is not part of the original dataset. In other embodiments, The cancer status of sequence group mapping 132 and cell-free fragments was obtained from the training. This is part of dataset 122.

[0193] According to some implementations, one or more of the above-mentioned elements are the aforementioned A set of instructions stored in one or more Mori devices to perform the functions described above. Corresponds to the above-identified module, data, or program (e.g., instruction set). A set (or a set of separate software programs, procedures, datasets, or modules) is a set of separate software programs, procedures, datasets, or modules. It does not need to be implemented in this way, and therefore, various subsets of these modules and data The set can be combined in various implementation forms, or it may be rearranged. In some implementations, non-persistent memory 111 may optionally be the module identified above. It stores a subset of the data structures. Furthermore, in some embodiments, memory This stores additional modules and data structures not described above. In the embodiment, one or more of the identified elements are required by the visualization system 100. Visualization system 1 A computer system addressable by the visualization system 100, separate from the one in 00. It is stored in the stem.

[0194] Figure 1 depicts "System 100," but this figure represents the implementation configuration described herein. Rather than a structural schematic diagram, it provides a functional description of the various features that can exist in a computer system. This is intended to be done. In practice, as will be recognized by those skilled in the art, they are shown separately. The items can be combined, and some items can be separated. Furthermore, in Figure 1, non-persistent Specific data and modules in Mori 111 are depicted, but these datasets The and / or some or all of the module may reside in persistent memory 112.

[0195] The system in accordance with this disclosure has been disclosed with reference to Figure 1, and next, the method in accordance with this disclosure will be disclosed. This will be explained in detail with reference to Figures 2A and 2B and Figures 3A and 3B. Either method determines whether the subject has cancer or whether the subject has cancer. To determine the possibility, U.S. Patent Application No. 15, filed on October 25, 2017 International patent No. 793,830 and / or International Filing Date of October 24, 2017 Publication No. PCT / US17 / 58099 (as incorporated herein by reference) Using or in conjunction with any of the assays or algorithms disclosed in (the following document) It will become clear that it is possible to move.

[0196] Identifying features for estimating cell source proportions

[0197] Block 202. One aspect of the present disclosure comprises one or more processors and one or more The memory that stores one or more programs for execution by the processor This is performed on a computer system to estimate the proportion of cell sources for a given subject. This provides a method for identifying multiple features for this purpose.

[0198] In some embodiments, the cell source ratio of block 202 in Figure 2A is the common primary site Corresponds to the first cancer state. In some embodiments, the cell source proportion is a tumor of a certain type of cancer. The proportion of ulcers, or the proportion thereof. In some embodiments, the proportion of cell sources is a predetermined number of Corresponds to the tumor percentage of the first cancer state of the stage. In some embodiments, the cell source percentage It is derived from one or more types of human cells.

[0199] Subject and cancer status

[0200] Block 204. In Block 204 of Figure 2A, this method is used for electronic training. The process proceeds by obtaining a training dataset. The training dataset consists of multiple training datasets. For each training target, at least a) the corresponding training a) Each of the multiple cell-free fragments, the corresponding methylation pattern of each cell-free fragment, and b) each The training subjects include signs of the target cancer, where the state of the target cancer is the first It is either a state of cancer or a state of a second type of cancer.

[0201] According to block 206, in some embodiments, multiple training subjects are 10 It consists of ~1000 training subjects. In some embodiments, multiple training The target group includes at least 10 training subjects, at least 25 training subjects, and a minimum of At least 50 training targets, at least 100 training targets, at least 25 0 training targets, at least 500 training targets, at least 750 training Training target, at least 1000 training targets or at least 1500 training It consists of training targets. In some embodiments, there are multiple training targets, 10 to 1 00,000 training targets, 100-50,000 training targets, or 1 It consists of 00 to 10,000 training targets.

[0202] In some embodiments, in multiple training subjects, the first cancer state and The number of training subjects with a second cancerous condition is balanced (for example, multiple trainees The training subjects include substantially the same number of training subjects with the same cancer status as each subject. For example, multiple training subjects include at least 50 trainees who have a first cancer condition. If the training subjects are included, the multiple training subjects have a second cancerous condition. This includes 50 training subjects, or multiple training subjects with a primary cancer condition. If the training subjects include at least 500 training subjects having, the plurality of training subjects This also includes at least 500 training subjects who have a secondary cancer condition. In this embodiment, 5% to 95% of the trainees have the first cancer state, and the remainder have the second. The cancerous condition is present. In some embodiments, 20 percent of the training subjects Approximately 80 percent have the first type of cancer, and the remainder have the second type of cancer. In some embodiments, 30 to 70 percent of the training subjects are first One has a cancerous state, and the rest have a second cancerous state. In some embodiments, the tray 40 to 60 percent of those undergoing the study had a primary cancer, and the remainder had a secondary cancer. The cancerous condition is present. In some embodiments, 45 percent of the training subjects 55 percent have a primary cancer, and the remainder have a secondary cancer.

[0203] Referring to block 208, in some embodiments, the first cancer state is from cancer Therefore, the second cancer state is the absence of cancer. In some embodiments, the first cancer The conditions include adrenal cancer, biliary tract cancer, bladder cancer, bone / bone marrow cancer, brain cancer, breast cancer, cervical cancer, Colon cancer, esophageal cancer, stomach cancer, head / neck cancer, hepatobiliary tract cancer, kidney cancer, liver cancer, lung cancer, egg cancer Focal cancer, pancreatic cancer, pelvic cancer, pleural cancer, prostate cancer, kidney cancer, skin cancer, stomach cancer, testicular cancer Cancer, thymic cancer, thyroid cancer, uterine cancer, lymphoma, melanoma, multiple myeloma, or leukemia The first is a disease, and the second cancer state is the absence of cancer. In some embodiments, the first is The condition is that there is a certain stage of adrenal cancer, a certain stage of biliary tract cancer, and a certain stage of bladder cancer. Hmm, bone / bone marrow cancer at a certain stage, brain cancer at a certain stage, breast cancer at a certain stage, Stage 1 cervical cancer, stage 2 colorectal cancer, stage 3 esophageal cancer, stage 4 Stomach cancer, head / neck cancer at a certain stage, hepatobiliary tract cancer at a certain stage Kidney cancer, liver cancer at a certain stage, lung cancer at a certain stage, ovarian cancer at a certain stage, Stage 1 pancreatic cancer, stage 2 pelvic cancer, stage 3 pleural cancer, stage 4 Prostate cancer, kidney cancer at a certain stage, skin cancer at a certain stage, stomach cancer at a certain stage , a certain stage of testicular cancer, a certain stage of thymic cancer, a certain stage of thyroid cancer, Stage uterine cancer, stage lymphoma, stage melanoma, stage Multiple myeloma is a type of leukemia at a certain stage, and the second cancerous state is the absence of cancer. be.

[0204] In some forms, the second type of cancer is adrenal cancer, biliary tract cancer, bladder cancer, Bone / bone marrow cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, stomach cancer, head / neck Hmm, hepatobiliary tract cancer, kidney cancer, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, pelvic cancer, pleural cancer, Prostate cancer, kidney cancer, skin cancer, stomach cancer, testicular cancer, thymic cancer, thyroid cancer, uterine cancer, It is one of the following: lymphoma, melanoma, multiple myeloma, or leukemia. Several implementation forms In terms of the second stage of cancer, there is adrenal cancer at a certain stage, biliary tract cancer at a certain stage, and Stage bladder cancer, stage 1 bone / bone marrow cancer, stage 1 brain cancer, stage 1 Breast cancer at a certain stage, cervical cancer at a certain stage, colon cancer at a certain stage, esophageal cancer at a certain stage Cancer, a certain stage of stomach cancer, a certain stage of head / neck cancer, a certain stage Hepatobiliary tract cancer, kidney cancer at a certain stage, liver cancer at a certain stage, lung cancer at a certain stage, Ovarian cancer at a certain stage, pancreatic cancer at a certain stage, pelvic cancer at a certain stage, Pleural cancer, prostate cancer at a certain stage, kidney cancer at a certain stage, skin cancer at a certain stage Cancer, a certain stage of stomach cancer, a certain stage of testicular cancer, a certain stage of thymic cancer, Stage thyroid cancer, stage uterine cancer, stage lymphoma, stage One of the following: melanoma, multiple myeloma at a certain stage, or leukemia at a certain stage be.

[0205] In some embodiments, the cancer state in question is a first cancer state, a second cancer state , and one of the third cancerous conditions. In some embodiments, multiple trays The cancer status of each training subject is determined from the status of multiple cancers. They are selected separately. In some such embodiments, multiple training subjects are multiple The training subjects include at least a minimum number of subjects that have each cancer state. In one embodiment, the minimum number of training subjects having each cancer state is at least 1 0, at least 20, at least 30, at least 40, at least 50, at least 6 0, at least 70, at least 80, at least 90, at least 100, at least 150, at least 200, at least 250, at least 300, at least 350, At least 400, at least 450, or at least 500 training subjects ru.

[0206] In some embodiments, multiple cancer conditions are at least 5, at least 10, and This includes at least 20 unique cancer conditions. In some embodiments, multiple cancers This includes 22 specific cancer conditions.

[0207] In some embodiments, each of the multiple cancer conditions is adrenal cancer, biliary tract cancer, etc. Hmm, bladder cancer, bone / bone marrow cancer, brain cancer, breast cancer, cervical cancer, colon cancer, esophageal cancer, stomach cancer Hmm, head / neck cancer, hepatobiliary tract cancer, kidney cancer, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, pelvic cancer Hmm, pleural cancer, prostate cancer, kidney cancer, skin cancer, stomach cancer, testicular cancer, thymus cancer, thyroid cancer It is one of the following: uterine cancer, lymphoma, melanoma, multiple myeloma, or leukemia. In some embodiments, each of the multiple cancer states is a certain stage of adrenal cancer, A certain stage of bile duct cancer, a certain stage of bladder cancer, a certain stage of bone / bone marrow cancer, Brain cancer at a certain stage, breast cancer at a certain stage, cervical cancer at a certain stage Colon cancer, esophageal cancer at a certain stage, stomach cancer at a certain stage, head / neck cancer at a certain stage Hmm, hepatobiliary tract cancer at a certain stage, kidney cancer at a certain stage, liver cancer at a certain stage, Stage 1 lung cancer, stage 1 ovarian cancer, stage 1 pancreatic cancer, stage 1 bone cancer Distal cancer, pleural cancer at a certain stage, prostate cancer at a certain stage, kidney cancer at a certain stage, A certain stage of skin cancer, a certain stage of stomach cancer, a certain stage of testicular cancer, a certain stage Thymic cancer of a certain stage, thyroid cancer of a certain stage, uterine cancer of a certain stage, thyroid cancer of a certain stage Melanoma, melanoma at a certain stage, multiple myeloma at a certain stage, or white blood cell carcinoma at a certain stage It is one of the blood diseases.

[0208] Cell-free fragment acquisition and methylation sequencing

[0209] Referring again to block 204, for each training target, each corresponding training Among the multiple cell-free fragments used for methylation, the corresponding methylation pattern of each cell-free fragment is: (i) One sample containing each fragment from the corresponding biological sample obtained from each training subject (ii) In each fragment, This includes the methylation state of each CpG site among the corresponding multiple CpG sites.

[0210] In some embodiments, the corresponding biological sample is a liquid biological sample. In one embodiment, the corresponding biological sample is a blood sample. So, the corresponding biological samples are blood, whole blood, plasma, serum, urine, and brain spinal cord of the subject being trained. It consists of cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites. In some embodiments, The corresponding biological samples are blood, whole blood, plasma, serum, urine, and brain / spinal cord samples used for training. It consists of fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites.

[0211] In some embodiments, one of the corresponding biological samples from the training subject is used. Multiple nucleic acid samples are obtained from cell-free nucleic acid samples (e.g., liquid biological samples). (These are the cells obtained from biological samples.) In some embodiments, cell-free nucleic acids are obtained from biological samples. is any form of nucleic acid as defined in this disclosure, or any combination thereof. For example, In some embodiments, cell-free nucleic acids obtained from biological samples are a mixture of RNA and DNA. It is an object.

[0212] Multiple cell-free fragments corresponding to each training target are used in the biological sample. In some embodiments derived from cell-free nucleic acids from a liquid biological sample (e.g., a liquid biological sample) It is advantageous that cell-free nucleic acids show an evaluable proportion of cell sources. Several implementations Morphologically, regarding the first or second cancer status of the corresponding training subject. The cell source percentage should be at least 2 percent, at least 5 percent, and at least 10 percent. cents, at least 15 percent, at least 20 percent, at least 25 percent at least 50 percent, at least 75 percent, at least 90 percent The percentage is at least 95 percent, or at least 98 percent.

[0213] In some embodiments, the biological sample is prepared for sequencing analysis by providing cell-free nuclei. It is treated to extract the acid. In some non-limiting examples, in some embodiments, self Lei nucleic acid fragments are obtained from biological samples (e.g.,) collected from subjects in K2 EDTA tubes. Extracted from a blood sample. If the biological sample is blood, the sample will be taken 2 days after collection. Within the allotted time, first rotate the biological sample in a double-rotation system at 1000g for 10 minutes, and then the obtained blood The plasma is processed by rotating it at 2000g for 10 minutes. After that, the plasma is divided into 1 mL. Stored in aliquots at -80°C. In this way, for the purpose of cell-free nucleic acid extraction. Next, an appropriate amount of plasma (e.g., 1-5 mL) is prepared from the biological sample. In such embodiments, cell-free nucleic acids are obtained using the QIAamp Circulating Nucleic Acid Kit (Qiagen). It is extracted and eluted in DNA suspension buffer (Sigma). In some embodiments, it is purified. Cell-free nucleic acids are stored at -20°C until use. For example, Swanton, et al., 2 017, "Phylogenetic ctDNA analysis depicts early stage lung cancer evolution," Na See ture, 545(7655):446-451 (this is incorporated herein by reference). (Assuming...) To prepare cell-free nucleic acids from biological methods for sequencing purposes, Other equivalent methods may be used, all of which fall within the scope of this disclosure.

[0214] In one embodiment, a cell-free nucleic acid fragment converts unmethylated cytosine to uracil. The process is carried out in such a way. In one embodiment, the method is carried out without converting methylated cytosine to non-methylated cytosine. This involves using bisulfite treatment of DNA to convert cytosine uracil to uracil. For example, EZ DN A Methylation(trademark)-Gold, EZ DNA Methylation(trademark)-Direct, or EZ DNA Methyl ation (trademark) - Lightning Kit (Zymo Research Corp (Irvine, California)) Commercially available kits (such as those available from) are used for bisulfite conversion. Another embodiment Therefore, the conversion of unmethylated cytosine to uracil is achieved using enzymatic reactions. For example, The conversion is done using non-metamorphic methods such as APOBEC-Seq (NEBiolabs, Ipswich, Massachusetts). Using commercially available kits for the conversion of chylated cytosine to uracil is possible. can.

[0215] A sequencing library is prepared from the converted cell-free nucleic acid fragments. (Optional) The sequencing library uses multiple hybridization probes, Cell-free nucleic acid fragments or genomic regions that are informative about cellular origin are enriched. Hybridization probes are specifically designated cell-free nucleic acid fragments, or standard They hybridize into target regions, and these fragments are then used for subsequent sequencing and analysis. or is a short oligonucleotide that concentrates the region. In some embodiments, Ebridization probes provide valuable information about cellular origins by targeting specific CpG sites. It is used to perform target high-depth analysis of the net. Once prepared, sequencing The library or a portion of it is sequenced to obtain multiple sequence reads.

[0216] In some embodiments, sequence reads obtained from the biological sample in question are used as a reference set. (For example, data obtained from multiple reference groups, such as a control cohort of healthy subjects) It will be normalized. (From "Method and system for selecting," published on September 19, 2019) U.S. Patent Application No. 20, entitled "Managing, and Analyzing Data of High Dimensionality" No. 19-0287649 (This is incorporated herein by reference in its entirety) (This document) discloses several methods of normalization.

[0217] In some embodiments, the number of sequence reads is at least 100, at least 500 , at least 1000, at least 2000, at least 3000, at least 4000 , at least 5000, at least 6000, at least 7000, at least 8000 , at least 9,000, at least 10,000, at least 20,000, at least Includes 50,000, at least 100,000, or at least 1 million sequence reads In some embodiments, multiple sequence reads are at least 5 million, at least 1 Includes 0 million, or at least 100 million, sequence reads.

[0218] In some embodiments, among a plurality of training targets, each training target is Multiple cell-free fragments for the runners: at least 100, at least 500, at least Also 1000, at least 2000, at least 3000, at least 4000, and at least 5000, at least 6000, at least 7000, at least 8000, and at least 9000, at least 10,000, at least 20,000, at least 50,000 0, at least 100,000, at least 1 million, at least 5 million, or less It contains at least 10 million cell-free fragments. In some embodiments, multiple training Among the training targets, the number of self-free fragments for each training target is small. 100 each, at least 500, at least 1000, at least 2000, at least 3000, at least 4000, at least 5000, at least 6000, at least 7000, at least 8000, at least 9000, at least 10,000, less 20,000 each, at least 50,000, at least 100,000, at least 1 It contains 0 million, at least 5 million, or at least 10 million cell-free fragments.

[0219] In some embodiments, the first training target among multiple training targets is: Having a first number of cell-free fragments, and having a first corresponding plurality of cell-free fragments, The second training target in the training target is a second number of self that is different from the first number. Having a second corresponding set of cell-free fragments including a Lee fragment (for example, several implementations) Morphologically, each training target has multiple self-free fragments for different training purposes. ).

[0220] In some embodiments, the corresponding number of self-free fragments for training are each , having an average length of less than 500 nucleotides. In some embodiments, the corresponding trace The multiple cell-free fragments for the run are 100, 200, 300, 400, and 50 respectively. They have an average length of less than 0, 600, 700, 800, 900, or 1000 nucleotides. do.

[0221] In some embodiments, the sequencing includes methylation sequencing.

[0222] In some embodiments, methylation sequencing is performed on each fragment, or multiple 5-methylcytosine (5mC) and / or 5-hydroxymethylcytosine (5mC) To detect hmC). In some such embodiments, the methylation sequencing is performed. In the sequence reads of each fragment, one or more unmethylated cytosines or one or more This further involves the conversion of multiple methylated cytosines to one or more corresponding uracils. In some embodiments, one or more uracils are one or more corresponding It is detected as thymine during methylation sequencing. In some embodiments, 1 Conversion of one or more unmethylated cytosines or one or more methylated cytosines This includes chemical transformations, enzymatic transformations, or combinations thereof. In some embodiments, The cytosine conversion is described in the "Systems and Methods for De U.S. Patent Application No. 62 / 877,755, entitled "Termining Tumor Fraction" (This is a reference to U.S. Patent Application No. 62 / 877,755) This shall be done as described herein (which shall be incorporated herein by reference).

[0223] In some embodiments, among the multiple corresponding CpG sites in each fragment, each CpG site The methylation status is determined by (i) if each CpG site is methylated, methylation sequencing will be performed. Therefore, if it is determined that the state is methylated, then (ii) each CpG site is not methylated It is in a non-methylated state if determined by methylation sequencing, and / (iii) The methylation status of each CpG site can be referred to as methylated or unmethylated. If none of the above apply, it will be flagged as "Other".

[0224] In some embodiments, methylation sequencing (e.g., determining the methylation pattern) is performed. The method used for this purpose is paired-end sequencing. In some embodiments, Methylation sequencing is a single-read sequencing method. Several implementations In this context, methylation sequencing is a type of whole-genome methylation sequencing (for example, whole-genome sequencing). This is (Mbaisulfite sequencing).

[0225] Whole-genome sequencing assays involve sequencing the entire genome or a significant portion of the entire genome. This generates sequence reads and determines major mutations such as copy number variations or copy number anomalies. This refers to a physical assay that can be used for the whole genome. Such physical assays can be used for the whole genome. Sequencing technology or whole-exome sequencing technology can be employed.

[0226] In some embodiments, whole-genome methylation sequencing was performed, for example, in March 2019. The application filed on the 13th of the month and currently published as US2019 / 0287652 is "Anomal U.S. Patent Application No. 16 / 352, entitled "usual fragment detection and classification" As described in Article 602 (which is incorporated herein by reference in its entirety) As shown, identify one or more methylation state vectors.

[0227] In some embodiments, sequencing is performed from nucleic acids (e.g., cell-free nucleic acids). This includes any form of sequencing that can be used to obtain several defined sequence reads. This is not limited to the Roche 454 platform, Applied Biosyst ems SOLiD platform, Helicos True Single Molecule DNA sequencing technology Which high-throughput sequencing system, Affymetrix Inc.'s sequencing bar I Hybridization Platform, Pacific Biosciences' single-molecule real-time SMRT technology, 54 Life Sciences, Illumina / Solexa and Helicos Biosciences Sequencing-by-synthesis platform, Applied Biosystems' sequencing Global ligation platform, Life Technologies' ION TORRENT technology, This includes sequencing and / or nanopore sequences. In some embodiments, sequencing This is sequencing-by-synthesis and reversible terminator-based sequencing. (For example, Illumina's Genome Analyzer; Genome Analyzer II; HISEQ 2000; HISEQ 2) Includes 500 (Illumina, San Diego, California).

[0228] In some embodiments, whole-genome methylation sequencing is used to sequence a portion of the genome. Used to determine. In some embodiments, a portion of the genome is the genome (e.g., At least 10 percent, 20 percent, 30 percent, 40 percent of the human reference genome. Percent, 50 percent, 60 percent, 70 percent, 80 percent, 90 percent It is 100%, 95%, 99%, 99.9%, or all. In some embodiments, whole-genome methylation sequencing generates multiple sequence reads. Each sequence read in the multiple sequence reads has a sequence length of 1000 base pairs or less. In some embodiments, whole-genome methylation sequencing is performed across a portion of the genome. At least 5 times, at least 10 times, at least 15 times, at least 20 times, at least 25 times, at least 30 times, at least 50 times, at least 100 times, or at least Obtain sequencing coverage of a portion of the genome that is 200 times greater. Several implementations In this state, whole-genome methylation sequencing is at least 5 times more efficient across the entire genome. At least 10 times, at least 15 times, at least 20 times, at least 25 times, at least 30 times, at least 50 times, at least 100 times, or at least 200 times the sequence Obtain single coverage.

