Epigenetic analysis of cell-free DNA

Abstract

Measurement of the amount (e.g., relative frequency) of specific sequence motifs of cell-free DNA fragments in a biological sample can be used to analyze the biological sample. Specific sequence motifs or sequence lengths in certain genomic regions may indicate histone modifications. Sequence motifs and / or lengths may be used to measure a property of a sample (e.g., a concentration fraction of tissue types or characteristics of tissue types), to measure the amount of histone modifications, to determine a condition of an organism based on such measurements, and to enrich clinically relevant DNA of