[0229] In some embodiments, methylation sequencing is performed using multiple nucleic acid probes. This is sequencing, and each of the multiple sequence groups (e.g., the target genomic region) is a multiple sequence. It is associated with at least one nucleic acid probe from a number of nucleic acid probes.

[0230] In some embodiments, target sequencing is performed using multiple nucleic acid probes. Target sequencing targets a portion of the genome (for example, the human reference genome). At least 5 times, at least 10 times, less than the portion (for example, the area the probe maps) At least 15 times, at least 20 times, at least 25 times, at least 30 times, at least 5 0x, at least 100x, at least 250x, at least 500x, or at least This also achieves 1000 times the sequencing coverage. In some embodiments, the target sequence The scanning is performed across selected regions within the target genome, increasing the resolution by at least 100 times, and less At least 200 times, at least 500 times, at least 1,000 times, at least 2,00 0x, at least 3,000x, at least 4,000x, at least 5,000x, small At least 10,000 times, at least 15,000 times, at least 20,000 times, at least At least 25,000 times, at least 30,000 times, at least 40,000 times, or Achieve sequencing coverage of at least 50,000 times.

[0231] In some embodiments, targeted panel sequencing is performed, for example, whole-genome sequencing. More efficient than (for example, the use of materials for sequencing, the time required for sequencing) (Regarding the length of time, etc.) and yet significant regarding the target region in the reference genome. It is beneficial because it obtains information. In other words, in some embodiments, the target panel The task of screening is to determine the tumor percentage (and / or tumor origin) for a subject. While making the problem easier to handle computationally, from basic data (for example, cell-free nucleic acid level and It helps to obtain as much information as possible (both across genomic regions). For example, The reference genome (for example, the human reference genome) contains approximately 28 million CpG sites, but the reference genome Targeted methylation panels are designed to have fewer CpG sites (e.g., 10,000-50 Includes CpG sites with 0,000 CpGs and CpG sites with 100,000 to 3,000,000 CpGs.

[0232] In some embodiments, at least one of the multiple probes is a small It is designed to bind to and concentrate nucleic acids in a biological sample containing at least one predetermined CpG site. It is measured. In some implementations, each probe of multiple probes is at least 1 Designed to bind to and concentrate nucleic acids in biological samples containing one predetermined CpG site. ru.

[0233] In some embodiments, each probe of a plurality of probes touches a certain number of predetermined CpG sites. It is designed to target nucleic acids. For example, in some embodiments, multiple One or more probes in the probe have a predetermined CpG site of 50 or less, and 40 or less The designated CpG sites, the designated CpG sites of 30 or less, the designated CpG sites of 25 or less, the sites of 22 or less A specified CpG site, a predetermined CpG site of 20 or less, a predetermined CpG site of 18 or less, a predetermined C of 15 or less pG sites, 12 or fewer predetermined CpG sites, 10 or fewer predetermined CpG sites, 5 or fewer predetermined CpG sites Designed to bind to and concentrate nucleic acids in biological samples containing 3 or fewer predetermined CpG sites. It is being done.

[0234] In some embodiments, for targeted methylation sequencing, multiple probes are used. Includes 1,000 to 2,000,000 probes. In some embodiments, multiple probes Lobe has over 1,000 probes, over 2,000 probes, and over 3,000 probes. Lobes, over 4,000 probes, over 5,000 probes, over 10,000 probes Lobes, including 20,000 or more probes, or 30,000 or more probes. How many? In that embodiment, the number of probes is between 1,000 and 30,000. In some embodiments, the number of probes is at least 5,000, at least 10,0 00, at least 20,000, at least 30,000, at least 40,000, small At least 50,000, at least 100,000, at least 200,000, few 300,000, at least 400,000, at least 500,000, and at least Also 600,000, at least 700,000, at least 800,000, at least Includes 900,000, or at least 1,000,000 probes.

[0235] It should be understood that multiple probes may include any other number of probes, A non-limiting example of this would be less than 1,500,000 probes, or 1,400,000 probes. Below 1,300,000 probes, below 1,200,000 probes, 1,10 Less than 0,000 probes, less than 1,000,000 probes, less than 900,000 probes Below, 800,000 probes or less, 700,000 probes or less, 600,000 probes Below 500,000 probes, below 400,000 probes, below 300,000 probes Below lobe, below 200,000 probes, below 100,000 probes, 90,000 Below 80,000 probes, below 70,000 probes, below 60,000 probes Lobes or less, 50,000 probes or less, 40,000 probes or less, 30,000 probes Below 20,000 probes, below 10,000 probes, below 9,000 probes Below, 8,000 probes or less, 7,000 probes or less, 6,000 probes or less, 5 ,000 probes or less, 4,000 probes or less, 4,000 probes or less, 2,000 Examples include probes or fewer, or 1,000 probes or fewer.

[0236] In some embodiments, multiple probes analyze 0.5–50 megabases of the reference genome. Multiple gene targets to be collectively covered (e.g., a portion of the reference genome and / or genes) The target panel is targeted. In some embodiments, multiple genes of multiple probes are targeted. The targets are 5-40 megabases of the reference genome, 10-30 megabases of the reference genome, and the reference genome Nomu's 15-35 megabases, reference genome's 20-30 megabases, reference genome's 25- It collectively covers 35 megabases, or 30-40 megabases of the reference genome.

[0237] In some embodiments, multiple probes constitute a target cancer assay panel. The targeted cancer assay panel is well known in this field, for example, on April 2, 2019. A patent application titled "Methylated Markers and Targeted Methylation Probe Panels" was filed. International Patent Application No. PCT / US2, published as International Publication No. 2019 / 195268. Filed in patent no. 019 / 025358 on September 27, 2019, "Methylated Markers and International Publication No. 2020 / 069350, titled "Targeted Methylation Probe Panel" International patent application number PCT / US2019 / 053509, published in 2020. The application filed on the 24th of the month, titled "Detecting Cancer, Cancer Tissue or Origin, or Cancer Type International patent application number PC, published as International Publication No. 2020 / 154682, titled " T / US2020 / 015082 (These are referred to herein by reference in their entirety.) It is described as being incorporated. For example, in some embodiments, the target is The assay panel provides information related to determining tumor prevalence and / or diagnosing cancer. Multiple probes capable of capturing fragments (cell-free nucleic acids) that can be provided. (or probe pair) in some embodiments. Multiple probes, at least 50, 100, 500, 1,000, 2,000, 2,5 00, 5,000, 6,000, 7,500, 10,000, 15,000, 20,000 Includes 0, 25,000, or 50,000 pairs of probes. In other embodiments, the target is Multiple probes in the assay panel include at least 500, 1,000, and 2,000 probes. 0, 5,000, 10,000, 12,000, 15,000, 20,000, 30,0 Includes 00, 40,000, 50,000, or 100,000 probes. In this embodiment, multiple probes collectively number at least 100,000, 200,000, 400,000, and 6 0 million, 800,000, 1 million, 2 million, 3 million, 4 million, 5 million, 6 million, 700 Containing 10,000, 8,000,000, 9,000,000, or 10,000,000 nucleotides. Several embodiments So, the probe (or probe pair) differs in cancer and non-cancer samples. It is specifically designed to target one or more chilled genomic regions.

[0238] For example, multiple probes in a targeted cancer assay panel are used in cancerous samples. A probe that can selectively bind and enrich cfDNA fragments that are methylated differently. It may contain . In this case, the tumor proportion can be determined by sequencing of the concentrated fragments. This may provide information related to the diagnosis of cancer. Furthermore, the probe can detect abnormal methylation patterns. Genomes determined to have a pattern of hypermethylation or hypomethylation By designing the detection to target specific regions, further selectivity and specificity can be provided.

[0239] In some embodiments, the number of probes (or probe pairs) in a plurality of probes is small. At least 25bp, 30bp, 35bp, 40bp, 45bp, 50bp, 60bp, 70 Target genomic regions containing bp, 80bp, or 90bp. In some embodiments, The probes in the multiple probes identify a genomic region containing at least five methylation sites. Target. In some embodiments, the probes in a plurality of probes are 20, 15, 1 Target genomic regions containing methylation sites less than 0, 8, or 6. Several implementations In this state, the probes in multiple probes are methylated in non-cancerous or cancerous samples. At least 80, 85 methylation sites (e.g., CpG) that are either methylated or demethylated. The target is a genomic region containing 90%, 92%, 95%, or 98% of the target.

[0240] Filtering of cell-free fragments

[0241] In some embodiments, this method involves filling multiple cell-free fragments with one or more fills. This further includes applying the condition. Thus, in some embodiments, one or All cell-free fragments obtained from methylation sequencing of multiple nucleic acid samples are Used to identify multiple features for estimating the proportion of target cell sources, and / or It is not used to estimate the proportion of the target cell source. In some embodiments This is because nucleic acid fragments (e.g., cell-free nucleic acids) differ in terms of the amount of information they contain. In some embodiments, only nucleic acid fragments having the desired amount of information are used for feature identification and / or or are retained for the estimation of cell source proportions (for example, fragments that do not provide relevant information are discarded). (to be). In some embodiments, the feature is that one of the multiple filtering conditions is This is determined from cell-free fragments that satisfy multiple filter conditions (for example, each filter The condition evaluates the amount of information in the fragments. Multiple filtering methods exist, for example, in May 2020. The application filed on the 22nd of the month, "Systems and Methods for Determining Whether a Subject has International patent application number PCT / US entitled "A Cancer Condition Using Transfer Learning" Filed under number 2020 / 034317 on March 13, 2019, and currently under US2019 The document published as issue / 0287652, titled "Anomalous fragment detection and classification," is available. This is described in detail in U.S. Patent Application No. 16 / 352,602, titled "ication" ( These are incorporated herein by reference, respectively. Typical examples are provided below.

[0242] P-value filtering based on methylation vectors

[0243] In some embodiments, the filter conditions in a plurality of filter conditions are a plurality of The requirement is that each cell-free fragment has a corresponding p-value that is below a threshold. The p-value is as follows: Titled "Learning Whether a Subject Has a Cancer Condition Using Transfer Learning" Example 5 of International Patent Application No. PCT / US2020 / 034317, and 2019 The application was filed on March 13 and is currently published as US2019 / 0287652, U.S. Patent Application No. 16 / 3 entitled "Anomalous fragment detection and classification" Nos. 52,602 (these are incorporated into this specification in their entirety by reference, respectively) It is determined by p-value filtering as described in ( ). The purpose of the Luther condition is to identify abnormal methylated self-methylation based on the corresponding methylation state vector. The goal is to accept and use the free fragments. For example, each cell-free fragment in the sample And, using the methylation state vector corresponding to the fragment, the fragment is (for example, obtained from there) (Through analysis of the sequence reads) abnormal methylation relative to the expected methylation state vector is observed. A determination is made as to whether or not it is happening (for example, here the expected methylation state The cell type is determined from sequence analysis of a cohort of healthy subjects. The generation of methylation state vectors of free fragments is, for example, in U.S. Patent Application Publication 2019 / 0 This is disclosed in Patent No. 287652, which is incorporated herein by reference in its entirety. It shall be assumed that...

[0244] In some embodiments, a healthy cohort includes at least 20 subjects, and multiple clusters. The rufree fragment contains at least 10,000 different corresponding methylation patterns. In some embodiments, the healthy cohort consists of at least 10, at least 20, and at least 30, at least 40, at least 50, at least 60, at least 70, at least Includes 80, at least 90, or at least 100 objects. In some embodiments, The healthy cohorts were divided into 1-10, 10-50, 50-100, 100-500, and 500-1. Includes 000 or more than 1000 targets. In some embodiments, multiple cell-free sections. The options are 1-1000, 1000-2000, 2000-4000, 4000-6000, 6000~8000, 8000~10,000, 10,000~20,000, 20,0 Includes different corresponding methylation patterns ranging from 00 to 50,000, or greater than 50,000.

[0245] In some embodiments, the p-value threshold is 0.001 to 0.20. The threshold is 0.01 (for example, in such an embodiment, p must be < 0.01). (Must be). In some embodiments, the threshold is 0.001, 0.005, 0.01, The threshold is 0.015, 0.02, 0.05, or 0.10. In some embodiments, the threshold is 0.015, 0.02, 0.05, or 0.10. The value is 0.0001 to 0.20. In some embodiments, the p-value threshold is set by multiple self Among the Lee fragments, the corresponding methylation pattern of each cell-free fragment is 0.10 or less, 0.05 If the p-value is below or less than 0.01, then the methylation pattern from the target is as follows: I feel fulfilled.

[0246] In such embodiments, only self-free fragments with p-values ​​below a threshold are used for feature identification. and / or contribute to the estimation of the cell source proportion. For example, in some embodiments, multiple Cell-free fragments are obtained from the plurality of cell-free fragments, with each fragment containing a corresponding plurality of CpG sites. The corresponding methylation patterns (e.g., methylation state vectors) across the range satisfy the p-value threshold. The data is filtered by removing each cell-free segment that has a p-value that does not add up.

[0247] In some embodiments, the abnormal fragment exceeds the threshold number of CpG sites, and methylation of the CpG sites occurs. Fragments exceeding the threshold rate (high methylation), or exceeding the threshold rate of demethylation of the CpG site ( It is identified as a (lowly methylated) fragment. For example, the minimum CpG site and / or fragment length described below. Refer to the filter conditions based on the following. In some embodiments, methylation and / or The threshold rate for non-methylated CpG sites is at least 50%, at least 60%, at least 70%, at least 80%, at least 85%, at least 90%, or at least 9 It is 5%. In some embodiments, the threshold of methylated and / or unmethylated CpG sites The percentage range is 50% to 100%.

[0248] In some embodiments, a Markov model (e.g., a Hidden Markov Model "HMM") is used Use this to give a set of probabilities for the likelihood of observing the following states in the array, each For each state of the methylation pattern of a cell-free fragment, the sequence of methylation states (for example, Each cell-free fragment is categorized as follows: (including "M" for tilization and "U" for demethylation) The probability of observing this is determined. In some embodiments, the set of probabilities is traced to the HMM. This can be achieved through training. Such training is conducted in a cohort of non-cancer subjects. The initial training of the observed methylation state sequence (e.g., methylation pattern) obtained from Given a dataset, what are the statistical parameters (e.g., the probability of transitioning from the first state to the second state)? (Transition probability) and / or the probability that a given methylation state is observed for each CpG site ( This includes calculating the output probability. In some embodiments, the HMM is supervised Training (for example, using samples whose underlying sequence is known, similar to the observed state) It is trained using (the model). In some alternative embodiments, the HMM is trained using unsupervised training. Training (e.g., Viterbi learning, maximum likelihood estimation, expected value maximization training, and / or It is trained using Baum-Welch training. For example, Baum-Welch training. Expectation maximization algorithms such as the Jerch algorithm are derived from the observed sample sequence. Estimate the transition probability and output probability, and parameterize the sequence that best explains the observed sequence. A probabilistic model is generated. In such an algorithm, the expected number of correctly predicted states is... The likelihood function is calculated repeatedly until it is maximized. For example, Yoon, 2009, "Hidden Markov Model" s and their Applications in Biological Sequence Analysis," Curr.Genomics.Sep; 10 See (6):402-415, doi: 10.2174 / 138920209789177575.

[0249] Smallest bag size

[0250] In some embodiments, the filter conditions in the multiple filter conditions are each cell The requirement is that the free fragment has a bag size greater than an integer threshold. In other words, In some embodiments, the filter conditions in one or more filter conditions This means that each cell-free fragment in a group of cell-free fragments corresponds to each fragment in the biological sample. In the corresponding sequence reads measured from one or more nucleic acid samples, The requirement is that it be represented by a threshold number of array reads. For example, threshold If the integer is 1, the filter condition is that each cell-free fragment is measured from the biological sample. In the corresponding sequence reads, the requirement is that they are represented by more than one sequence read. The method involves applying the following. In some embodiments, the threshold integer is 1, 2, 3, 4, 5 , 6, 7, 8, 9, 10, or an integer from 10 to 100. In some embodiments, The threshold integers are 1-10, 10-20, 20-30, 30-40, 40-50, and 50-6. These are 0, 60-70, 70-80, 80-90, or 90-100. Several implementations Morphologically, the threshold integers are 100-500, 500-1000, or greater than 1000. .

[0251] In some embodiments, the filter conditions in the multiple filter conditions are each cell The requirement is that the free fragment has a bag size greater than an integer threshold, and here Then, the array reads within each bag (for example, representing each cell-free fragment) are multiple cell-free This is obtained from nucleic acid sequencing. For example, in some embodiments, one Alternatively, the filter conditions in multiple filter conditions refer to each of the multiple cell-free fragments. A free fragment contains one or more nucleic acid samples, each containing a corresponding fragment in the biological sample. In this case, the requirement is applied that it be represented by a threshold number of cell-free nucleic acids. In some embodiments, the threshold integer is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 , or an integer between 10 and 100. In some embodiments, the threshold integer is between 1 and 10 , 10-20, 20-30, 30-40, 40-50, 50-60, 60-70, 70- The threshold integer is 80, 80-90, or 90-100. In some embodiments, the threshold integer is The ranges are 100-500, 500-1000, or over 1000.

[0252] Minimum number of CpG sites

[0253] In some embodiments, the filter conditions in one or more filter conditions and The requirement is that each cell-free fragment of a group of cell-free fragments has a threshold number of CpG sites. The method involves applying the following. In some embodiments, the number of thresholds for CpG sites is at least 1 , 2, 3, 4, 5, 6, 7, 8, 9, or 10 CpG sites. Several embodiments So, the number of threshold values ​​for CpG sites is 1-10, 10-20, 20-30, 30-40, 40- This refers to a CpG site with 50 or more CpG values.

[0254] In some embodiments, the filter conditions in one or more filter conditions and This means that each cell-free fragment has a base pair length less than a threshold number. The following is required. In some embodiments, the threshold number of base pairs is 1000, 2000, 30 The base pairs are 00 or 4000. In some embodiments, the threshold number of base pairs is 1 00, 200, 300, 400, 500, 600, 700, 800, 900, or 10 It is 00 base pairs. In some embodiments, the threshold number of base pairs is 1000, 20 These are consecutive base pairs of 00, 3000, or 4000. In some embodiments, a salt The threshold numbers for the base pair are 100, 200, 300, 400, 500, 600, 700, and 8 in length. These are consecutive base pairs of 00, 900, or 1000.

[0255] In some embodiments, the filter conditions in the multiple filter conditions are each cell The free fragment covers the first threshold number of CpG sites and, in terms of base pairs, is less than the second threshold in length. The requirement is that the first threshold is 1 CpG site and the second threshold is 1000 salts. If it is a base pair, each cell-free fragment covers more than 1 CpG site and 1000 base pairs. It must be less than the length. In some embodiments, each cell-free fragment is a specific Within the fragment length (e.g., second threshold length), at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 1 CpG sites 0, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 It must be covered. In some embodiments, each cell-free fragment is a certain number ( For example, while spanning the CpG site of the first threshold, the lengths are 500, 1000, 2000, It must be a sequence of fewer than 3000 or 4000 base pairs. In other words, In some embodiments, the filter conditions in a plurality of filter conditions are each The cell-free fragment contains at least one cell within a sequence of fewer than 500 nucleotides in the reference genome. CpG sites, at least 2 CpG sites, at least 3 CpG sites, at least 4 C pG site, at least 5 CpG sites, at least 6 CpG sites, at least 7 CpG areas position, at least 8 CpG sites, at least 9 CpG sites, at least 10 CpG sites , at least 11 CpG sites, at least 12 CpG sites, at least 13 CpG areas The site must include at least 14 CpG sites or at least 15 CpG sites. .

[0256] High methylation or low methylation

[0257] In some embodiments, the filter conditions in the multiple filter conditions are each cell The requirement is that the free fragments are highly methylated. In some embodiments, multiple The filtering condition in this context means that each cell-free fragment is demethylated. This is a requirement. In some embodiments, the filter condition is a genomic region (e.g., It depends on the sequence group. For example, hypermethylation associated with one or more cancerous conditions. Several regions of the human genome that have a state, and one or more cancerous conditions are associated with this. Several regions of the human genome that have a hypomethylated state were filed on April 2, 2019. An international conference titled "Methylated Markers and Targeted Methylation Probe Panels" was held. International Patent Application No. PCT / US201, published as Publication No. 2019 / 195268 Filed on January 24, 2020, under patent number 9 / 025358, "Detecting Cancer, Cancer International Publication No. 2020 / 154682, titled "Tissue or Origin, or Cancer Type" International patent application number PCT / US2020 / 015082, and 201, which were published as such. The patent application filed on September 27, 2009, for "Methylated Markers and Targeted Methylation Probe Pa International patent application number published as International Publication No. 2020 / 069350, titled "nel" PCT / US2019 / 053509 (These are all referenced in their entirety in this document.) It is disclosed in (which shall be incorporated into the document). Therefore, some implementations of this disclosure Morphologically, one or more sequence groups in multiple genomic regions are each classified as International Public Release No. 2 Publication No. 019 / 195268, International Publication No. 2020 / 154682, and / or International Publication This represents the corresponding genomic region in the region disclosed in Patent Publication No. 2020 / 069350, and multiple The filter conditions in the numerical filter conditions are (a) International Publication No. 2019 / 1952 Publication No. 68, International Publication No. 2020 / 154682, and / or International Publication No. 2020 / 0 As indicated by No. 69350, elevated CpG sites associated with one or more cancerous conditions When selecting a fragment to map to a set of sequences representing a human genome region that has a chilled state This requires selecting highly methylated cell-free nucleic acids, and (b) International Publication No. 2 Publication No. 019 / 195268, International Publication No. 2020 / 154682, and / or International Publication As indicated by Publication No. 2020 / 069350, associated with one or more cancerous conditions Fragments that map to sequences representing human genome regions with a low methylation state at CpG sites. If you choose this option, you will need to select low-methylated cell-free nucleic acids.

[0258] In some embodiments, multiple filter conditions include satisfying a p-value threshold and The rufree fragments are required to be highly methylated. In some embodiments, multiple The filter conditions are that the p-value threshold is met and that the cell-free fragments are demethylated. It is necessary to have multiple filter conditions in each array group. They are different. For example, for one of several sets of sequences, there are multiple filter conditions. However, this requires that the p-value threshold is met and that the cell-free fragment is hypomethylated. On the other hand, for the second set of sequences among the multiple set of sequences, multiple filter conditions apply to the p-value threshold. The value must be satisfied, and the cell-free fragment must be highly methylated.

[0259] Cancer state

[0260] In some embodiments, the filter conditions in the multiple filter conditions are each cell The free fragment must meet the threshold for cancer status (for example, each cancer status of each cell-free fragment). The requirement is that the probability associated with exceeds a predetermined threshold. Morphologically, each cancer state has a different, predefined threshold. For example, Systems and Methods for Using Neural Networks to filed on March 31, 2020 U.S. Patent Application No. 63 / 003,087, entitled Determine a Cancer State (see reference) As described in the entirety of this specification, each genome region To determine the probability of cancer in a region (e.g., a set of sequences), a trained neural network is used. A network (for example, one trained on multiple reference subjects) is used. .

[0261] In some such embodiments, for each of the multiple sequence groups, each sequence group is given For each of the multiple cell-free fragments being slapped, the corresponding training The pre-programmed neural network is based on the methylation pattern of each cell-free fragment. , a predicted value which is the probability that a cell-free fragment is associated with a cancerous state (e.g., the presence of cancer). The methyl of each cell-free fragment is calculated. Therefore, in some such embodiments, the methyl of each cell-free fragment is calculated. The pattern is scored using a trained neural network. Here, the score output by the trained neural network A includes the probability that the cell-free fragment has a cancerous state and / or that the cell-free fragment is cancerous. This includes calculations based on probabilities related to the state of each self (e.g., the presence of cancer). A Lee fragment is defined as a score that exceeds the conditions defined above (for example, a fixed threshold value). If the rate is satisfied, the filter condition is passed (for example, special for estimating the cell source ratio). Selected for use in identifying characteristics, and / or for estimating cell source proportions. (Selected for this purpose). Each cell-free fragment is selected based on the score obtained, which meets the conditions defined above (e.g., For example, if the probability of falling below a fixed threshold is not met, the filter condition is not met (e.g. (For example, it will be discarded.)

[0262] In some such embodiments, the threshold is positive or negative. The thresholds are 0.1-1, 1-5, 5-10, 10-50, 50-100, or 100. It is greater than. In some embodiments, the thresholds are -0.1 to -1, -1 to -5, and -5 to -1 0, -10 to -50, -50 to -100, or less than -100. Several implementation forms In this state, the threshold is zero. In some embodiments, each sequence group corresponds to each cancer state. Each has a threshold (for example, each subset of the sequence group is associated with each cancer state). .

[0263] In some embodiments, any combination of the disclosed filter conditions is imposed. In several embodiments, the methylation patterns of the cell-free fragments are disclosed herein. It contains one or more cell-free fragments that satisfy one or more filter conditions. ru.

[0264] Mapping of fragments and sequences

[0265] Block 210. In Block 210, the method is performed on each of the multiple cell-free fragments. Each cell-free fragment is mapped to one of several sequence sets, thereby self This process proceeds by obtaining multiple training sets of Lie fragments. Each row represents a corresponding region of the human reference genome. Each training set of cell-free fragments This is mapped to a different set of sequences among several sets of sequences.

[0266] In some embodiments, the mapping is implemented in Smith-Water, for example, as implemented in Arioc. man gap alignment, or Burrows-Wheeler transformation as implemented in Bowtie, for example. It is performed using [this method]. Other suitable alignment programs include BarraCUDA, BBMap, and BF This includes, but is limited to, AST, BigBWA, BLASTN, BLAT, BWA, BWA-PSSM, and CASHX. That's not the case. For example, Langmead and Salzberg, 2012, Nat Methods 9, pp.357-359; Li and Durbin, 2009, “Fast and accurate short read alignment with Burrows-Wheeler "transform," Bioinformatics 25(14), 1754-1760; and Smith and Yun, 2017, "Evalua tion alignment and variant-calling software for mutation identification in C. el egans by whole-genome sequencing," PLOS ONE, doi.org / 10.1371 / journal.pone.01744 See section 46 (these will be incorporated herein by reference). In some embodiments, each cell-free fragment is mapped to a set of sequences within a set of sequences. This can lead to mismatches. In some embodiments, the mapping is at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, less at least 7, at least 8, at least 9, at least 10, or more than 10 mismatches Includes Chi.

[0267] In some embodiments, referring to block 212, the multiple array groups are 1000~ It consists of or includes a group of 100,000 sequences. In some embodiments, multiple The set of number sequences consists of or includes sequences ranging from 15,000 to 80,000. In some embodiments, the multiple sequence groups consist of 25,000 to 65,000 sequence groups. or including them. In some embodiments, the number of sequence groups is 45,000~ It consists of or includes 65,000 sequences.

[0268] In some embodiments, the number of sequence groups is at least 1000 sequence groups, at least 2 500 sequence groups, at least 5000 sequence groups, at least 10,000 sequence groups, at 20,000 sequence groups, at least 30,000 sequence groups, at least 40,000 sequence groups, At least 50,000 sequence groups, at least 60,000 sequence groups, at least 70,000 0 sequence group, at least 80,000 sequence group, at least 90,000 sequence group, at It includes 100,000 sequence groups or at least 110,000 sequence groups.

[0269] Furthermore, in some embodiments, according to block 214 in Figure 2A, multiple array groups Each sequence group has an average of 10 to 1200 residues (for example, each sequence group has 10 to 12 (corresponding to a portion of the human reference genome consisting of 00 nucleotides). In some embodiments Each of the multiple sequence groups has, on average, 10 to 10,000 residues. In that embodiment, each of the multiple sequence groups has, on average, 10 to 500 residues. In some embodiments, each of the multiple sequence groups consists of an average of 10 to 100 residues. It has. In some embodiments, each of the multiple sequence groups has an average of 25 to 10 It has 0 residues. In some embodiments, each of the multiple sequence groups has, on average, 5 It has 000 to 10000 residues.

[0270] In some embodiments, each sequence group of multiple sequences has fewer than 10 residues and fewer than 20 residues. residues of less than 30, residues of less than 40, residues of less than 50, residues of less than 60, residues of less than 70 residues, residues less than 80, residues less than 90, residues less than 100, residues less than 200, 30 Less than 0 residues, less than 400 residues, less than 500 residues, less than 600 residues, less than 700 residues residues, residues less than 800, residues less than 900, residues less than 1000, residues less than 2000 Less than 3000 residues, less than 4000 residues, less than 5000 residues, less than 6000 residues , including less than 7000 residues, less than 8000 residues, or less than 9000 residues.

[0271] Referring to block 216, in some embodiments, each of the multiple sequence groups is: 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, Includes 18, 19, 20, or more CpG sites. In some embodiments, multiple Each sequence group is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 , including 15, 16, 17, 18, 19, 20 or more consecutive CpG sites. In several embodiments, each of the multiple sequence groups is a sequence of 2 to 100 consecutive sequences in the human reference genome. It consists of CpG sites. In some embodiments, each of the multiple sequence groups comprises 2 to 50 It consists of consecutive CpG sites. In some embodiments, each sequence group of a plurality of sequence groups is 5 It consists of 0 to 100 consecutive CpG sites. In some embodiments, each of the multiple sequence groups The column group consists of at least two consecutive CpG sites.

[0272] In some embodiments, multiple sequences are used in relation to a reference genome (e.g., mammalian, human, etc.). It is constructed by dividing all or part of it into groups of arrays of equal size, and here Each sequence group represents an intrinsic, equally sized portion of the reference genome. In some embodiments, Multiple sequence sets are equivalent to all or part of a reference genome (e.g., mammal, human, etc.). Alternatively, it can be constructed by dividing it into a group of arrays of unequal sizes, where each array Each group represents a specific portion of the reference genome.

[0273] In some embodiments, multiple sequences are used in relation to a reference genome (e.g., mammalian, human, etc.). Constructed by dividing all or part of into sets of arrays of equal or unequal size. Here, each sequence group represents a corresponding portion of the reference genome. In such embodiments, The corresponding portion of the reference genome represented by one set of sequences among multiple sets of sequences is multiple The sequence group may overlap with the corresponding region of the reference genome represented by another set of sequences within that set. In such embodiments, the set of sequences is a reference genome (e.g., mammalian, human, etc.). It is constructed by dividing all of ) into sets of arrays of equal or unequal size, Here, each sequence group represents the corresponding overlapping or non-overlapping portion of the reference genome. In some embodiments, multiple sequence groups are part of a reference genome (e.g., mammals, hi). It is constructed by dividing a (such as a) into groups of arrays of equal or unequal size, In this context, each sequence group represents either an overlapping or non-overlapping portion of the reference genome.

[0274] In some embodiments, a group of sequences is involved in the absence or presence of cancer. While at least a portion of the NM region is represented by its multiple sequence sets, other parts of the reference genome The region is constructed so that it is not represented by its set of sequences. Regardless of the approach, each The sequence group represents a specific portion of the reference genome. In some embodiments, the size of such sequence group This refers to the reference genome in the ranges of 30 bps to 5000 bps, 30 bps to 4000 bps, and 30 bps. s~3000bps, 30bps~2000bps, 30bps~1000bps, also The range is 40 bps to 800 bps. In an alternative embodiment, a large number of such sequence groups The size ranges from 10,000 bps to 100,000 bps and 20,000 bps of the reference genome. ~300,000bps, 30,000bps~500,000bps, 40,000b ps~1,000,000bps, 50,000bps~5,000,000bps, The range is 100,000 to 25,000,000 bps.

[0275] In some embodiments, the portion of the reference genome is chromosomes 1-22 of the reference genome, or At least 25 percent of the reference genome, at least 30 percent, at least 35 percent - cents, at least 40 percent, at least 45 percent, at least 50 percent cents, at least 55 percent, at least 60 percent, at least 65 percent , at least 70 percent, at least 75 percent, at least 80 percent T, at least 85 percent, at least 90 percent, at least 95 percent Or at least 99 percent. In some such embodiments, each sequence group This refers to the 10,000 to 100,000 base pairs and the 20,000 to 300 base pairs of the reference genome. 000 bases, 30,000 bases to 500,000 bases, 40,000 bases to 1,000, 000 bases, 50,000 bases to 5,000,000 bases, or 100,000 bases to This represents 25,000,000 base pairs.

[0276] In some embodiments, each sequence group is a reference genome identified as being associated with cancer. It represents a specific part of something.

[0277] In some embodiments, the sequence group was compared to a non-cancer control in cfDNA. Cancer and / or tissue-specific methylation patterns identified as being associated with cancer This represents a specific region of the irradiated genome.

[0278] In some embodiments, each sequence group is an enhancer, promoter in the reference genome. -, 5'UTR, exon, exon / inhibitor boundary, intron, intron / exon boundary , representing all or part of a 3'UTR region, CpG shelf, CpG shore, or CpG island. For a proper definition of such a domain, see, for example, Cavalcante and Santor, 2017, “ann See "otatr: genomic regions in context," Bioinformatics 33(15) 2381-2383. Here, such annotations have been documented for several different species. ru.

[0279] In some embodiments, genomic regions that are highly variable or have low mapping potential are For example, using the method disclosed in Jensen et al, 2013, PLoS One 8; e57381, multiple It will be excluded from the presentation of the sequence group within the sequence group. Furthermore, regarding the analysis of mapping possibilities... See Li and Freudenberg, 2014, Front.Genet.5, p.318.

[0280] Selection of human genome regions to be used in the sequence set

[0281] In some embodiments of this disclosure, each of the sequence groups is a cancer-specific methylated sequence. Derived from a panel of genomic regions designed for turn target selection. Several embodiments So, each of these genomic regions is referred to in the "Detecting Canc" patent application filed on January 24, 2020. International publication No. 2020 / 154 titled "Cancer Tissue or Origin, or Cancer Type" International patent application number PCT / US2020 / 015082, published as No. 682 ( (The sequence list mentioned in this document, including the sequence list mentioned in this document, shall be incorporated herein by reference.) This is derived from Table 2. SEQ ID NO 45 of PCT / US2020 / 015082. 2,706–483,478 are specific hypermethylated or hypomethylated target genomic regions. This provides further information regarding these SEQ ID NO records, which help in identifying cancer types. By combining these factors, we can identify target genomic regions that may be methylated differently in the sample. SEQ ID NO 452, 706-4 of PCT / US2020 / 015082. Target genomic regions 83 and 478 are listed in PCT / US2020 / 015082, specifically in list 6. This leads to the conclusion that many of the same target genome regions are related to PCT / US2020 / 015082. It can also be seen in St 1-5 and List 7-16. Entries for each SEQ ID are: The chromosomal location of the target genomic region for hg19, and the cfDNA fragments enriched from that region. The sequence of a single DNA strand in the target genomic region, whether it is hypermethylated or hypomethylated, and A combination or multiple cancer types that are methylated differently in their genomic regions. The combinations are shown. The methylation status of several target genome regions can be determined by multiple combinations. To distinguish between species, each entry is specified in PCT / US2020 / 015082 (in which it is stated). Table 3 of the sequence listings (including those provided) shows the first cancer type and one or more second cancers. Identify the species.

[0282] In some embodiments, the sequence group of the present disclosure includes PCT / US2020 / 015 Lists 1-16, 1-3, 13-16, 12, and 4 of issue 082, Or at least 200, 500, 1,000 in any one of Lists 8-11 5,000, 10,000, 15,000, 20,000, 30,000, 40,000 or include separate sets of sequences for each of the 50,000 target genomic regions. In several embodiments, the plurality of sequences of this disclosure include PCT / US2020 / 015082 Any combination of one or more items from Lists 1-16 (e.g., Lists 1-3, List 1) At least two of the following (3-16, List 12, List 4, or List 8-11, etc.) 00, 500, 1,000, 5,000, 10,000, 15,000, 20,000, For each of the 30,000, 40,000, or 50,000 target genomic regions It contains separate sets of sequences.

[0283] In some embodiments, the sequence group of the present disclosure includes PCT / US2020 / 015 At least 20% of the target genomic region in any one of the listings 1-16 in issue 082, To each of 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 95% This includes a separate set of sequences. In some embodiments, the set of sequences of this disclosure includes: List 1-16 of PCT / US2020 / 015082 (for example, (List 1-3, List 13-16, List 12, List 4, or List 8-11, etc.) At least 20%, 30%, 40%, or 50% of the target genomic region in any combination. This includes separate sets of sequences for 60%, 70%, 80%, 90%, or 95% respectively. It can be done.

[0284] Additional selection of human genome regions to be used in the sequence set

[0285] In some embodiments of this disclosure, each of the sequence groups is cancer-specific methylation. Derived from a panel of genomic regions designed for pattern target selection. In the application, each such genomic region is referred to in the "Methylat" patent application filed on September 27, 2019. International Publication No. 2020 / 0 entitled "ed Markers and Targeted Methylation Probe Panel" International patent application number PCT / US2019 / 053509, published as patent number 69350. Derived from Table 2, which includes the sequence list mentioned therein, by reference to this specification. It shall be incorporated into.

[0286] The sequence listing in International Publication No. 2020 / 069350 contains the following information: (1) SEQ ID NO, (2)(a) chromosome or contig where the CpG site is located, (b) opening of the region Array identifiers that specify the start and end positions, arrays corresponding to (3)(2), and ( 4) Whether the region is included based on its high or low methylation score. It is found. The chromosome number, start position, and stop position are based on the known human reference genome GRCh37 / h Provided based on gl9. The GRCh37 / hgl9 sequence was developed by the Santa Cruz Genomics Institute. Provided by the National Center for Biotechnology Information (NCBI), the Genome Reference It is available from the ce Consortium and Genome Browser.

[0287] Generally, the sequence group is a set of marks included in List 1-8 of International Publication No. 2020 / 069350. It may include any of the CpG sites that fall within the start / stop range of any of the target regions.

[0288] In some embodiments, the sequence group of the present disclosure includes International Publication No. 2020 / 0693 At least 200, 500, 1,000, 5, in any of the listings 1-8 of No. 50. 000, 10,000, 15,000, 20,000, 30,000, 40,000, This includes separate sets of sequences for each of the 50,000 target genomic regions. In that embodiment, the plurality of sequences of the present disclosure include the sequence of International Publication No. 2020 / 069350 At least 200, 500, 1,000, 5 in any combination of List 1-8 000, 10,000, 15,000, 20,000, 30,000, 40,000, This includes separate sets of sequences for each of the 50,000 target genomic regions.

[0289] In some embodiments, the sequence group of the present disclosure includes International Publication No. 2020 / 0693 At least 20% of the target genomic region in any one of the list 1-8 in issue 50, 30 For each of %, 40%, 50%, 60%, 70%, 80%, 90%, or 95% This includes a separate set of sequences. In some embodiments, the set of sequences of this disclosure includes international Any combination of target genome regions from List 1-8 in Publication No. 2020 / 069350 At least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 9% Each of the 5% includes a separate set of sequences.

[0290] In some embodiments of this disclosure, each of the sequence groups is a cancer-specific methylated sequence. Derived from a panel of genomic regions designed for turn target selection. Several implementations Morphologically, each such sequence group is a "Methylated Marke" filed on April 2, 2019. International Publication No. 2019 / 19526, titled "rs and Targeted Methylation Probe Panels" International patent application number PCT / US2019 / 025358, published as application number 8 (the entire application) The genome in any of Tables 1-24 (which are incorporated into this book by reference) Corresponds to the domain.

[0291] In some embodiments, each of the sequences in this disclosure is related to International Publication No. 2019 / 195268. Tables 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 , noted in one or more of 17, 18, 19, 20, 21, 22, 23 and / or 24 It is mapped to the genomic region.

[0292] In some embodiments, the entirety of the multiple sequences of this disclosure together constitutes International Publication No. 2 At least one of the genomic regions in Table 1-24 of issue 019 / 195268 Mapped to 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 95% It is configured to be configured such that the sequence in a group of sequences is configured in some such embodiments. The group is a single unique entity in any of the Tables 1-24 of International Publication No. 2019 / 195268. It is mapped to the corresponding genomic region. In some such embodiments, the present disclosure The sequence groups in the multiple sequence groups are shown in Table 1-24 of International Publication No. 2019 / 195268. Unique correspondences of 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 in any combination It is mapped to the corresponding genomic region.

[0293] In some such embodiments, the sequence group in the plurality of sequence groups of the present disclosure is internationally public. A single in either Table 2-10 or 16-24 of Publication No. 2019 / 195268 It is mapped to a specific corresponding genomic region. In some such embodiments, multiple The sequence groups in the sequence groups of numbers are shown in Table 2-10 of International Publication No. 2019 / 195268 or 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 in any combination of 16-24 It is mapped to a specific corresponding genomic region.

[0294] In some embodiments, one or more sequence groups in the plurality of sequence groups of this disclosure are Together, Tables 1, 2, 3, 4, 5, 6, and 7 of International Publication No. 2019 / 195268, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, At least 30%, 40%, and 50% of the genomic region in 22, 23, and / or 24 It is configured to map to 60%, 70%, 80%, 90%, or 95%.

[0295] Assignment of cell-free fragments to cancer status

[0296] Block 218. Referring to Block 218 in Figure 2B, this method is used for cell-free fragments. Each self within each training set of the self-free fragments of the multiple training sets This proceeds by assigning the cancerous state of the self-free fragment to the self-free fragment, and here the self-free fragment - The cancerous state of the fragments is determined when the methylation pattern of each cell-free fragment is input into the classification device. It is a function of the output of the classification device, representing either the first cancer state or the second cancer state.

[0297] In some embodiments, the classification device has the following form:

number

[0298] In some such embodiments, JPEG2026048644000008.jpg1096 is the first model of the first stage of cancer.

[0299] In some such embodiments, JPEG2026048644000009.jpg1096 is a second model of the second cancer state. In some embodiments, the first and second models Regarding Dell, "fragment" refers to the methylation pattern of each cell-free fragment. In the application method, the cancerous state of the cell-free fragments of each fragment is determined when R(fragment) satisfies the threshold. A first cancer state is assigned to each case. In some embodiments, the threshold is 1 to 10. It is an arbitrary value. In some embodiments, the threshold is 1, 2, 3, 4, 5, 6, 7, 8, It is 9 or 10.

[0300] In some embodiments, the first model is a first mixture comprising a first plurality of submodels The first model is a second hybrid model that includes a second set of submodels, and the second model is a second hybrid model that includes a second set of submodels. Each of the submodels of the first and second submodels is self-reactive in the corresponding biological sample. —Represents an independent, corresponding methylation model for the source of the fragment.

[0301] In some embodiments, the cancer state in question is one of several cancer states. (For example, if multiple cancer conditions include N cancer conditions). Some such actual In terms of implementation, the classification device has the following form:

number

[0302] In some such embodiments, JPEG2026048644000011.jpg1163 is a third model of the third cancer state in a group of cancer states. In one embodiment JPEG2026048644000012.jpg13102 is the Nth model for the Nth cancer state in a group of cancer states.

[0303] Examples of mixture models for use according to embodiments of this specification are provided on May 13, 2019. A U.S. patent application filed on [date] entitled "Model-Based Featurization and Classification" Application No. 62 / 847,223 (the entirety of which is incorporated herein by reference) It is described in ).

[0304] In some embodiments, each independent corresponding methylation model is a binomial model, a beta-binary model. It is one of the following: a model, an independent-part model, or a Markov model. Several implementations In this state, two or more of the first submodels are independent site models. Furthermore, two or more of the submodels in the second set of submodels are independent site models.

[0305] For example, the "Identifying Methylation Patterns tha" patent application filed on February 28, 2020. U.S. Patent Application No. 62 / 98, titled "Discriminate or Indicate a Cancer Condition" Patent No. 3,443 (which is incorporated herein by reference in its entirety) covers the subject matter. This paper discloses several methods for identifying methylation patterns that distinguish specific cancer states. In some embodiments, each cancer state in a group of cancer states (e.g., origin) Cancers that result in a targeting of the reference genome or a subset of the reference genome (e.g., targeting a specific panel) Each pattern of abnormal methylation (e.g., modified methylation) is evaluated by methyl sequencing. This method corresponds to the (synthesis pattern). To determine the state of cancer in a specific target, this method is intended to achieve the objective. Evaluate multiple genomic regions, and for each genomic region of the multiple genomic regions, Generates corresponding counts of fragments having methylation patterns that map to the nom region. (For example, possible methylation patterns that can be identified in fragments that map to each genomic region) (There is a count for each fragment for each segment). Next, this method calculates the number of segments for the subject. Fragment counts across multiple genomic regions, corresponding to methylation patterns in different cancer states. Compared to a database (e.g., a library) (for example, each cancer state is compared to multiple g (Having a corresponding fragment count for each subset of genomic regions within the genomic region), This determines the cancerous state with almost certainty. Here, the cancerous state is defined as cancer versus non-cancer. This corresponds to a species and / or origin tissue. In some embodiments, this method is downstream For input to applications (for example, to estimate the proportion of tumors in a given area, and / or It is used to identify the state of the cancer in question (or to determine the minimum residual disease). In some embodiments, the group of sequences used in this disclosure is U.S. Patent Application No. 62 / 9 Related to any single or any combination of cancers evaluated in No. 83,443 The methylation pattern identified in U.S. Patent Application No. 62 / 983,443 Selected to represent a part of Nomu.

[0306] Another example is the "Model-Based Featurization and U.S. Patent Application No. 15 / 931,022, entitled "Classification" (see reference for the entire application) (which shall be incorporated herein) methylation features corresponding to distinct cancer conditions To identify the methylation status of a genomic region (e.g., the sequence mapped to the genomic region), We disclose the development of a probabilistic model using (determined from fragments represented by reads). In some embodiments, the group of sequences used in this disclosure is U.S. Patent Application No. 15 / 93 In case 1,022, any single or any combination of cancers were evaluated. Geno, including chilling patterns, identified in U.S. Patent Application No. 15 / 931,022 It is selected to represent the "mu" part.

[0307] Other methods for classifying cancers using nucleic acid fragments include, for example, December 13, 2019. The patent application for "Cancer Classification using Patch Convolutional Neural Networks" was filed. U.S. Patent Application No. 62 / 948,129, filed on March 13, 2019, titled " Method and System for Selecting, Managing, and Analyzing Data of High Dimensiona U.S. Patent Application No. 16 / 352,739, titled "lity", filed on May 31, 2019. "Convolutional Neutral Network Systems and Methods for Data Classification" U.S. Patent Application No. 16 / 428,575, titled " and filed on March 4, 2020 "Systems and Methods for Cancer Condition Determination using Autoencoders" Examples include those disclosed in U.S. Patent Application No. 62 / 985,258, titled ", Each of these shall be incorporated here by reference, with the whole being included.

[0308] In some embodiments, the classification device is a multivariate logistic regression, a neural network, etc. Work, convolutional neural networks, support vector machines (SVMs), decision trees This could be a regression algorithm or a supervised clustering model.

[0309] Logistic regression algorithms, including multivariate logistic regression, are described by Agresti, An In Introduction to Categorical Data Analysis, 1996, Chapter 5, pp.103-144, John Wiley This information is disclosed in & Son, New York, and is incorporated herein by reference.

[0310] Neural network algorithms including convolutional neural network algorithms The term "zum" is found in Vincent et al., 2010, “Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion,” J Mach Lea rn Res 11, pp. 3371-3408; Larochelle et al., 2009, “Exploring strategies for tr aining deep neural networks,” J Mach Learn Res 10, pp. 1-40; and Hassoun, 1995, Fundamentals of Artificial Neural Networks, Massachusetts Institute of Technology These are disclosed in GY, and are incorporated herein by reference, respectively.

[0311] The SVM algorithm is described in Cristianini and Shawe-Taylor, 2000, “An Introduction to Support Vector Machines,” Cambridge University Press, Cambridge; Boser et al., 1992, “A training algorithm for optimal margin classifiers,” in Proceedings o f the 5 th Annual ACM Workshop on Computational Learning Theory, ACM Press, Pitts burgh, Pa., pp. 142-152; Vapnik, 1998, Statistical Learning Theory, Wiley, New Y ork; Mount, 2001, Bioinformatics: sequence and genome analysis, Cold Spring Harb or Laboratory Press, Cold Spring Harbor, NY; Duda, Pattern Classification, Sec. ond Edition, 2001, John Wiley & Sons, Inc., pp. 259, 262-265; and Hastie, 2001, The Elements of Statistical Learning, Springer, New York; and Furey et al., 2000 These are described in Bioinformatics 16, 906-914, and their entirety can be found by reference. The body shall be incorporated herein. When used for classification, SVM shall (for example, A predetermined set of binary-labeled data training sets (based on tumor percentage values) Separate the data using the hyperplane furthest from the data with the markers. If linear separation is not possible, use SV M is combined with a technique called a "kernel" that automatically implements nonlinear mapping to the feature space. It is possible to operate them in conjunction. The hyperplane found by the SVM in the feature space is in the input space. This corresponds to a nonlinear decision boundary.

[0312] Decision trees are described in Duda, 2001, Pattern Classification, John Wiley & Sons, Inc., New York. This is generally explained in k, pp. 395-396, which is incorporated herein by reference. The tree-based method divides the feature space into a set of rectangles, and then each A model (something like a constant) is fitted to this. In some embodiments, the decision tree runs This is dam forest regression. One specific algorithm that can be used is classification. This is a regression tree (CART). Other specific decision tree algorithms include ID3, C4.5, MART, Examples include, but are not limited to, random forests. CART, ID 3, and C4.5 are from Duda, 2001, Pattern Classification, John Wiley & Sons, Inc., Ne This is described in York, pp. 396–408 and pp. 411–412, which is referred to herein by reference. It shall be incorporated. CART, MART, and C4.5 are based on Hastie et al., 2001, The Eleme. This is described in "Studies of Statistical Learning," Springer-Verlag, New York, Chapter 9. The entirety of this will be incorporated into this book by reference. Random forests are Brei man, 1999, “Random Forests--Random Features,” Technical Report 567, Statistics It is documented in the Department, UC Berkeley, September 1999, and the entire document is referenced by This specification shall be incorporated herein.

[0313] Clustering is described in Duda and Hart, Pattern Classification and Scene Analysis, 19 Pages 211-256 of 73, John Wiley & Sons, Inc., New York (hereinafter referred to as "Duda 1973"). It is described in Duda 1973, and the entire text is incorporated here by reference. As described in Section 6.7, the challenges of clustering are that within a dataset It is described as finding natural groupings. Identifying natural groupings. To that end, we will address two problems. First, the similarity (or dissimilarity) between two samples. The measurement method is determined. This index (similarity index) is used to determine if samples within a certain cluster are similar. This is used to ensure that samples are more similar to each other than samples within other clusters. Then, a mechanism is decided upon to divide the data into clusters using a similarity index.

[0314] The measure of similarity is discussed in section 6.7 of Duda 1973, and the clustering method One way to begin the analysis is to define a distance function and apply it to all the samples in the training set. It is stated that this involves calculating a matrix of distances between pairs of elements. Distance is a good measure of similarity. If so, the distance between reference entities within the same cluster is equal to the distance between reference entities within different clusters. It will be significantly smaller than the distance between entities. However, Duda 1973, page 215 As stated on the page, clustering does not require the use of distance measurement. For example To compare two vectors x and x', we use a nonmetric similarity function s(x,x'). It is possible to use it. Traditionally, s(x,x') is a symmetric function, and its value is such that x and x' It becomes larger when there is some kind of "similarity". Example of a non-metric similarity function s(x,x') This is described on page 218 of Duda 1973.

[0315] After selecting a method to measure the "similarity" or "dissimilarity" between points in the dataset, Clustering requires a criterion function to measure the clustering quality of any arbitrary division of data. This is essential. Data clustering is the process of splitting a dataset in a way that maximizes a criterion function. It is used in [this way]. See page 217 of Duda 1973. The reference function is in section 6 of Duda 1973. This is explained in section 8.

[0316] Recently, Duda et al., Pattern Classification, 2 nd edition, John Wiley & Sons, Inc. New York published the book. Pages 537-563 describe the details of clustering. It is being done. More detailed information on clustering technology can be found in Kaufman and Rousseeuw, 1 990, Finding Groups in Data: An Introduction to Cluster Analysis, Wiley, New York k, NY; Everitt, 1993, Cluster analysis (3d ed.), Wiley, New York, NY; and Ba cker, 1995, Computer-Assisted Reasoning in Cluster Analysis, Prentice Hall, Uppe These can be found in Saddle River, New Jersey, and are listed in this document by reference. It shall be incorporated into the document. Specific exemplary clusterings that may be used in this disclosure In terms of clustering techniques, there is hierarchical clustering (nearest neighbor algorithm, furthest neighbor algorithm, hierarchical clustering). Condensation using a homogeneous connected algorithm, centroid algorithm, or sum of squares algorithm (Collective clustering), k-mean clustering, fuzzy k-mean clustering Examples include gorym and Jarvis-Patrick clustering, but are not limited to these. No. Such clustering is based on the set of first features {p1,...,pN-K}( Or it may be for the principal component derived from the set of first features. Several implementations In this case, clustering involves how the training set is clustered. An unteachered class where no predetermined opinion is imposed on whether an eel cluster should be formed. Taring is included.

[0317] Identification of features

[0318] Block 220. Referring to Block 220 in Figure 2B, this method involves each of the multiple sequence groups. Regarding the sequence group, (a) of the multiple training targets, the target of each training target (b) the state of (a) and the training of the corresponding cell-free fragments mapped to each sequence group. To determine the corresponding association Scale I between the cancer status of each cell-free fragment in the set Therefore, proceed.

[0319] In some embodiments, with respect to block 222, the measure of relevance is correlation. Referring to block 224, in some embodiments, the correlation is expressed by the Pearson correlation coefficient. Yes. Referring to block 226, in some embodiments, the correlation is an adjusted correlation coefficient. This is performed using weighted correlation, reflection correlation coefficient, or scaled correlation coefficient.

[0320] In some embodiments, the measure of relevance is the calculation of mutual information. For example, Song et al. al., 2012, “Comparison of co-expression measures: mutual information, correlati See, “on, and model-based indices,” BMC Bioinformatics 13, 328. For example, In some embodiments, mutual information is calculated according to Figure 8. The training target label Y (cancer type A or B in the case of two cancer types) and sequence group features. The mutual information between X is calculated by the mutual information. In fact, Figure 8 shows that the subject is cancer type A or Under the assumption that the probability of having either B is the same (P(Y=A)=P(Y=B)) This provides a method for calculating mutual information in the following specific embodiments. In some particular embodiments, the measure of relevance is The mutual information is calculated as follows:

number

[0321] In some such embodiments, i and j are a set of cancer states (e.g., It is an independent index for the status of primary and secondary cancers. Several implementations In terms of form x i Among multiple training subjects, trainees with cancer status i This is the number of targets (for example, i is the first cancer state, or alternatively, i is the second...) (e.g., having cancer). In some embodiments, y j These are mapped to each set of sequences. Among multiple training subjects having one or more cell-free fragments, the cancer state j This is the number of training subjects assigned (for example, j is the first cancer state, or Alternatively, j could be a second cancerous condition. The measurement will take the following form.

number

[0322] In some such embodiments, the measure of relevance is at least a) the first stage of cancer Each sequence group having a state and assigned to the first cancer state contains one or more self a) Number of training subjects having Lee fragments, b) Having a first cancer state but a second cancer Trainers having one or more cell-free fragments within each sequence group assigned to a state c) Number of subjects to be treated, each of the following: having a second cancerous state and assigned to a second cancerous state The number of training subjects that have one or more cell-free fragments in the sequence group, and d) Although it has a second cancerous state, it has one or more within each sequence group assigned to the first cancerous state. This is determined based on the number of training subjects that have multiple self-free fragments.

[0323] In one embodiment, the function

number

number

[0324] In some embodiments, when two possible cancer states are present, the relevance scale Degrees are a distance metric. Table 1 provides examples of such distance metrics.

[0325] [Table 1] JPEG2026048644000019.jpg124153

[0326] In some embodiments, the calculation of the relevance measure is performed for each of the multiple training subjects. The training target is a group of sequences that have one of several cancer conditions. Then, the measure of relevance is determined. In some such embodiments, the measure of relevance is, It is calculated as follows:

number

[0327] JPEG2026048644000021.jpg81153

[0328] Block 228. This method, referring to Block 228 in Figure 2B, involves determining the proportion of the target cell source. By identifying multiple features for estimation as subsets of multiple sequence sets, the process continues. In a subset of multiple sequence groups, each sequence group has a corresponding correlation scale for each sequence group. It meets the selection criteria based on degree.

[0329] In some embodiments, the selection criterion is one of the top N items on the relevance scale. A selection of arrays is specified, where N is a positive integer greater than or equal to 50. In some embodiments, In some embodiments, N is 500 to 5000. Yes. In some embodiments, N is at least 100, at least 200, and at least 300, at least 400, at least 500, at least 600, at least 700 , at least 800, at least 900, at least 1000, at least 1100, small At least 1200, at least 1300, at least 1400, or at least 150 It is 0.

[0330] In some embodiments, referring to block 230, the selection criterion is a measure of relevance. This specifies the selection of a group of sequences that have one of the top N sequences, where N is a positive integer greater than or equal to 50. The number is such that (for example, at least 50 sequences with the highest relevance measure are characterized) (and is selected).

[0331] In some embodiments, the number of features may be at least 10, at least 50, and at least 100 each, at least 200, at least 300, at least 400, at least 50 0, at least 600, at least 700, at least 800, at least 900, less At least 1000, at least 1100, at least 1200, at least 1300, less The features include at least 1400, or at least 1500, characteristics. In some embodiments, Multiple features include 500-5000, 800-1500, or more than 1500 features. Born.

[0332] Estimation of cell source proportion

[0333] In some embodiments, multiple features (e.g., distribution) are used to estimate the proportion of the target cell source. After identifying a subset of the column group, this method tests the pair based on at least several features. This further includes estimating the proportion of elephant cell sources.

[0334] In some embodiments, the method electronically generates multiple self-free fragments for testing. The corresponding methylation pattern of each cell-free fragment is obtained (for example, cancer classification is desired). The procedure involves performing an estimate of the proportion of cell sources or tumors from the test subjects, In this case, the corresponding methylation pattern of each cell-free fragment is obtained from (i) the test subject. Methylation sequence of one or more nucleic acid samples containing each fragment in a biological sample (ii) Determined by testing, and of the multiple corresponding CpG sites in each fragment, each CpG site This procedure includes the methylation state of each of several cell-free fragments for testing. Mapping to a set of sequences within a set of sequences, thereby enabling multiple tests of self-free fragments. Further including obtaining a set, each test set of cell-free fragments is one of several sequence groups It is mapped to different sets of sequences. This procedure involves first determining the methylation pattern of each cell-free fragment. The cancer state of the cell-free fragment, which is a function of the output of the classification device when input into the classification device, Each cell-free fragment in each test set of multiple test sets of cell-free fragments This is followed by assigning to a free fragment. This procedure is performed on a subset of multiple sequence sets. In each test set of cell-free fragments, the first cancer state was assigned to the test. Calculate the first representative value of the number of cell-free fragments from the target, and subsets of multiple sequence groups. In each test set of self-free fragments across the test area, the self-free fragments from the test subject This procedure includes calculating the second representative value of a number. This procedure uses the first and second representative values. Then, the proportion of cell sources for the test subjects is estimated.

[0335] In some embodiments, the second cancer state includes the absence of cancer and is subject to test. The estimated cell source proportions include the tumor proportions for the test subjects.

[0336] For example, in some embodiments, the tumor percentage is estimated from the biological sample under test ( For example, one or more methylation status patterns in cfDNA and / or plasma are found in tumors. The origin of this tumor is the frequency of such tumor-derived methylation patterns relative to the ratio of cancer cells to normal cells. It is calculated based on the assumption that it is directly proportional to the sum (for example, the tumor percentage).

[0337] There are various methods for determining such percentages, some of which were discussed on December 18, 2019. The patent application "Systems and Methods for Estimating Cell Source Fractions using Methy U.S. Patent Application No. 16 / 719,902, entitled “Lation Information” and 2020 The patent application filed on April 16th, "Systems and Methods for Tumor Fraction Estimation from Sma It is described in U.S. Patent Application No. 16 / 850,634, titled "II Variants," and Both will be incorporated into this book in their entirety by reference.

[0338] In some embodiments, the first representative value is a self across a subset of multiple sequence groups. Multiple test subjects were assigned a primary cancer state in each test set of Lee fragments. Arithmetic mean, weighted mean, midrange, mid-hinge, ternary mean of the number of self-free fragments from This is the mean, Windsor mean, average, or mode. In some embodiments, the second representative The value is the value in each test set of a cell-free fragment spanning a subset of multiple array sets. Arithmetic mean, weighted mean, midrange, and midrange of the number of self-free fragments from multiple test subjects. This is the Dohinge, ternary mean, Windsor mean, mean, or mode. Several implementations In this context, estimating the cell source ratio involves dividing the first representative value by the second representative value. In some embodiments, each of the training targets among multiple training targets The cancer state is selected from multiple cancer states. In some embodiments, The central tendency scale is determined for each cancer state among several cancer states. In such embodiments, estimating the cell source proportion is a first representative value that is a central tendency of other This includes dividing by the sum of each scale.

[0339] In some embodiments, the proportion of tumors under test is 0.003 to 1.0. In some embodiments, the proportion of tumors being tested is in the range of 0.001 to 1.0. In some embodiments, the proportion of tumors under test is at least 0.001, at least 0 0.005, at least 0.01, at least 0.05, at least 0.1, at least 0 0.2, at least 0.3, at least 0.4, at least 0.5, at least 0.6, less It is at least 0.7, at least 0.8, at least 0.9, or at least 1.0.

[0340] In some embodiments, determining the proportion of the target cell source (e.g., tumor) is the target Further identification of the cancer that is the origin of the disease. In some embodiments, the first and / or second The state of cancer includes the tissue of origin (for example, the place where the cancer is thought to have originated). In some embodiments, the first and / or second cancer state includes the cancer stage (e.g.) For example, this includes stages I, II, III, or IV.

[0341] Some conventional cancers are the source of cancer, including non-cancerous cancers, breast cancer, lung cancer, and prostate cancer. Cancer, colorectal cancer, kidney cancer, uterine cancer, pancreatic cancer, esophageal cancer, lymphoma, head / neck cancer, egg Focal cancer, hepatobiliary tract cancer, melanoma, cervical cancer, multiple myeloma, leukemia, thyroid cancer, bladder Selected from the group consisting of gastric cancer, nasopharyngeal cancer, liver cancer, or a combination thereof. This includes the first stage of cancer.

[0342] In some embodiments, the originating cancers include breast cancer, lung cancer, prostate cancer, and colorectal cancer. Hmm, kidney cancer, uterine cancer, pancreatic cancer, esophageal cancer, lymphoma, head / neck cancer, ovarian cancer, hepatobiliary cancer Cholesterol cancer, melanoma, cervical cancer, multiple myeloma, leukemia, thyroid cancer, bladder cancer, stomach cancer, A small number of cases are selected from the group consisting of nasopharyngeal cancer, liver cancer, or a combination thereof. This includes, at the very least, the first and second stages of cancer.

[0343] In some embodiments, the first and / or second cancer state is a stage Breast cancer, lung cancer at a certain stage, prostate cancer at a certain stage, colorectal cancer at a certain stage, A certain stage of kidney cancer, a certain stage of uterine cancer, a certain stage of pancreatic cancer, a certain stage Esophageal cancer of a certain stage, lymphoma of a certain stage, head / neck cancer of a certain stage Ovarian cancer, hepatobiliary tract cancer at a certain stage, melanoma at a certain stage, cervical cancer at a certain stage Hmm, multiple myeloma at a certain stage, leukemia at a certain stage, thyroid cancer at a certain stage, Bladder cancer at a certain stage, stomach cancer at a certain stage, nasopharyngeal cancer at a certain stage This includes liver cancer, or a combination of these.

[0344] In some embodiments, determining the proportion of the cell source to be tested (e.g., tumor) is This further includes providing treatment recommendations (e.g., cancer treatment) to the subjects of testing, and the treatment recommendations are , the proportion of cellular sources (e.g., how far the disease has progressed) and the low prevalence of the originating cancer It is based on both in part.

[0345] In some embodiments, this method is used to monitor the progression of a disease or to assess the therapeutic effect. To monitor (for example, therapeutic efficacy), one or more time points (for example, before treatment) This further includes determining the proportion of the cell source (e.g., tumor) to be tested (or later). For example, in some embodiments, the tumor rate over time (e.g., at a second, later point in time) An increase indicates disease progression, and conversely, in some embodiments, it indicates the progression over time (e.g., second, A reduction in the tumor percentage (at a later point in time) indicates that the treatment was successful.

[0346] For example, in some embodiments, the method is used when the value of the proportion of the cell source under test is small. This further includes applying a treatment regimen to the test subject, at least partially. In some embodiments, the treatment regimen includes applying an anti-cancer drug to the subject of test. In some embodiments, cancer drugs are hormones, immunotherapy, X-rays, or anticancer drugs. It is a drug. In some embodiments, the anticancer drug is lenalidomide, pembrolizumab, Trastuzumab, bevacizumab, rituximab, ibrutinib, human papillomavirus Quadrivalent (types 6, 11, 16, and 18) vaccine, pertuzumab, pemetrexed, niro Tinib, nilotinib, denosumab, abiraterone acetate, promacta, imatinib, evero Limus, palbociclib, erlotinib, bortezomib, or bortezomib, These are generic equivalents.

[0347] In some embodiments, the subjects of test have been treated with cancer drugs, and this method This method uses the proportion of the cell source of the test subject to evaluate the response of the test subject to cancer drugs. This further includes the following. In some embodiments, the cancer drug is a hormone, an immunotherapy, X It is a cancer drug or an anticancer agent. In some embodiments, the cancer drug is lenalidomide pembrolizumab, trastuzumab, bevacizumab, rituximab, ibrutinib, Topapillomavirus quadrivalent (types 6, 11, 16, and 18) vaccine, pertuzumab, Pemetrexed, nilotinib, nilotinib, denosumab, abiraterone acetate, promacta Imatinib, everolimus, palbociclib, erlotinib, bortezomib, bortezomib MIB, or its generic equivalent.

[0348] In some embodiments, the test subject has been treated with cancer drugs, The method involves testing to determine whether to intensify or discontinue cancer drugs in the test subjects. This further includes using the proportion of the target cell source. In some embodiments, the subject of test is , having undergone surgical intervention to treat cancer, this method corresponds to surgical intervention This further includes using the cell source percentage of the test subject to evaluate the state of the test subject. nothing.

[0349] In some embodiments, this method involves multiple points in time (e.g., two or more) across an epoch. It is repeated at each of the following points in time (at point 1, at three or more points in time, at four or more points in time), and by For the test subjects at each time point, the proportion of multiple cell sources (e.g., tumors) was considered. Obtain the proportion of a cell source (e.g., tumor), and use the proportions of the multiple cell sources (e.g., tumors) The state or progression of the target disease during the epoch is determined by the first cell source over the epoch (for example, The determination is made by an increase or decrease in the percentage of tumors.

[0350] In some such embodiments, an epoch is a period of several months, and there are multiple points in time. Each point in time is a different point in time within a period of several months. In some embodiments, the period of several months is The intervals are 1-4 months, 4-8 months, 8-12 months, 12-18 months, and 18 The period is between 1 month and 24 months, or more than 24 months. In some embodiments, the period is several months. It is less than 4 months old.

[0351] In some embodiments, an epoch is a period of several years, and each of several time points is: These are different points in time within a period of several years. In some embodiments, the period of several years is 2 to 10 years. It is a year. In some embodiments, the period of a year is 1 to 5 years, 5 to 10 years, 10 years to 15 years, 15-20 years, or more than 20 years.

[0352] In some embodiments, an epoch is a period of several hours, and each of several time points is , different points in time within a period of several hours. In some embodiments, the period of several hours is 1 hour The interval is ~6 hours. In some embodiments, the period of several hours is 1 hour ~ 3 hours, 3 hours ~6 hours, 6-9 hours, 9-12 hours, 12-18 hours, 18-24 hours It is between 24 hours or more.

[0353] In some embodiments, the method involves the proportion of the target first cell source (e.g., tumor) being epochal. If a change of only a threshold amount is observed across the test, the diagnosis of the subject being tested should be changed. The method further includes the following. In some embodiments, the method involves the first cell source of the target (e.g., tumor ) If the proportion changes by a threshold amount over epochs, the prognosis of the subject is considered to be Further includes modifications. In some embodiments, the method involves the first cell source of the subject (e.g.) For example, if the tumor percentage changes by a threshold amount over an epoch, This further includes modifying the treatment of the elephant. In some of the embodiments described above, the threshold is 1 percent Over cents, over 5 percent, over 10 percent, over 20 percent, over 30 percent, The threshold is greater than 40 percent or greater than 50 percent. In some embodiments, the threshold This is more than 2 times, more than 3 times, more than 4 times, or more than 5 times.

[0354] In certain embodiments, this method is used before cancer treatment (e.g., before surgical resection or therapeutic intervention). This applies not only to the first point in time (which is when cancer is present), but also to after cancer treatment (for example, after surgical resection or therapeutic intervention). This is also performed at a second time point, and the cell source is determined by the method disclosed at each time point (for example, The methods disclosed are used to monitor the effectiveness of treatment by comparing tumor rates. For example, if the tumor rate at the second time point decreases compared to the tumor rate at the first time point... In total, the treatment is considered successful. However, the tumor rate at time 2 is higher than at time 1. If the tumor rate increases compared to the rate of other tumors, the treatment is considered unsuccessful. Other embodiments So, both the first and second time points are before cancer treatment (e.g., surgical resection or therapeutic intervention). (before). In yet another embodiment, both the first and second time points are after cancer treatment ( For example, before surgical resection or therapeutic intervention, this method is used to determine the effectiveness or efficacy of the treatment. It is used to monitor fruit loss. In yet another embodiment, a biological sample (cf DNA samples are obtained from the test subject (e.g., cancer patient) at the first and second time points. For example, to monitor the progression of cancer, you can check whether the cancer is in remission (for example, after treatment). To determine whether or not the disease persists, to monitor or detect residual disease or disease recurrence, Alternatively, the data may be analyzed to monitor the effectiveness of a treatment (e.g., therapy).

[0355] A person skilled in the art can test a biological sample at any number of time points (e.g., cancer The patient's cancer status is determined by obtaining information from the patient and analyzing it according to the methods disclosed herein, for example, tumor It will be easy to understand that it can be monitored (by percentage). In some embodiments, The first and second time points are time units ranging from approximately 15 minutes to a maximum of approximately 30 years, for example, approximately 30 Minutes, for example, about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 1 5, 16, 17, 18, 19, 20, 21, 22, 23, or approximately 24 hours, for example, approximately 1, 2, 3, 4, 5, 10, 15, 20, 25 or about 30 days, for example, about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 months, or for example, about 1, 1.5, 2 2.5, 3, 3.5, 4, 4.5, 5, 5.5, 6, 6.5, 7, 7.5, 8, 8.5 , 9, 9.5, 10, 10.5, 11, 11.5, 12, 12.5, 13, 13.5, 1 4, 14.5, 15, 15.5, 16, 16.5, 17, 17.5, 18, 18.5, 1 9, 19.5, 20, 20.5, 21, 21.5, 22, 22.5, 23, 23.5, 2 4, 24.5, 25, 25.5, 26, 26.5, 27, 27.5, 28, 28.5, 2 9, 29.5, or approximately 30 years are divided into time units. In other embodiments, biological Target samples are collected at least once every three months, at least once every six months, and at least once a year. times, at least once every two years, at least once every three years, at least once every four years, or less It can be obtained from patients at least once every five years.

[0356] Determining the estimated cell source proportion of the test subjects

[0357] Block 302. Referring to Block 302 in Figure 3A, the subject (for example, the test subject) A method for estimating the cell source proportion is provided. In some embodiments, the subject is human. In some embodiments, the subject is a man or woman of any stage (e.g., male, female). , or children). In some embodiments, the proportion of the target cell source is a single cell source. Derived from. In some embodiments, the proportion of the cell source in question is derived from two or more cell sources. In some embodiments, the cell source proportion is as described above with respect to block 202. That is the case.

[0358] Block 304. Referring to Block 304, this method is electronically used to control multiple self Each cell of a Lee fragment (for example, multiple cell-free fragments originating from the biological sample in question) This is followed by obtaining the corresponding methylation pattern of the free fragment, where each cell free The corresponding methylation pattern of the fragment is determined by (i) each fragment in the biological sample obtained from the subject. Determined by methylation sequencing of one or more nucleic acid samples, (i i) The methylation status of each CpG site among the corresponding multiple CpG sites in each fragment is included. In one embodiment, referring to block 306, the multiple cell-free fragments are 500 nu It has an average length less than a creotide. In some embodiments, the cell-free fragment is block As mentioned above, regarding Ku204, it originates from a biological sample.

[0359] In some embodiments, the biological sample is the subject's blood, whole blood, plasma, serum, urine, or cerebral vertebrae. It includes or consists of cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites. In such embodiments, the biological sample may be the blood, whole blood, plasma, serum, urine, or cerebral vertebral column of the subject. Cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites, as well as other components of the subject (for example, It may also contain solid tissue, etc.

[0360] Such biological samples contain cell-free nucleic acid fragments (e.g., cfDNA fragments). In several embodiments, the biological sample is cell-free for preparation for sequencing analysis. The nucleic acids are processed to be extracted. In some non-limiting examples, the cells are processed in some embodiments. Free nucleic acid fragments are obtained from biological samples (e.g., blood) collected from subjects in K2 EDTA tubes. Extracted from the sample. If the biological sample is blood, the sample should be taken 2 hours after collection. Within that time, first rotate the biological sample in a double-rotation system at 1000g for 10 minutes, and then the obtained plasma The plasma is processed by rotating it at 2000g for 10 minutes. Afterwards, the plasma is mixed with 1 mL of ali. It is stored at -80°C in a coated container. In this way, for the purpose of cell-free nucleic acid extraction, A suitable amount of plasma (e.g., 1-5 mL) is prepared from the biological sample. In such embodiments, cell-free nucleic acids are extracted using the QIAamp Circulating Nucleic Acid Kit (Qiagen). It is then eluted into DNA suspension buffer (Sigma). In some embodiments, purified ce Nucleic acid free should be stored at -20°C until use. For example, Swanton, et al., 2017. “Phylogenetic ctDNA analysis depicts early stage lung cancer evolution,” Natu See re, 545(7655): 446-451 (which is incorporated herein by reference). (Assuming...) To prepare cell-free nucleic acids from biological methods for sequencing purposes, Other equivalent methods may be used, all of which fall within the scope of this disclosure.

[0361] In some embodiments, cell-free nucleic acid fragments obtained from biological samples are used in this disclosure. It is any defined form of nucleic acid, or a combination thereof. For example, several real In this application, cell-free nucleic acids obtained from biological samples are a mixture of RNA and DNA. .

[0362] In one embodiment, the cell-free nucleic acid fragment contains unmethylated cytosine and uracil It is processed to convert to. In one embodiment, the method converts methylated cytosine. Instead, use bisulfite treatment of DNA to convert unmethylated cytosine to uracil. Example For example, EZ DNA Methylation (trademark) - Gold, EZ DNA Methylation (trademark) - Direct or an EZ DNA Methylation (trademark) - Lightning kit (Zymo Research Corp (California) Commercially available kits (such as those from Irvine, Arizona) are used for bisulfite conversion. In another embodiment, the conversion of unmethylated cytosine to uracil is achieved using an enzymatic reaction. It is accomplished. For example, the conversion is APOBEC-Seq (NEBiolabs, Ipswick, Massachusetts). Commercially available kits for the conversion of unmethylated cytosine, such as (C), to uracil. You can use it.

[0363] A sequencing library is prepared from the converted cell-free nucleic acid fragments. (Optional) The sequencing library uses multiple hybridization probes, Cell-free nucleic acid fragments or genomic regions that are informative about cellular origin Hybridization probes are used for specific cell-free nucleic acid fragments, or Hybridize the target region and use the resulting fragments for subsequent sequencing and analysis. Or a short oligonucleotide to concentrate over a region. In some embodiments, Hybridization probes provide valuable information about cellular origins at specific CpG sites. It is used to perform targeted high-depth analysis of the set. Once prepared, sequencing The grave library or a portion of it is sequenced to obtain multiple sequence reads. .

[0364] In some embodiments, sequencing includes methylation sequencing. In that embodiment, the methylation sequencing is paired-end sequencing. In some embodiments, methylation sequencing is single-read sequencing. In some embodiments, methylation sequencing is performed using whole-genome methylation sequencing. In some embodiments, methylation sequencing involves multiple nucleic acid probes. The target sequencing used involved multiple sequence groups, with each sequence group corresponding to multiple nucleic acid probes. It is associated with at least one corresponding nucleic acid probe. In some embodiments, Each of the multiple sequence groups contains at least two corresponding nucleic acids among the multiple nucleic acid probes. It is associated with the probe.

[0365] In some embodiments, multiple nucleic acid probes (for example, used for target sequencing) The probes include over 1,000 nucleic acid probes, over 2,000 nucleic acid probes, Over 3,000 nucleic acid probes, over 4,000 nucleic acid probes, over 5,000 nucleic acids Probes, over 10,000 nucleic acid probes, over 20,000 nucleic acid probes or 3 It contains more than 0,000 nucleic acid probes. In some embodiments, it contains 1,000 nuclei There are multiple nucleic acid probes, ranging from acid probes to nucleic acid probes, totaling around 30,000.

[0366] In some embodiments, methylation sequencing (for example, as described herein) or carried out according to any methylation sequencing method known in the art. (This results in one or more 5-methylcytosine (5mC) and / or in each fragment. Alternatively, 5-hydroxymethylcytosine (5hmC) can be detected.

[0367] In some embodiments, methylation sequencing involves the sequence reads of each fragment. , one or more unmethylated cytosines or one or more methylated cytosines This includes conversion to one or more corresponding uracils. In some embodiments, 1 One or more uracils, as one or more corresponding thymines, form a methylated sequence. Detected during sync. In some embodiments, one or more unmethylated citrosin The conversion of one or more methylated cytosines can be done through chemical conversion, enzymatic conversion, and or a combination of those is included.

[0368] In some embodiments, among the multiple corresponding CpG sites in each fragment, each CpG site The methylation status is: a) If each CpG site is methylated, methylation sequencing will If it is determined that the site is methylated, then b) if the site is not methylated in each CpG, If determined by chill sequencing, it is in a non-methylated state, and c) each CpG site If the methylation state cannot be called either methylated or unmethylated, select "Other". It will be flagged as such.

[0369] Block 308. Referring to Block 308, this method is used for each of the multiple cell-free fragments. The cell-free fragment is mapped to a set of sequences within a set of sequences, thereby enabling multiple self-free fragments. - A set of fragments is obtained, and each set of cell-free fragments is a different sequence from among multiple sequence groups. It will be mapped.

[0370] In some embodiments, referring to block 310, the multiple array groups are 1000~ It consists of 100,000 sequence groups. In some embodiments, the number of sequence groups is 1500 It consists of a group of sequences from 0 to 80,000. In some embodiments, multiple groups of sequences are as described above. As explained with respect to block 210, it consists of any number of arrays.

[0371] Referring to block 312, in some embodiments, each of the multiple sequence groups is: On average, it has 10 to 1200 residues. In some embodiments, multiple sequences Each sequence group has, on average, 10 to 10,000 residues. In some embodiments, Each of the multiple sequence groups has, on average, 10 to 500 residues. In the application morphology, each sequence group has an average of 10 to 100 residues. In some embodiments, each of the multiple sequence groups has, on average, 25 to 100 residues. It has. In some embodiments, each of the multiple sequence groups has, on average, 5000 to 1 It has 0,000 residues.

[0372] Furthermore, with respect to block 314, in some embodiments, each of the multiple sequence groups 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 1 It includes or consists of 7, 18, 19, 20 or more CpG sites. In that embodiment, each of the multiple sequence groups is a set of 2 to 100 sequences in the human reference genome. It consists of consecutive CpG sites. In some embodiments, each sequence group of multiple sequence groups is 2~ It consists of 50 consecutive CpG sites. In some embodiments, each of the sequence groups of multiple sequence groups It consists of 50 to 100 consecutive CpG sites. In some embodiments, multiple sequences Each sequence group in the group consists of at least two consecutive CpG sites.

[0373] Block 316. Referring to Block 316, this method involves multiple cells of a self-free fragment. Each self-free fragment within each training set This is followed by assigning the cancerous state of the cell-free fragments to the cancer of the cell-free fragments. The state is the output of the classifier when the methylation pattern of each cell-free fragment is input into the classifier. It is a function of either the first cancer state or the second cancer state. Block 318 Referring to the above, in some embodiments, the first cancer state is cancer, and the second cancer is The state is the absence of cancer. In some embodiments, the first cancer state is cancer, The second cancer state is the absence of cancer. In some embodiments, the cell-free fragment is This condition is one of several cancerous conditions (see, for example, block 206). As mentioned above.

[0374] In some embodiments, classification systems used to assign the state of self-free fragments The setup involves a first model for the first cancer state and a second model for the second cancer state. Including the first model, the first hybrid model is a first mixed model that includes a first set of submodels, and the second model This is a second mixed model that includes a second set of submodels, and the first and second set of submodels Each submodel of the model relates to the source of cell-free fragments in the corresponding biological sample. Each represents an independent, corresponding methylation model. In some embodiments, the classification device is It has the form of formula (1) or formula (3).

[0375] Block 320. Referring to Block 320 in Figure 3B, this method involves multiple sequences. In each set of cell-free fragments, the first cancer state is assigned to the subject. The method further includes calculating a first representative value for the number of Rufley fragments. In some embodiments, Referring to block 322, the first representative value is each of the cell-free fragments across multiple sequence sets. In a set, the arithmetic of the number of self-free fragments from a subject assigned the first cancer state. Average, weighted average, midrange, mid-hinge, ternary mean, Windsor mean, average, t is the mode.

[0376] Block 324. Referring to Block 324, this method involves cells across multiple array sets. Cell free from the subject assigned the second cancer state in each set of free fragments. This further includes calculating a second representative value for the number of fragments. In some embodiments, blocks Referring to 326, the second representative value is for each set of cell-free fragments across multiple sequence sets. In the first cancer state, the arithmetic mean of the number of cell-free fragments from the assigned subject, plus Multiple mean, midrange, mid-hinge, ternary mean, Windsor mean, mean, or mode It is a value.

[0377] Block 328. Referring to Block 328, this method provides the first representative value and the second representative value. The process proceeds by estimating the proportion of the target cell source using the values. In some embodiments, The cell source percentage includes the tumor percentage. In some embodiments of block 330, Estimating the tumor percentage involves dividing the first representative value by the second representative value.

[0378] In some embodiments, the cell source proportion is related to the disease associated with the cell source in the test subject. For example, as a basis or partial basis for determining treatment options for treating cancer. It is used as a basis for monitoring the treatment. In some embodiments, the cell source ratio is used as a basis for monitoring the treatment. It is used as a foundation. In some embodiments, given the estimated cell source proportion of the subject, It is possible to determine that a particular treatment option is not effective, or is unlikely to be effective, for the target patient. Yes, there is. For example, checkpoint immunotherapy involves cytotoxic T cells becoming dysfunctional, and apo It would not be effective if ptosis occurs. Such a situation is, for example, the target organism If multiple fragments from the sample are determined to originate from cytotoxic T cells in the blood, As shown, in some embodiments, the estimated cell source percentage is used to monitor the minimum residual disease dose. Help the ring.

[0379] Those skilled in the art will recognize the embodiments disclosed in the previous section (for example, "to estimate cell source proportions"). (See "Identification of Characteristics") Any of the above refers to the estimated cell source percentage of the test subject as described herein. It is recognized that this can be applied to any combination of methods and embodiments for determining Hey. [Examples]

[0380] Example 1 - Increase in median ctDNA rate by cancer stage

[0381] Referring to Figure 4, the subjects are those with stage I cancer, regardless of the type of cancer they have. They are grouped into II, III, and IV. In Figure 4, the x-axis represents which of each object The graph shows whether the patient has a cancer stage, and the y-axis shows the observed ctDNA rate for each subject. The method used to calculate the proportion of cfDNA in each subject within the cohort is The process involves obtaining a first set of nucleic acid fragment sequences in electronic form from a biological sample, and the biological sample is self-reactive. —Contains nucleic acid molecules.

[0382] Figure 4 shows the types of cancer among subjects with cell-free sequence reads indicating underlying cancer. Regardless, this analysis shows how the ctDNA rate changes depending on the cancer stage. Therefore, Figure 4 is determined by the clinical stage classification (stages 1-4). As the disease becomes more severe, more evidence is needed regarding the proportion of cellular sources (larger ctDNA). Figure 4 shows that the rate is found in cfDNA. (See Example 3 for details) This is a general case throughout, while this slope This indicates the presence of a violation (outlier) in the direction. Such outliers are shown in Figure 4. This is suggestive and best explained by clinical misclassification. Therefore, Figure 4 is the basis This shows the typical expected cell source percentage in cfDNA, which is a fundamental component of the underlying disease. Figure 4 also shows that there are individuals in stage 4 that exhibit a very low shedding rate. This indicates that different substates exist within Stage 4.

[0383] Figure 4 shows the shedding rate (ctDNA) as a basis for setting meaningful and useful thresholds. This indicates that the rate can be used.

[0384] Example 2 - Acquisition of multiple sequence reads

[0385] Figure 5 shows a method for preparing nucleic acid samples for sequencing according to one embodiment. This is a flowchart of Method 500. Method 500 includes, but is not limited to, the following steps. It is not done. For example, any step of method 500 is a quantitative measure for quality control. This may include steps or other laboratory assay procedures known to those skilled in the art.

[0386] In block 502, a nucleic acid sample (DNA or RNA) is extracted from the target. The sample may be any subset of the human genome, including the whole genome. Extracted from subjects known to have or suspected to have cancer. That's fine too. The sample can be blood, plasma, serum, urine, feces, saliva, other types of body fluids, or so Any combination of these may be included. In some embodiments, for obtaining a blood sample The method (e.g., syringe or finger prick) may yield a tissue biopsy that may require surgery. The procedure may be less invasive than the one used for the extraction. The extracted sample contains cfDNA and / or It may contain ctDNA. In healthy individuals, the human body naturally removes cfDNA and other cellular debris. This can be done. If the subject has cancer or disease, the ctDNA in the extracted sample can be examined. It may be present at a level detectable for diagnostic purposes.

[0387] In block 504, the sequencing library is prepared. During library preparation, Through pteroligation, a unique molecular identifier (UMI) is assigned to a nucleic acid molecule (e.g., a DNA molecule). It is attached to the end of the DNA fragment. UMI is a short nucleus that is attached to the end of the DNA fragment during adapter ligation. It is an acid sequence (e.g., 4-10 base pairs). In some embodiments, UMI is a specific DNA It functions as a unique tag that can be used to identify sequence reads derived from fragments. It is a degenerate base pair. During PCR amplification following adapter ligation, UMI is added. It is replicated along with the DNA fragment. This allows for downstream analysis to use the same original fragment. The resulting sequence reads can be identified.

[0388] In block 506, the target DNA sequence is enriched from the library. During enrichment, Hybridization probes (also referred to as "probes" in this specification) are used in cancer (and The presence or absence of disease, the state of cancer, or the classification of cancer (e.g., cancer class or tissue of origin). It is used to target and pull down informational nucleic acid fragments. In this workflow, the probe anneals (or hives) to the target (complementary) strand of DNA. The target chain is sometimes designed to reduce the "positive" chain (e.g., mR). Even if it is a chain that is transcribed into NA and then translated into protein, the complementary "negative" It may also be a strand. The probe length is in the range of 10, 100, or 1000 base pairs. It is possible. In one embodiment, the probe is designed based on a methylation site panel. In this application, the probe is a genome suspected to correspond to a specific cancer or other type of disease. To analyze specific mutations or target regions (for example, in humans or other organisms), It is designed based on a panel of genes. Furthermore, the probe targets the overlapping portion of the target region. It may be covered. In block 408, these probes are general to nucleic acid samples. It is used for performing sequence reads.

[0389] Figure 6 is a graph illustrating the process for obtaining sequence reads according to one embodiment. Figure 6 depicts an example of nucleic acid segment 800 from a sample. Here, nucleic acid segment 600 can be a single-stranded nucleic acid segment, such as a single strand. In the application, nucleic acid segment 600 is a double-stranded cfDNA segment. (Illustrated example) So, there are three regions of nucleic acid segments that can be targeted by different probes: 605A, 05B and 605C are depicted. Specifically, the three regions 605A, 605B, Each of 605C and 605C includes an overlapping position on nucleic acid segment 600. An example of this position is shown in Figure 5 as the cytosine ("C") nucleotide base 602. It is there. Cytosine nucleotide base 602 is near the first end of region 605A, region 605 It is located near the center of B and the second end of region 605C.

[0390] In some embodiments, one or more (or all) of the probes are used to target specific gamma or specific genomes (e.g., human or another organism) suspected to correspond to other types of diseases Based on gene panels or methylation site panels for analyzing specific mutations or target regions. It is designed to encompass all expression levels of the genome, also known as "whole exome sequencing." Instead of sequencing genes, target gene panels or methylation site panels are used. By using it, method 600 increases the sequencing depth of the target region. It can be used, where depth is the number of times a given target sequence in the sample has been sequenced. This refers to counting the number of elements. Increasing the sequencing depth allows for the input of nucleic acid samples. It is possible to reduce the amount of force required.

[0391] Hybridization of 600 nucleic acid samples using 1 or 1 probes Then, target sequence 670 is identified. As shown in Figure 6, target sequence 670 is a hybrid. This is the nucleotide sequence of region 605 targeted by the nucleotide probe. Furthermore, target sequence 670 can also be described as a hybridized nucleic acid fragment. For example, Target sequence 670A is the region 60 targeted by the first hybridization probe. Corresponding to 5A, target sequence 670B is targeted by the second hybridization probe. Corresponding to region 605B, the target sequence 670C is the third hybridization phase. This corresponds to region 605C targeted by the lobe. Cytosine nucleotide base 602 Each region 605A-C targeted by the hybridization probe Considering its position, each target sequence 670 has a specific position on the target sequence 670. It contains the nucleotide base corresponding to synnucleotide base 602.

[0392] After the hybridization step, the hybridized nucleic acid fragments are captured, and It can also be amplified using PCR. For example, the target sequence 670 can be enriched and then... A concentrated sequence 680 that can be scanned can be obtained. Several embodiments Then, each enriched sequence 680 is replicated from target sequence 670. Target sequences 670A and 6 The enriched sequences 680A and 680C, which are amplified from 70C respectively, also have 6 sequence reads each. Contains a thymine nucleotide base located near the 80A or 680C end. Used below. In this case, the reference allele (for example, cytosine nucleotide base 602) is mutated. Substitute mutant nucleotide bases (e.g., thymine nucleotide bases) in the enriched sequence 680. It is considered an allele. Furthermore, each enriched sequence 680B amplified from the target sequence 670B is considered to be each enriched It contains a cytosine nucleotide base located near or at the center of the contracted sequence 680B.

[0393] In block 508 of Figure 5, the sequence read is an enriched DNA sequence, for example, the enriched DNA sequence shown in Figure 6. It is generated from a reduced sequence of 680. The sequencing data is generated by means known in the art. It may be obtained from enriched DNA sequences by, for example, the synthesis technique (Illumin a) Pyro-sequencing (454 Life Sciences), Ion Torrent sequence technology cing), single-molecule real-time sequencing (Pacific Biosciences), sequencing SOLiD sequencing, nanopore sequencing (Oxford Nanopore Next-generation sequencing (NGS), including paired-end sequencing (e Technologies), or e Technologies. ) may include technology. In some embodiments, a reversible dye terminator is used. We perform ultra-parallel sequencing using sequencing bisynthesis.

[0394] In some embodiments, the sequence reads determine the alignment position information. Alignment to a reference genome may be performed using methods known in the aforementioned technical field. Inment position information is the start nucleotide base and end nucleotide base of a given sequence read. It can indicate the start and end positions of regions in the reference genome corresponding to base pairs. The position information includes the length of the sequence reads, which can be determined from the start and end positions. A region in the reference genome may be associated with a gene or a segment of a gene.

[0395] JPEG2026048644000022.jpg50153

[0396] Example 3 - Cell-Free Genome Atlas Study (CCGA) Cohort

[0397] The subject from CCGA[NCT02889978] was used in the embodiments of this disclosure. CCGA is 14 In locations with over 0 participants, we enrolled over 15,000 demographically balanced participants. This is a multicenter observational cfDNA-based early cancer detection study.

[0398] In this example, we look at one of the CCGA substudies. New treatment history is defined at registration. Subjects diagnosed with no cancer (C, Case) and participants not diagnosed with cancer (non-cancer) Blood was collected from [NC] (control). This pre-planned substudy included: 878 cases and 580 control groups across 20 types of tumors and all clinical stages. The sample includes 169 assay controls (n=1627).

[0399] All samples were analyzed using the following method: 1) Paired targeting of cfDNA and leukocytes (WBCs) Quensing (60,000X, 507 gene panel); WBC by joint caller 1) Removal of somatic mutations and residual technical noise; 2) Paired whole-genome sequencing of cfDNA and WBCs. (WGS, 35X); Novel machine learning algorithm generates cancer-related signal scores; Joint analysis identifies shared events; 3) cfDNA whole-genome bisulfite sequencing WGBS, 34X); Generate normalized scores using abnormally methylated fragments. Target Up In SEI, somatic variants (SNVs / indels) of cfDNA matched non-tumor WBCs, while in NC, This accounted for 76% of all mutants, and 65% of mutant C. Somatic mosaicism (e.g., clonal hematopoiesis) In line with this, the number of mutants that matched WBC increased with age; some have been It was a non-standard loss-of-function mutation that had not been reported. After removing the mutant from the WBC, the standard The driver somatic cell mutant was highly specific to C (for example, NC with the mutant was 0). In contrast, for EGFR and PIK3CA, C was 11 and 30 respectively. Similarly, in WGS Of the eight NCs in which somatic cell copy number changes (SCNAs) were detected, four were derived from whole cell backs (WBCs). Yes, it was. CCGA's WGBS data provided informational high-fragment and low-fragment levels of Cp. The ratio of G (1:2) was identified; and the methylation score was calculated using a subset of this ratio. In all assays, a consistent "cancer-like" signal was observed in <1% of NC participants. (Indicates the possibility of undiagnosed cancer). Increased in NC vs. Stage I-III vs. Stage IV. A trend was observed (nonsyn. SNVs / indels per Mb [Mean±SD] NC: 1.01±0.86, stages I -III: 2.43±3.98; stage IV: 6.45±6.79; WGS score NC: 0.00±0.08, I-III: 0.27±0 .98; IV: 1.95± 2.33; methylation score NC: 0±0.50; I-III: 1.02±1.77; IV: 3.94 (±1.70). These data make it possible to achieve >99% specificity for invasive cancer. This demonstrates the potential of cfDNA assays for early cancer detection. .

[0400] Example 4 - Example of a cell source

[0401] In some embodiments, the cell source of any embodiment of this disclosure is a primary cell of a common origin. This is the first cancerous state of the site. In some embodiments, the first cancerous state is breast cancer. Lung cancer, prostate cancer, colon cancer, kidney cancer, uterine cancer, pancreatic cancer, esophageal cancer, lymphoma, Head / neck cancer, ovarian cancer, hepatobiliary tract cancer, melanoma, cervical cancer, multiple myeloma, leukemia, thyroid gland This could be a bladder cancer, stomach cancer, or a combination thereof.

[0402] In some embodiments, the cell source of any embodiment of this disclosure is a tumor of a certain cancer species, This is a part of it. In some embodiments, the tumor is adrenocortical carcinoma, pediatric adrenocortical Hmm, AIDS-related cancer tumors, Kaposi's sarcoma, anal cancer-related tumors, appendiceal cancer-related tumors, Astrocytoma, pediatric (brain cancer) tumors, atypical malformations / rhabdoid tumors, central nervous system ( Brain cancer tumors, basal cell carcinoma of the skin, cholangiocarcinoma-related tumors, bladder cancer tumors, pediatric bladder cancer tumors Scars, bone cancers (e.g., Ewing's sarcoma, osteosarcoma, malignant fibrous histiocytoma), tissue, brain tumors, breast Cancer tissue, pediatric breast cancer tissue, pediatric bronchial tumors, Burkitt lymphoma tissue, carcinoid tumors Gastrointestinal tumors, pediatric carcinoid tumors, cytoplasmic type of tumors of unknown primary origin, pediatric cytoplasmic type of tumors of unknown primary origin, pediatric heart (Cardiac) tumors, central nervous system tumors (e.g., atypical malformations in children / rhabdoid brain tumors), small Pediatric germ cell tumors, pediatric germ cell tumors, cervical cancer tissue, pediatric cervical cancer tissue, bile duct cancer tissue, small Childhood chordoma tissue, chronic myeloproliferative neoplasms, colorectal cancer tumors, pediatric colorectal cancer tumors, pediatric craniopharynx Tumor tissue, ductal carcinoma in situ (DCIS), pediatric embryonic tumors, endometrial cancer (uterine cancer) tissue, small Childhood ependymoma tissue, esophageal cancer tissue, pediatric esophageal cancer tissue, sensory neuroblastoma (head and neck cancer) tissue, small Extracranial germ cell tumors, extragonadal germ cell tumors, ocular cancer tissue, intraocular melanoma, retinoblastoma, eggs Tuberoid cancer tissue, gallbladder cancer tissue, stomach cancer tissue, pediatric stomach cancer tissue, gastrointestinal carcinoma Idoid tumors, gastrointestinal stromal tumors (GIST), pediatric gastrointestinal stromal tumors, germ cell tumors (e.g., pediatric stromal tumors) Central nervous system germ cell tumors, pediatric extracranial germ cell tumors, extragonadal germ cell tumors, ovarian germ cell tumors, and (Testicular cancer tissue), head and neck cancer tissue, pediatric cardiac tumors, hepatocellular carcinoma (HCC) tissue. Islet cell tumors. Surgery (pancreatic neuroendocrine tumor), kidney or renal cell carcinoma (RCC) tissue, laryngeal cancer tissue, leukemia, Liver cancer tissue, lung cancer (non-small cell and small cell) tissue, pediatric lung cancer tissue, male breast cancer tissue. Malignant fibrous histiocytoma and osteosarcoma of bone, melanoma, pediatric melanoma, intraocular melanoma, pediatric intraocular melanoma Merkel cell carcinoma, malignant mesothelioma, pediatric mesothelioma, metastatic cancerous tissue, metastatic squamous cell carcinoma of unknown primary origin Cervix cancer, median linea cell type with NUT gene mutations, oral cancer (head and neck cancer) tissue, multiple Endocrine tumor syndrome tissue, multiple myeloma / plasmacytic tumor, myelodysplastic syndrome tissue, myelodysplasia Adult / myeloproliferative neoplasms, chronic myeloproliferative neoplasms, nasal cavity and paranasal sinus cancer tissue, nasopharyngeal cancer (NP) C) Tissue, neuroblastoma tissue, non-small cell lung cancer tissue, oral cancer tissue, lips and oral cancer and Oropharyngeal cancer tissue, osteosarcoma and malignant fibrous histiocytoma tissue, ovarian cancer tissue, pediatric ovarian cancer Tissue, pancreatic cancer tissue, pediatric pancreatic cancer tissue, papillomatous (pediatric larynx) tissue, paraganglioma tissue, Childhood paraganglioma tissue, paranasal sinus and nasal cavity cancer tissue, parathyroid cancer tissue, penile cancer tissue, pharynx Head cancer tissue, pheochromocytoma tissue, pediatric pheochromocytoma tissue, pituitary tumor, plasma cell tumor / multiple Myeloma, pleuropulmonary blastoma, primary central nervous system (CNS) lymphoma, primary peritoneal cancer tissue, prostate Cancer tissue, rectal cancer tissue, retinoblastoma, pediatric rhabdomyosarcoma, salivary gland cancer tissue, sarcoma (for example, Childhood hemangiomas, osteosarcomas, uterine sarcomas, etc., Sézary syndrome (lymphoma) tissue, skin cancer tissue pediatric skin cancer tissue, small cell lung cancer tissue, small intestine cancer tissue, cutaneous squamous cell carcinoma, primary tumor of unknown origin Squamous cell carcinoma, cutaneous T-cell lymphoma, testicular cancer tissue, pediatric testicular cancer tissue, throat cancer (for example) (Nasopharyngeal cancer, oropharyngeal cancer, hypopharyngeal cancer) tissue, thymoma or thymic carcinoma, thyroid cancer tissue , renal pelvis and ureteral transitional cell carcinoma tissue, tissue of unknown primary cell type, ureter or renal pelvis tissue, transitional cell Cellular carcinoma (kidney (renal cell) cancer tissue, urethral cancer tissue, endometrial uterine cancer tissue, uterine sarcoma tissue) Vaginal cancer tissue, pediatric vaginal cancer tissue, hemangiomas, vulvar cancer tissue, Wilms' tumor, or others This is a pediatric kidney tumor.

[0403] In some embodiments, the cell source of any embodiment of the present disclosure is in a first cancerous state. In some such embodiments, the first cancer state is a stage of breast cancer. Lung cancer at a certain stage, prostate cancer at a certain stage, colorectal cancer at a certain stage Stage 1 kidney cancer, stage 1 uterine cancer, stage 1 pancreatic cancer, stage 1 esophageal cancer Cancer, lymphoma at a certain stage, head / neck cancer at a certain stage, ovarian cancer at a certain stage , hepatobiliary tract cancer at a certain stage, melanoma at a certain stage, cervical cancer at a certain stage, Stage 1 multiple myeloma, stage 2 leukemia, stage 3 thyroid cancer, stage 4 It is either a stage bladder cancer or a stage of stomach cancer.

[0404] In some embodiments, the cell source of any embodiment of the present disclosure is milk at a predetermined stage Hmm, lung cancer at a specified stage, prostate cancer at a specified stage, colorectal cancer at a specified stage , kidney cancer of a specified stage, uterine cancer of a specified stage, pancreatic cancer of a specified stage, Esophageal cancer of a specified stage, lymphoma of a specified stage, head / neck cancer of a specified stage, Ovarian cancer at a specified stage, hepatobiliary tract cancer at a specified stage, melanoma at a specified stage, Cervical cancer at a specified stage, multiple myeloma at a specified stage, leukemia at a specified stage, Thyroid cancer at a specified stage, bladder cancer at a specified stage, or stomach cancer at a specified stage That is the case.

[0405] In some embodiments, the cell source of any embodiment of this disclosure is from non-cancerous tissue. In some embodiments, the cell source of any embodiment of this disclosure is derived from healthy tissue. These are derived from cells. In some embodiments, the cells of any embodiment of the present disclosure The source is the breast, lungs, prostate, large intestine, kidneys, uterus, pancreas, esophagus, lymph, ovaries, cervix, surface It comes from healthy tissue such as skin, thyroid gland, bladder, stomach, or a combination thereof.

[0406] In some embodiments, the cell source of any embodiment of the present disclosure is derived from one tissue type. In some embodiments, the cell source of any embodiment of the present disclosure is two or more tissue types. It is induced by... In some embodiments, the tissue type is one or more cell types (e.g., Includes a combination of healthy non-cancerous cells and cancerous cells. In some embodiments, the combination A cell type is a single cell type (for example, either cancerous cells or healthy, non-cancerous cells). Includes.

[0407] In some embodiments, the cell source of any embodiment of this disclosure is one cell type, 2 One cell type, three cell types, four cell types, five cell types, six cell types, seven cell types It consists of 8 cell types, 9 cell types, 10 cell types, or more than 10 cell types.

[0408] In some embodiments, the cell source of any embodiment of this disclosure is hepatocytes. In such embodiments, the cell source is hepatocytes, hepatic astrocytes (ITO cells), These are upper cells, sinusoidal endothelial cells, or any combination thereof.

[0409] In some embodiments, the cell source of any embodiment of this disclosure is gastric cells. In such an embodiment, the cell source is parietal cells.

[0410] In some embodiments, the cell source of any embodiment of this disclosure is one or more types These are human cells. In some such embodiments, the cell source is adaptive NK cells, adipose cells. Cells, alveolar cells, Alzheimer's type II astrocytes, amacrine cells, ameloblasts Astrocytes, B cells, basophils, basophil-activated cells, basophil-promoting cells, Betz Cells, bilayer ganglion cells, Böttcher cells, cardiomyocytes, CD4+ T cells, cementoblasts, Cerebellar granule cells, bile duct cells, gallbladder cells, chromaffin cells, cigar cells, club cells, ol Tycotropic cells, cytotoxic T cells, dendritic cells, enterochromaffin cells Cells, enterochromaffin cell-like cells, eosinophils, extraglomerular mesangial cells, fagot cells, Fat pad cells, gastric long cells, goblet cells, gonadotropin cells, hepatic stellate cells, hepatocytes, hyperdifferentiated Neutrophils, intraglomerular mesangial cells, juxtaglomerular cells, keratinocytes, renal proximal tubular brush Edge cells, Kupffer cells, mammary gland stimulating cells, Leydig cells, macrophages, macula densa Cells, mast cells, megakaryocytes, melanocytes, pleurocytes, monocytes, natural killer cells, NATURAL cells Natural killer T cells, glitter cells, neutrophils, osteoblasts, osteoclasts, osteocytes, eosinophilic cells Parathyroid cells, Paneth cells, parafollicular cells, parasol cells, parathyroid gland cells, parietal cells, Small cell neurosecretory cells, pegylated cells, pericytes, peritubulomyloid cells, platelets, podocytes Regulatory T cells, reticulocytes, retinal bipolar cells, retinal horizontal cells, retinal ganglion cells, preretinal cells Progenitor cells, sentinel cells, Sertoli cells, somatoammotrophic cells, growth somatotropic cells, astrocytes, supporting cells, T cells, T helper cells, telocytes, Tendon cells, thyroid-stimulated cells, transitional B cells, hairy cells (human), tufted cells, unipolar brush cells These include white blood cells, solid alveoli, or any combination thereof. In the embodiment, such cells of the cell source are healthy. In the alternative embodiment, the cells of the cell source are healthy. Cells like these are cancerous.

[0411] In some embodiments, the cell source of any embodiment of the present disclosure is such that the cell type is simple It is any combination of cell types, provided that it originates from one organ. In embodiments such as the breast, lung, prostate, colon / rectum, kidney, uterus, It is the pancreas, esophagus, blood, head / neck, ovaries, liver, cervix, thyroid gland, bladder, or stomach. In one embodiment, this single organ is healthy. In an alternative embodiment, this single organ is , the patient has cancer originating from a single organ. In yet another embodiment, this single organ is, The patient has cancer that originated in an organ other than a single organ and has metastasized to that single organ.

[0412] In some embodiments, the cell source of any embodiment of the present disclosure is such that the cell type is It is any combination of cell types, provided that they originate from a predetermined set of containers. In some such embodiments, this predetermined set of organs is the breast, lungs, prostate, colon / Rectum, kidneys, uterus, pancreas, esophagus, blood, head / neck, ovaries, liver, cervix, thyroid gland, bladder and These are any two organs within the set of the stomach and the ribs. In some embodiments, these predetermined organs The set is healthy. In an alternative embodiment, this set of predetermined organs is the set of predetermined organs The patient has cancer originating from one of the organs within the body. In an alternative embodiment, this A set of organs originates from organs other than the set of organs specified, and is transferred to the set of organs specified. He has contracted cancer that was transmitted to him.

[0413] In some embodiments, the cell source of any embodiment of the present disclosure is such that the cell type is It is any combination of cell types, provided that they originate from a predetermined set of containers. In some such embodiments, this predetermined set of organs is the breast, lungs, prostate, colon / Rectum, kidneys, uterus, pancreas, esophagus, blood, head / neck, ovaries, liver, cervix, thyroid gland, bladder and These are any three organs in the set of the stomach. In some embodiments, these predetermined organs The set is healthy. In an alternative embodiment, this set of predetermined organs is the set of predetermined organs The patient has cancer originating from one of the organs within the body. In yet another alternative embodiment, this The set of specified organs is derived from organs other than the set of specified organs, and the set of specified organs He has cancer that has metastasized to his body.

[0414] In some embodiments, the cell source of any embodiment of the present disclosure is such that the cell type is It is any combination of cell types, provided that they originate from a predetermined set of containers. In some such embodiments, this predetermined set of organs is the breast, lungs, prostate, colon / Rectum, kidneys, uterus, pancreas, esophagus, blood, head / neck, ovaries, liver, cervix, thyroid gland, bladder and Any four, five, six, or seven organs in the set of stomachs. In some embodiments, this predetermined set of organs is healthy. In alternative embodiments, This set of organs is affected by cancer originating from one of the organs within the set of organs. In yet another alternative embodiment, this predetermined set of organs is a set of predetermined organs. The patient has cancer that originated in an organ other than the designated set of organs and has metastasized to that set of organs.

[0415] In some specific embodiments, the cell source of any embodiment of this disclosure is leukocytes. In some such embodiments, the cell source is neutrophils, eosinophils, basophils, lymphocytes. These are B lymphocytes, T lymphocytes, cytotoxic T cells, monocytes, or any combination thereof. .

[0416] conclusion

[0417] Multiple examples are given for a component, operation, or structure described herein as a single example. It may be provided. Finally, the boundaries between the various components, operations, and data stores are somewhat This is optional, and certain operations are illustrated in the context of specific exemplary configurations. Other function assignments This is anticipated and may fall within the scope of implementation. Generally, in the exemplary configuration, separate components The structures and functions shown as constituent elements are implemented as a combined structure or constituent element. Similarly, structures and functions presented as single components may be considered as separate components. These may be implemented as elements. These and other variations, modifications, additions, and improvements are impleme...

Claims

1. A method for identifying multiple features to estimate the proportion of target cell sources, comprising: One or more processors and for execution by the one or more processors In a computer system having memory that stores one or more programs There, A) To obtain a training dataset in electronic format, and the training The dataset is for each training target of multiple training targets: a) The correspondence of each cell-free fragment among the multiple cell-free fragments for corresponding training The methylation pattern, b) Including the signs of the target cancer of each of the training subjects, The corresponding methylation pattern of each cell-free fragment is (i) each training pair One or more nucleic acid samples containing each fragment from the corresponding biological sample obtained from an elephant. (ii) Determined by methylation sequencing of the following multiple in each of the above fragments It includes the methylation status of each CpG site among the CpG sites, The cancer state of the subject is either the first cancer state or the second cancer state. thing; B) Each cell-free fragment of each of the multiple cell-free fragments is selected from one of the multiple sequence groups. We map to groups and thereby obtain multiple training sets of self-free fragments. Each of the plurality of sequence groups represents a corresponding portion of the human reference genome, and self Each training set of Lee fragments is mapped to a different sequence group from the aforementioned group of sequences. To be done; C) Each of the training sets of the self-free fragments Assigning the cancer status of each cell-free fragment within the training set. The cancerous state of the cell-free fragment is determined by the methylation pattern of each cell-free fragment. The first cancer state and It is one of the conditions of the second type of cancer; D) For each of the plurality of sequence groups, (a) of the plurality of training targets (b) the state of the target cancer for each training target, and the pair mapped to each of the sequence groups. Each self-free fragment in the training set of corresponding self-free fragments is To determine the corresponding correlation scale I between the states; E) Multiple features for estimating the proportion of the target cell source, sub-sequences of the multiple sequence groups To specify as a set, each sequence group in the subset of the plurality of sequence groups is The selection criteria based on the corresponding correlation scale for each of the aforementioned sequence groups are met. Methods that include...

2. A method according to claim 1, comprising the following steps, for determining the proportion of the cell source to be tested. The procedure further includes determining, The corresponding methylation pattern of each cell-free fragment among several cell-free fragments used for testing. The method involves obtaining the corresponding methylation pattern of each cell-free fragment in electronic format. (i) one or more of the biological samples obtained from the subject of the test, each containing a fragment. Determined by methylation sequencing of nucleic acid samples, (ii) correspondence in each of the aforementioned fragments It includes the methylation state of each of the multiple CpG sites; Each of the cell-free fragments of the multiple cell-free fragments used for testing is selected from the multiple array groups. We map to a set of sequences and thereby obtain multiple test sets of self-free fragments. Therefore, each test set of the cell-free fragment is a different sequence group from the aforementioned group of sequences. It is mapped to; When the methylation pattern of each of the aforementioned cell-free fragments is input to the classification device, the output of the classification device The cancer state of a cell-free fragment, which is a function of force, is applied to multiple test sets of cell-free fragments. Assigning each self-free fragment in each test set to each self-free fragment; In each test set of cell-free fragments spanning a subset of the aforementioned plurality of sequence groups, The first cancer state is defined as the first representative value of the number of cell-free fragments from the assigned test subjects. Calculating; Tests in each test set of self-free fragments across subsets of the aforementioned set of sequences To calculate a second representative value for the number of self-free fragments from the target; Using the first representative value and the second representative value, the proportion of the cell source of the test subject is estimated. A method that includes doing something.

3. The second cancer state is the absence of cancer, and the proportion of the cell source of the test subject is the test subject The method according to claim 2, wherein the tumor percentage is included.

4. A method according to claim 1, wherein the classification device is of formula: [Math 1] It has, in the formula, 【number】 This is the first model for the state of the first cancer, "Fragment" is the methylation pattern of each of the aforementioned cell-free fragments, 【number】 This is a second model for the second stage of cancer, When R(fragment) satisfies the threshold, the cancerous state of the cell-free fragment of each fragment is as follows: The method according to claim 1, wherein the state of n can be assigned.

5. The method according to claim 4, wherein the threshold is 1 to 10.

6. The threshold is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10, as described in claim 4. The method.

7. The aforementioned correlation scale I is [Math 2] It is calculated as follows, and in the formula, i and j are independent inputs for the set {first cancer state, second cancer state}. Dex, x ｉ This refers to the training subjects among the multiple training subjects that have cancer state i. It is a number, y ｊ This applies to each sequence group to which cancer state j is assigned among the multiple training subjects. The number of training subjects that have one or more mapped cell-free fragments. 、 【number】 The method according to claim 1.

8. The aforementioned measure of relevance is a measure of correlation, mutual information, or distance metric, claim The method described in item 1.

9. The method according to claim 1, wherein the measure of relevance is the Pearson correlation coefficient.

10. The aforementioned correlation measure is an adjusted correlation coefficient, a weighted correlation coefficient, a reflected correlation coefficient, or a scale variable. The method according to claim 1, wherein the correlation coefficient has been converted.

11. The plurality of sequence groups consist of 1,000 sequence groups to 100,000 sequence groups, according to claim 1 to The method described in any one of the ten items.

12. Claim 1 to, wherein the plurality of sequence groups consist of 15,000 to 80,000 sequence groups The method described in any one of the ten items.

13. Each of the plurality of sequence groups has, on average, 10 to 1200 residues, claim. The method described in any one of items 1 to 12.

14. Each of the aforementioned plurality of sequence groups has, on average, 10 to 10,000 residues, claim The method described in any one of items 1 to 12.

15. The first representative value is each test of a cell-free fragment across a subset of the plurality of sequence groups. In the set, self-reporting from multiple test subjects assigned the first cancer state - Arithmetic mean, weighted mean, midrange, mid-hinge, ternary mean, Windsor of the number of fragments The method according to claim 2, wherein the mean is the mean or the mode.

16. The second representative value is each test of the cell-free fragments across the subset of the plurality of sequence groups. Arithmetic mean, weighted mean, and minus the number of self-free fragments from multiple test subjects in a set. Midrange, mid-hinge, ternary mean, Windsor mean, mean, or mode. The method according to claim 2.

17. Estimating the cell source proportion includes dividing a first representative value by a second representative value, according to claim 2. Method of description.

18. The aforementioned multiple training targets range from 10 training targets to 1000 training targets. A method according to any one of claims 1 to 17, comprising the subject of the G.

19. The selection criterion specifies the selection of a group of sequences that have one of the top N relevance measures, where N is The method according to any one of claims 1 to 18, wherein the integer is 50 or greater.

20. The method according to claim 19, wherein N is 500 to 5000.

21. The method according to claim 19, wherein N is 800 to 1500.

22. The methylation sequencing is paired-end sequencing, according to claims 1 to 21. The method described in any one of the items.

23. Claims 1 to 2, wherein the methylation sequencing is single-read sequencing. The method described in any one of item 1.

24. The aforementioned multiple cell-free fragments for training are less than 500 nucleotides in length. The method according to any one of claims 1 to 23, having an equal length.

25. If the first cancerous state is cancer, and the second cancerous state is the absence of cancer, The method described in any one of the requests 1 to 24.

26. The first type of cancer mentioned above includes adrenal cancer, biliary tract cancer, bladder cancer, bone / bone marrow cancer, brain cancer, and breast cancer. Cancer, cervical cancer, colorectal cancer, esophageal cancer, stomach cancer, head / neck cancer, hepatobiliary tract cancer, kidney cancer, Liver cancer, lung cancer, ovarian cancer, pancreatic cancer, pelvic cancer, pleural cancer, prostate cancer, kidney cancer, skin cancer Cancer, stomach cancer, testicular cancer, thymic cancer, thyroid cancer, uterine cancer, lymphoma, melanoma, multiple cancers It is one of the myelomas or leukemias. The second cancer state is the absence of cancer, as described in any one of claims 1 to 24. The method.

27. The first cancer state described above is adrenal cancer at a certain stage, biliary tract cancer at a certain stage, and Stage bladder cancer, stage bone / bone marrow cancer, stage brain cancer, stage Breast cancer, cervical cancer at a certain stage, colon cancer at a certain stage, esophageal cancer at a certain stage Hmm, a certain stage of stomach cancer, a certain stage of head / neck cancer, a certain stage of hepatobiliary tract cancer, Kidney cancer at a certain stage, liver cancer at a certain stage, lung cancer at a certain stage Ovarian cancer, pancreatic cancer at a certain stage, pelvic cancer at a certain stage, pleural cancer at a certain stage , prostate cancer at a certain stage, kidney cancer at a certain stage, skin cancer at a certain stage, Stage 1 stomach cancer, stage 2 testicular cancer, stage 3 thymic cancer, stage 4 thymic cancer Adenocarcinoma, uterine cancer at a certain stage, lymphoma at a certain stage, melanoma at a certain stage, It is one of the following: multiple myeloma at a certain stage, or leukemia at a certain stage. The second cancer state is the absence of cancer, as described in any one of claims 1 to 24. The method.

28. Claim 1 states that the methylation sequencing is whole-genome methylation sequencing. Method of loading.

29. The methylation sequencing is a targeted sequencing method using multiple nucleic acid probes. Furthermore, each of the plurality of sequence groups is at least one nucleic acid probe of the plurality of nucleic acid probes The method according to claim 1, which is associated with a robe.

30. The plurality of nucleic acid probes consist of 1,000 or more nucleic acid probes and 2,000 or more nuclei Acid probes, over 3,000 nucleic acid probes, over 5,000 nucleic acid probes, 10 More than 1,000 nucleic acid probes, or 1,000 to 30,000 nucleic acid probes The method according to claim 29, which includes

31. Each of the aforementioned plurality of sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 , including 13, 14, 15, 16, 17, 18, 19, 20 or more CpG sites The method according to any one of claims 1 to 30.

32. Each of the aforementioned plurality of sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 , 13, 14, 15, 16, 17, 18, 19, 20, or more consecutive CpG sections The method according to any one of claims 1 to 29, comprising the position.

33. Each of the aforementioned sequence groups corresponds to 2 to 100 consecutive CpG sites in the human reference genome. The method according to any one of claims 1 to 30.

34. The method according to claim 1, wherein the corresponding biological sample is a liquid biological sample.

35. The method according to claim 1, wherein the corresponding biological sample is a blood sample.

36. The corresponding biological sample is the blood, whole blood, plasma, serum, or urine of the subject of training. The following is a claim, comprising cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites. Method of loading.

37. The corresponding biological sample is the blood, whole blood, plasma, serum, or urine of the subject of training. The claim 1, comprising cerebrospinal fluid, feces, saliva, sweat, tears, pleural fluid, pericardial fluid, or ascites. method.

38. Of the corresponding multiple CpG sites in each of the aforementioned fragments, the methylation state of each CpG site is: If each of the aforementioned CpG sites is methylated, it is determined by the methylation sequencing. If this occurs, it is in a methylated state. If each of the aforementioned CpG sites is not methylated, it is determined by the methylation sequencing. If this occurs, it is in a non-methylated state. The methylation state of each CpG site is determined to be either methylated or unmethylated. If a call cannot be made through sequencing, it will be flagged as "Other" in the billing process. The method described in any one of items 1 to 37.

39. The methylation sequencing revealed that one or more 5-methyl A claim for detecting cytosine (5mC) and / or 5-hydroxymethylcytosine (5hmC). The method described in 38.

40. The methylation sequencing involves one or multiple sequences in the sequence reads of each fragment. A number of unmethylated cytosines or one or more methylated cytosines, one corresponding The method according to claim 38, which includes conversion to multiple uracils.

41. The one or more uracils are, in addition to one or more corresponding thymines, the methyl The method according to claim 40, which is detected during chemical sequencing.

42. One or more unmethylated cytosines or one or more methylated cytosines The transformation includes a chemical transformation, an enzymatic transformation, or a combination thereof, claim. Method 40.

43. The aforementioned first model is a first mixed model that includes a first plurality of submodels, The aforementioned second model is a second hybrid model that includes a second set of submodels, Each of the first and second submodels is the corresponding biological sample A claim representing an independent, corresponding methylation model for the source of the cell-free fragments within. The method described in any one of paragraphs 4 to 42.

44. The independent corresponding methylation models are the binomial model, the beta-binomial model, and the independent part. The method according to claim 43, wherein the method is one of a positional model or a Markov model.

45. Two or more of the first submodels described above are independent site models. 、 Two or more of the second set of submodels are independent site models. The method according to claim 43.

46. Before mapping B), one or more filter conditions are applied to the plurality of cell-free fragments. The method according to any one of claims 1 to 45, further comprising applying the following:

47. One of the one or more filter conditions is the plurality of Apply a p-value threshold to the corresponding methylation pattern of each cell-free fragment among the cell-free fragments. In some cases, the p-value threshold is determined when the methylation pattern is observed in a cohort of non-cancer subjects. The method according to claim 46, which is representative of the frequency of

48. The method according to claim 47, wherein the p-value threshold is 0.001 to 0.

20.

49. The cohort includes at least 20 subjects, and the plurality of cell-free fragments are at least The method according to claim 47, comprising 10,000 different corresponding methylation patterns.

50. The p-value threshold is determined by the corresponding methyl group of each cell-free fragment among the plurality of cell-free fragments. If the transformation pattern has a p-value of 0.10 or less, 0.05 or less, or 0.01 or less, The method according to claim 47, which satisfies the methylation pattern from the subject.

51. One of the one or more filter conditions is the plurality of Each cell-free fragment contains one of the corresponding fragments in the biological sample. Alternatively, in the corresponding sequence reads measured from multiple nucleic acid samples, the threshold number The method according to claim 46, which applies the requirement that it be represented by sequence reads. 。

52. The threshold number is 2, 3, 4, 5, 6, 7, 8, 9, 10, or an integer from 10 to 100. The method according to claim 51.

53. One of the one or more filter conditions is the plurality of Each cell-free fragment contains one of the corresponding fragments in the biological sample. Alternatively, in multiple nucleic acid samples, the requirement is that it is represented by a threshold number of cell-free nucleic acids. The method according to claim 46, wherein the item is applied.

54. The threshold number is 2, 3, 4, 5, 6, 7, 8, 9, 10, or an integer from 10 to 100. The method according to claim 53.

55. One of the one or more filter conditions is the plurality of The requirement that each cell-free fragment of a cell-free fragment has a threshold number of CpG sites is applied. The method according to claim 46.

56. The threshold number of the CpG site is at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 The method according to claim 55, wherein the CpG site is a single CpG site.

57. One of the one or more filter conditions is the plurality of Each cell-free fragment of the cell-free fragment must have a length less than a threshold number of base pairs. The method according to claim 46.

58. The base pairs of the aforementioned threshold number have a length of 1,000, 2,000, 3,000, or 4,000. The method according to claim 57, wherein there are zero consecutive base pairs.

59. The acquisition, mapping, assignment, calculation of the first and second representative values, and the epoch The estimation of the proportion of the cell source of the test subject at each of the multiple time points across these points in time is repeated. Therefore, for the test subject at each time point, the corresponding cell source ratio from among multiple cell source ratios is determined. To obtain; Using the aforementioned proportions of multiple cell sources, the state or progression of the disease being tested during the epoch is determined. This is determined by an increase or decrease in the proportion of the first cell source over the aforementioned epoch. The method according to claim 2 or 3, further comprising:

60. The aforementioned epoch is a period of several months, and each of the aforementioned multiple time points is within the period of several months. The method according to claim 59, which is a different point in time within the interval.

61. The method according to claim 60, wherein the aforementioned period of several months is less than four months.

62. The aforementioned epoch is a period of several years, and each of the aforementioned multiple points in time falls within the aforementioned period of several years. The method according to claim 59, which is at a different point in time.

63. The method according to claim 62, wherein the aforementioned period of several years is 2 to 10 years.

64. The aforementioned epoch is a period of several hours, and each of the multiple time points is within the period of several hours. The method according to claim 59, which is a different point in time within the interval.

65. The method according to claim 64, wherein the aforementioned period of several hours is 1 to 6 hours.

66. It was observed that the proportion of the first cell source of the aforementioned target changed by a threshold amount over the aforementioned epoch. Claims 59 to 65 further include changing the diagnosis of the test subject if such a condition is met. The method described in any one of the items.

67. It was observed that the proportion of the first cell source of the aforementioned target changed by a threshold amount over the aforementioned epoch. Claims 59 to 65 further include changing the prognosis of the test subject if such a change occurs. The method described in any one of the items.

68. It was observed that the proportion of the first cell source of the aforementioned target changed by a threshold amount over the aforementioned epoch. Claims 59 to 65 further include changing the treatment of the subject under test if such a condition is met. The method described in any one of the items.

69. The aforementioned thresholds are greater than 10%, greater than 20%, greater than 30%, greater than 40%, greater than 50%, greater than 2 times, greater than 3 times, The method according to any one of claims 66 to 68, wherein the ratio is more than 5 times.

70. The proportion of tumors in the test subject is 0.003 to 1.0, according to any of claims 59 to 69. The method described in any one of the items.

71. Based at least partially on the value of the cell source percentage of the test subject, the treatment of the test subject The method according to claim 2 or 3, further comprising applying a therapeutic regimen.

72. Claim 71, wherein the treatment regimen includes applying an oncology drug to the subject of the test. Methods used.

73. Claim 7, wherein the aforementioned cancer drug is a hormone, immunotherapy, X-ray examination, or anticancer agent. The method described in 2.

74. The aforementioned cancer drugs are lenalidomide, pembrolizumab, trastuzumab, and bevacizumab. , rituximab, ibrutinib, human papillomavirus tetravalent (6, 11, 16, and Type 18 vaccine, pertuzumab, pemetrexed, nilotinib, nilotinib, denosum B, abiraterone acetate, promacta, imatinib, everolimus, palbociclib, el Rotinib, bortezomib, bortezomib, or their generic equivalents, please The method described in item 72.

75. The subject of the test is being treated with an anticancer drug, and the method is the cell source of the subject of the test. The claim further includes using a percentage to evaluate the response to the aforementioned cancer drug. The method described in item 2 or 3.

76. Claim 7, wherein the aforementioned cancer drug is a hormone, immunotherapy, X-ray examination, or anticancer agent. The method described in 5.

77. The aforementioned cancer drugs are lenalidomide, pembrolizumab, trastuzumab, and bevacizumab. , rituximab, ibrutinib, human papillomavirus tetravalent (6, 11, 16, and Type 18 vaccine, pertuzumab, pemetrexed, nilotinib, nilotinib, denosum B, abiraterone acetate, promacta, imatinib, everolimus, palbociclib, el Rotinib, bortezomib, bortezomib, or their generic equivalents, please The method described in item 75.

78. The subject of the test is being treated with an anticancer drug, and the method is used to test the cells of the subject of the test. Using the source ratio, it is determined whether to increase or discontinue the cancer drug in the test subject. The method according to claim 2 or 3, further comprising determining

79. The subject of the test has undergone surgical intervention to address the cancer, and the method is Using the cell source proportion of the test subject, the state of the test subject in response to the surgical intervention was determined. The method according to claim 2 or 3, further comprising evaluating the state.

80. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. Genome regions listed in one or more of Lists 1-8 in Issue 2020 / 069350 The method according to any one of claims 1 to 79, corresponding to the method described in claim 1 to 79.

81. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. Genome regions listed in one or more of Lists 1-8 in Issue 2020 / 069350 A method according to any one of claims 1 to 80, which maps to at least 30% of 。

82. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. Genome regions listed in one or more of Lists 1-8 in Issue 2020 / 069350 The following is mapped to at least 50-95% of claims 1 to 81 The method.

83. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. One to ten unique in one or more of Lists 1 to 8 of Issue 2020 / 069350 The method according to any one of claims 1 to 82, which is mapped to the corresponding genomic region. Law.

84. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. A single, unique correspondence in one or more of Lists 1-8 of Issue 2020 / 069350 The method according to any one of claims 1 to 83, which is mapped to a genomic region.

85. For each of the training targets of the plurality of training targets, the training multi Claim 1, the number of cell-free fragments comprises at least 100,000 cell-free fragments. The method described in any one of paragraphs 84 to 84.

86. For each of the training targets of the plurality of training targets, the training multi Claim 1, the number of cell-free fragments comprises at least 100,000 cell-free fragments. The method described in any one of paragraphs 84 to 84.

87. For each of the aforementioned training targets, multiple training targets The cell-free fragment comprises at least 1 million cell-free fragments, according to claims 1 to 84. The method described in any one of the items.

88. Each of the above-mentioned sequence groups contains fewer than 100 nucleic acid residues, fewer than 500 nucleic acid residues, Nucleic acid residues less than 1000, less than 2500, less than 5000, 1 Nucleic acid residues less than 0,000, nucleic acid residues less than 25,000, nuclei less than 50,000 Acid residues, fewer than 100,000 nucleic acid residues, fewer than 250,000 nucleic acid residues, or 5 The method according to any one of claims 1 to 87, comprising less than 00,000 nucleic acid residues. 。

89. One or more processors, One or more programs for execution by the aforementioned one or more processors are described below. To identify multiple features for estimating the proportion of target cell sources, including the memory it holds, A computer system comprising the one or more programs: A) An instruction for obtaining an electronic training dataset, the tray The training dataset is for each of the multiple training targets: a) The correspondence of each cell-free fragment among the multiple cell-free fragments for corresponding training The methylation pattern, b) Including the signs of the target cancer of each of the training subjects, The corresponding methylation pattern of each cell-free fragment is (i) each training pair One or more nucleic acid samples containing each fragment from the corresponding biological sample obtained from an elephant. (ii) Determined by methylation sequencing of the following multiple in each of the above fragments It includes the methylation status of each CpG site among the CpG sites, The cancer state of the subject is either the first cancer state or the second cancer state. Commands and; B) Each cell-free fragment of each of the multiple cell-free fragments is selected from one of the multiple sequence groups. To map to groups and thereby obtain multiple training sets of self-free fragments The instruction is such that each of the plurality of sequence groups represents a corresponding portion of the human reference genome, Each training set of cell-free fragments is matched to a different sequence group from the aforementioned group of sequences. Ping, command and; C) Each of the training sets of the self-free fragments To assign the cancer status of each cell-free fragment to each cell-free fragment in the training set. The instruction is that the cancerous state of the cell-free fragment is the methylation of each cell-free fragment. The first cancer state is a function of the output of the classification device when a pattern is input to the classification device. and one of the conditions of the second cancer, command; D) For each of the plurality of sequence groups, (a) of the plurality of training targets (b) the state of the target cancer for each training target, and the pair mapped to each of the sequence groups. Each self-free fragment in the training set of corresponding self-free fragments is Instructions for determining the corresponding correlation scale I between states; E) Multiple features for estimating the proportion of the target cell source, sub-sequences of the multiple sequence groups An instruction for identifying as a set, wherein each sequence in the subset of the plurality of sequence groups The group satisfies the selection criteria based on the corresponding association measure for each of the aforementioned sequence groups. A computer system, including instructions.

90. When executed by the processor, the processor estimates the proportion of the target cell source. A non-uniform program code instruction that stores a method for identifying multiple features of a particular object. A time-sensitive computer-readable storage medium, wherein the method is: A) To obtain a training dataset in electronic format, and the training The dataset is for each training target of multiple training targets: a) The correspondence of each cell-free fragment among the multiple cell-free fragments for corresponding training The methylation pattern, b) Including the signs of the target cancer of each of the training subjects, The corresponding methylation pattern of each cell-free fragment is (i) each training pair One or more nucleic acid samples containing each fragment from the corresponding biological sample obtained from an elephant. (ii) Determined by methylation sequencing of the following multiple in each of the above fragments It includes the methylation status of each CpG site among the CpG sites, The cancer state of the subject is either the first cancer state or the second cancer state. thing; B) Each cell-free fragment of each of the multiple cell-free fragments is selected from one of the multiple sequence groups. We map to groups and thereby obtain multiple training sets of self-free fragments. Each of the plurality of sequence groups represents a corresponding portion of the human reference genome, and self Each training set of Lie fragments is mapped to a different set of sequences from among several sets of sequences. ru, koto to; C) Each of the training sets of the self-free fragments Assigning the cancer status of each cell-free fragment within the training set. The cancerous state of the cell-free fragment is determined by the methylation pattern of each cell-free fragment. The first cancer state and It is one of the conditions of the second type of cancer; D) For each of the plurality of sequence groups, (a) of the plurality of training targets (b) the state of the target cancer for each training target, and the pair mapped to each of the sequence groups. Each self-free fragment in the training set of corresponding self-free fragments is To determine the corresponding correlation scale I between the states; E) Multiple features for estimating the proportion of the target cell source, sub-sequences of the multiple sequence groups To specify as a set, each sequence group in the subset of the plurality of sequence groups is The selection criteria based on the corresponding correlation scale for each of the aforementioned sequence groups are met. Non-temporary computer-readable storage media, including [specific data / information].

91. A method for estimating the proportion of target cell sources: One or more processors and for execution by the one or more processors In a computer system having memory that stores one or more programs There, In electronic format, the corresponding methylation pattern of each cell-free fragment among multiple cell-free fragments The means to obtain the corresponding methylation pattern of each cell-free fragment is (i ) One or more nucleic acid samples containing each fragment from the biological sample obtained from the subject. (ii) Determined by methylation sequencing of the following multiple in each of the above fragments The CpG sites include the methylation state of each CpG site; Each of the aforementioned cell-free fragments is mated to one of the multiple sequence groups. The process involves pinging and thereby obtaining multiple sets of self-free fragments, and self-free Each set of fragments is mapped to a different set of sequences from the aforementioned set of sequences; Each self-free fragment in each set of self-free fragments in the plurality of sets of self-free fragments - Assigning the cancerous state of a cell-free fragment to a fragment, the cell-free fragment The state of the cancer is determined when the methylation patterns of each cell-free fragment are input into the classification device. This is a function of the output of the device, which is either the first cancer state or the second cancer state. Toto; In each set of cell-free fragments spanning the plurality of sequence groups, the first cancer state This involves calculating a first representative value for the number of self-free fragments from the assigned target; Cell-free fragments from the target in each set of cell-free fragments across the aforementioned group of sequences Calculate the second representative value of the number of pieces; The proportion of the target cell source is estimated using the first representative value and the second representative value. Includes, method.

92. Claim 91, wherein the plurality of sequence groups consist of 1,000 sequence groups to 100,000 sequence groups. Method of loading.

93. Claim 91, wherein the plurality of sequence groups consist of 15,000 to 80,000 sequence groups. Method of description.

94. Each of the plurality of sequence groups has, on average, 10 to 1200 residues, claim. The method described in any one of items 91 to 93.

95. Each of the aforementioned plurality of sequence groups has, on average, 10 to 10,000 residues, claim The method described in any one of paragraphs 91 to 93.

96. The first representative value is the first in each set of cell-free fragments across the plurality of sequence groups. The cancer status is determined by the arithmetic mean, weighted mean, and minute mean of the number of cell-free fragments from the assigned subjects. Midrange, mid-hinge, ternary mean, Windsor mean, mean, or mode, The method according to any one of claims 91 to 95.

97. The second representative value is the pair of cell-free fragments in each set across the plurality of sequence groups. Arithmetic mean, weighted mean, midrange, mid-hinge, ternary number of self-free fragments from an elephant Any one of claims 91 to 95, which is the mean, Windsor mean, average, or mode. The method described in section [section number].

98. Estimating the cell source proportion includes dividing the first representative value by the second representative value. The method according to any one of claims 91 to 97.

99. Claims 91 to 98, wherein the methylation sequencing is paired-end sequencing. The method described in any one of the items.

100. Claim 91 to the present invention, wherein the methylation sequencing is single-read sequencing. The method described in any one of paragraphs 98.

101. Each of the aforementioned cell-free fragments has an average length of less than 500 nucleotides. The method according to any one of claims 91 to 100.

102. If the first cancer state is cancer, and the second cancer state is the absence of cancer, The method described in any one of the requests 91 to 101.

103. The first type of cancer mentioned above includes adrenal cancer, biliary tract cancer, bladder cancer, bone / bone marrow cancer, brain cancer, and breast cancer. Cancer, cervical cancer, colorectal cancer, esophageal cancer, stomach cancer, head / neck cancer, hepatobiliary tract cancer, kidney cancer, Liver cancer, lung cancer, ovarian cancer, pancreatic cancer, pelvic cancer, pleural cancer, prostate cancer, kidney cancer, skin cancer Cancer, stomach cancer, testicular cancer, thymic cancer, thyroid cancer, uterine cancer, lymphoma, melanoma, multiple cancers It is one of the myelomas or leukemias. The second cancer state is the absence of cancer, according to any one of claims 91 to 102. Method of description.

104. The first cancer state mentioned above includes adrenal cancer at a certain stage, biliary tract cancer at a certain stage, and Stage bladder cancer, stage bone / bone marrow cancer, stage brain cancer, stage Breast cancer, cervical cancer at a certain stage, colon cancer at a certain stage, esophageal cancer at a certain stage Hmm, a certain stage of stomach cancer, a certain stage of head / neck cancer, a certain stage of hepatobiliary tract cancer, Kidney cancer at a certain stage, liver cancer at a certain stage, lung cancer at a certain stage Ovarian cancer, pancreatic cancer at a certain stage, pelvic cancer at a certain stage, pleural cancer at a certain stage , prostate cancer at a certain stage, kidney cancer at a certain stage, skin cancer at a certain stage, Stage 1 stomach cancer, stage 2 testicular cancer, stage 3 thymic cancer, stage 4 thymic cancer Adenocarcinoma, uterine cancer at a certain stage, lymphoma at a certain stage, melanoma at a certain stage, It is one of the conditions of multiple myeloma at a certain stage, or leukemia at a certain stage. The second cancer state is the absence of cancer, according to any one of claims 91 to 102. Method of description.

105. Claim 91, wherein the methylation sequencing is whole-genome methylation sequencing. Method of description.

106. The methylation sequencing is a targeted sequencing method using multiple nucleic acid probes. Furthermore, each of the plurality of sequence groups is at least one of the plurality of nucleic acid probes The method according to claim 91, which is associated with a corresponding nucleic acid probe.

107. The plurality of nucleic acid probes consist of 1,000 or more nucleic acid probes and 2,000 or more nuclei Acid probes, over 3,000 nucleic acid probes, over 5,000 nucleic acid probes, 10 More than 1,000 nucleic acid probes, or 1,000 to 30,000 nucleic acid probes The method according to claim 106, including the method described in claim 106.

108. Each of the aforementioned plurality of sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 , including 13, 14, 15, 16, 17, 18, 19, 20 or more CpG sites The method according to any one of claims 91 to 107.

109. Each of the aforementioned plurality of sequence groups is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 , 13, 14, 15, 16, 17, 18, 19, 20, or more consecutive CpG sections The method according to any one of claims 91 to 107, including the position.

110. Each of the aforementioned sequence groups is a sequence from 2 to 100 consecutive CpG sites in the human reference genome. The method according to any one of claims 91 to 107.

111. The biological sample is a liquid biological sample, according to any one of claims 91 to 110. Method of description.

112. The claim according to any one of claims 91 to 111, wherein the biological sample is a blood sample. The method.

113. The aforementioned biological sample is the subject's blood, whole blood, plasma, serum, urine, cerebrospinal fluid, feces, or saliva. , including sweat, tears, pleural fluid, pericardial fluid, or ascites, as described in any one of claims 91 to 111 Method of loading.

114. The aforementioned biological sample is the subject's blood, whole blood, plasma, serum, urine, cerebrospinal fluid, feces, or saliva. , consisting of sweat, tears, pleural fluid, pericardial fluid, or ascites, according to any one of claims 91 to 111 Method of description.

115. Of the corresponding multiple CpG sites in each of the aforementioned fragments, the methylation state of each CpG site is: If each of the aforementioned CpG sites is methylated, it is determined by the methylation sequencing. If this occurs, it is in a methylated state. If each of the aforementioned CpG sites is not methylated, it is determined by the methylation sequencing. If this occurs, it is in a non-methylated state. The methylation state of each CpG site is determined to be either methylated or unmethylated. If a call cannot be made through sequencing, it will be flagged as "Other" in the billing process. The method described in any one of paragraphs 91 to 114.

116. The methylation sequencing results in one or more 5-methyl groups in each of the fragments. Cytosine (5mC) and / or 5-hydroxymethylcytosine (5hmC) are detected. The method according to any one of claims 91 to 115.

117. The methylation sequencing involves one or multiple sequences in the sequence reads of each fragment. A number of unmethylated cytosines or one or more methylated cytosines, one corresponding The method described in any one of claims 91 to 116, which includes conversion to multiple uracils. method.

118. The one or more uracils are, as one or more corresponding thymines, the methi The method according to claim 117, which is detected during sequencing.

119. The aforementioned one or more unmethylated cytosines or one or more methylated cytosines The aforementioned transformations include chemical transformations, enzymatic transformations, or combinations thereof. The method described in claim 117.

120. The classification device used to assign the state of the cell-free fragment is the first cancer The first model for the state of the first cancer and the second model for the state of the second cancer include: The aforementioned first model is a first mixed model that includes a first plurality of submodels, The aforementioned second model is a second hybrid model that includes a second set of submodels, Each of the first and second submodels is the corresponding biological sample Represents an independent, corresponding methylation model for the source of the cell-free fragment in The method according to any one of claims 91 to 119.

121. The independent corresponding methylation models are the binomial model, the beta-binomial model, and the independent part. The method according to claim 120, wherein the method is one of a positional model or a Markov model.

122. Two or more of the first set of submodels described above are independent part models, Two or more of the second set of submodels described above are independent site models. The method according to claim 120.

123. Before mapping B), one or more filter conditions are applied to the plurality of cell-free fragments. The method according to any one of claims 91 to 122, further comprising applying the following:

124. One of the one or more filter conditions is the plurality of Apply a p-value threshold to the corresponding methylation pattern of each cell-free fragment among the cell-free fragments. In some cases, the p-value threshold is determined when the methylation pattern is observed in a cohort of non-cancer subjects. The method according to claim 123, which is representative of the frequency of

125. The method according to claim 124, wherein the p-value threshold is 0.001 to 0.

20.

126. The cohort includes at least 20 subjects, and the plurality of cell-free fragments are at least The method according to claim 124, comprising 10,000 different corresponding methylation patterns.

127. The p-value threshold is determined by the corresponding methyl group of each cell-free fragment among the plurality of cell-free fragments. If the transformation pattern has a p-value of 0.10 or less, 0.05 or less, or 0.01 or less, The method according to claim 124, which satisfies the methylation pattern from the aforementioned target.

128. One of the one or more filter conditions is the plurality of Each cell-free fragment contains one of the corresponding fragments in the biological sample. Alternatively, in the corresponding sequence reads measured from multiple nucleic acid samples, the threshold number The method according to claim 123, which applies the requirement that the sequence reads are represented. Law.

129. The threshold number is an integer from 2, 3, 4, 5, 6, 7, 8, 9, 10, or 10 to 100. A method according to claim 128.

130. One of the one or more filter conditions is the plurality of Each cell-free fragment contains one of the corresponding fragments in the biological sample. Alternatively, in multiple nucleic acid samples, the requirement is that it is represented by a threshold number of cell-free nucleic acids. The method according to claim 123, wherein the item is applied.

131. The threshold number is 2, 3, 4, 5, 6, 7, 8, 9, 10, or an integer from 10 to 100. A method according to claim 130.

132. One of the one or more filter conditions is the plurality of The requirement that each cell-free fragment of a cell-free fragment has a threshold number of CpG sites is applied. The method according to claim 123.

133. The threshold number of the CpG site is at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 The method according to claim 132, wherein the CpG site is a number of CpG sites.

134. One of the one or more filter conditions is the plurality of The requirement is that each cell-free fragment of the cell-free fragment has a length less than a threshold number of base pairs. The method according to claim 123.

135. The threshold number of base pairs has a length of 1,000, 2,000, 3,000, or 4,000. The method according to claim 134, wherein the base pairs are consecutive.

136. Based at least partially on the value of the cell source percentage of the subject, a treatment regimen is administered to the subject. The method according to any one of claims 91 to 135, further comprising applying it.

137. Claim 136, wherein the treatment regimen includes applying an anticancer drug to the subject. Method of loading.

138. Claim 1, wherein the cancer drug is a hormone, immunotherapy, X-ray examination, or anticancer agent. Method 37.

139. The aforementioned cancer drugs are lenalidomide, pembrolizumab, trastuzumab, and bevacizumab. , rituximab, ibrutinib, human papillomavirus tetravalent (6, 11, 16, and Type 18 vaccine, pertuzumab, pemetrexed, nilotinib, nilotinib, denosum B, abiraterone acetate, promacta, imatinib, everolimus, palbociclib, el Rotinib, bortezomib, bortezomib, or their generic equivalents, please The method described in item 137.

140. The aforementioned subject is being treated with cancer drugs, The method uses the proportion of the target cell source to determine the response to the target cancer drug. The method according to any one of claims 91 to 135, further comprising evaluation.

141. Claim 1, wherein the cancer drug is a hormone, immunotherapy, X-ray examination, or anticancer agent. Method 40.

142. The aforementioned cancer drugs are lenalidomide, pembrolizumab, trastuzumab, and bevacizumab. , rituximab, ibrutinib, human papillomavirus tetravalent (6, 11, 16, and Type 18 vaccine, pertuzumab, pemetrexed, nilotinib, nilotinib, denosum B, abiraterone acetate, promacta, imatinib, everolimus, palbociclib, el Rotinib, bortezomib, bortezomib, or their generic equivalents, please The method described in item 140.

143. The aforementioned subject is being treated with cancer drugs, The method uses the proportion of the target cell source to increase the amount of the cancer drug in the target. Any one of claims 91 to 135 further includes deciding whether to intensify or suspend. Methods used.

144. The subject has undergone surgical intervention to address the cancer, The method uses the proportion of the target cell source to determine the target's response to the surgical intervention. The method according to any one of claims 91 to 135, further comprising evaluating the state.

145. The acquisition, mapping, assignment, calculation of the first and second representative values, and the epoch The estimation of the proportion of the target cell source at each of the multiple time points is repeated, and therefore Then, for the aforementioned target at each point in time, the corresponding cell source proportion is obtained from among multiple cell source proportions. to do; Using the aforementioned multiple cell source proportions, increase the proportion of the first cell source over the epoch or In the form of a decrease, the state or progression of the target disease condition during the aforementioned epoch is determined. The method according to any one of claims 91 to 144, including the method described herein.

146. The aforementioned epoch is a period of several months, and each of the aforementioned multiple time points is within the period of several months. The method according to claim 145, which is a different point in time within the interval.

147. The method according to claim 146, wherein the aforementioned period of several months is less than four months.

148. The aforementioned epoch is a period of several years, and each of the aforementioned multiple points in time falls within the aforementioned period of several years. The method according to claim 145, which is at a different point in time.

149. The method according to claim 148, wherein the aforementioned period of several years is 2 to 10 years.

150. The aforementioned epoch is a period of several hours, and each of the multiple time points is within the period of several hours. The method according to claim 145, which is a different point in time within the interval.

151. The method according to claim 150, wherein the aforementioned period of several hours is 1 to 6 hours.

152. It was observed that the proportion of the first cell source of the aforementioned target changed by a threshold amount over the aforementioned epoch. The method according to claim 145, further comprising changing the diagnosis of the subject in such cases.

153. It was observed that the proportion of the first cell source of the aforementioned target changed by a threshold amount over the aforementioned epoch. The method according to claim 145, further comprising changing the prognosis of the subject in such cases.

154. It was observed that the proportion of the first cell source of the aforementioned target changed by a threshold amount over the aforementioned epoch. The method according to claim 145, further comprising changing the treatment of the subject in such cases.

155. The aforementioned thresholds are greater than 10%, greater than 20%, greater than 30%, greater than 40%, greater than 50%, greater than 2 times, greater than 3 times, The method according to claim 152, 153, or 154, wherein the ratio is more than 5 times.

156. The method according to any one of claims 1 to 155, wherein the cell source proportion is the tumor proportion.

157. The method according to claim 156, wherein the tumor ratio is 0.003 to 1.

0.

158. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. Genome regions listed in one or more of Lists 1-8 in Issue 2020 / 069350 The method according to any one of claims 91 to 157, corresponding to the method described in claim 91 to 157.

159. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. Genome regions listed in one or more of Lists 1-8 in Issue 2020 / 069350 The following is mapped to at least 30% of claims 91 to 158 method.

160. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. Genome regions listed in one or more of Lists 1-8 in Issue 2020 / 069350 The following is mapped to at least 50-95% of any one of claims 91 to 159 Method of description.

161. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. One to ten unique in one or more of Lists 1 to 8 of Issue 2020 / 069350 Mapped to the corresponding genomic region, as described in any one of claims 91 to 160. The method.

162. The sequence groups in the aforementioned plurality of sequence groups are as shown in Tables 1-2 of International Publication No. 2019 / 195268.

4. Lists 1-16 of International Publication No. 2020 / 154682, and / or International Publication No. A single, unique correspondence in one or more of Lists 1-8 of Issue 2020 / 069350 The method according to any one of claims 91 to 161, which is mapped to a genomic region. 。

163. With respect to the aforementioned target, the plurality of cell-free fragments are at least 100,000 cells The method according to any one of claims 91 to 162, comprising a free fragment.

164. With respect to the aforementioned target, the plurality of cell-free fragments are at least 500,000 cells The method according to any one of claims 91 to 162, comprising a free fragment.

165. Regarding the above subject, the plurality of cell-free fragments are at least 1 million cell-free The method according to any one of claims 91 to 162, including a fragment.

166. Each of the above-mentioned sequence groups contains fewer than 100 nucleic acid residues, fewer than 500 nucleic acid residues, Nucleic acid residues less than 1000, less than 2500, less than 5000, 1 Nucleic acid residues less than 0,000, nucleic acid residues less than 25,000, nuclei less than 50,000 Acid residues, fewer than 100,000 nucleic acid residues, fewer than 250,000 nucleic acid residues, or 5 The following is a description of any one of claims 91 to 165, consisting of fewer than 00,000 nucleic acid residues. method.

167. One or more processors, One or more programs for execution by the aforementioned one or more processors are described below. A computer system for estimating the proportion of target cell sources, including the memory it holds. There is one or more of the above programs, In electronic format, the corresponding methylation pattern of each cell-free fragment among multiple cell-free fragments An instruction for obtaining a n, wherein the corresponding methylation pattern of each cell-free fragment is (i) one or more nucleic acid samples containing each fragment in the biological sample obtained from the subject Determined by methylation sequencing of the sample, (ii) the corresponding in each of the above fragments The instruction includes the methylation state of each of the multiple CpG sites; Each of the aforementioned cell-free fragments is mated to one of the multiple sequence groups. An instruction to ping and thereby obtain multiple sets of cell-free fragments, Each set of free fragments is mapped to a different set of sequences from the aforementioned set of sequences, Order and; Each self-free fragment in each set of self-free fragments in the plurality of sets of self-free fragments - Instructions for assigning the cancer state of a self-free fragment to a fragment, the self-free fragment - The cancerous state of the fragments was determined by inputting the methylation pattern of each cell-free fragment into a classification device. The output of the classification device is a function of the first cancer state and the second cancer state. to, command; In each set of cell-free fragments spanning the plurality of sequence groups, the first cancer state The instruction is to calculate a first representative value of the number of self-free fragments from the assigned target; Cell-free fragments from the target in each set of cell-free fragments across the aforementioned group of sequences Instructions for calculating the second representative value of the number of pieces; Using the first representative value and the second representative value, the life to estimate the proportion of the target cell source A computer system that includes the letter "Rei".

168. When executed by the processor, the processor estimates the proportion of the target cell source. on a non-temporary computer-readable storage medium that stores program code instructions to execute the law. And the aforementioned method is: In electronic format, the corresponding methylation pattern of each cell-free fragment among multiple cell-free fragments The means to obtain the corresponding methylation pattern of each cell-free fragment is (i ) One or more nucleic acid samples containing each fragment from the biological sample obtained from the subject. (ii) Determined by methylation sequencing of the following multiple in each of the above fragments The CpG sites include the methylation state of each CpG site; Each of the aforementioned cell-free fragments is mated to one of the multiple sequence groups. The process involves pinging and thereby obtaining multiple sets of self-free fragments, and self-free Each set of fragments is mapped to a different set of sequences from the aforementioned set of sequences; Each self-free fragment in each set of self-free fragments in the plurality of sets of self-free fragments - Assigning the cancerous state of a cell-free fragment to a fragment, the cell-free fragment The state of the cancer is determined when the methylation patterns of each cell-free fragment are input into the classification device. This is a function of the output of the device, which is either the first cancer state or the second cancer state. Toto; In each set of cell-free fragments spanning the plurality of sequence groups, the first cancer state This involves calculating a first representative value for the number of self-free fragments from the assigned target; Cell-free fragments from the target in each set of cell-free fragments across the aforementioned group of sequences Calculate the second representative value of the number of pieces; The proportion of the target cell source is estimated using the first representative value and the second representative value. Includes non-temporary computer-readable storage media.

Images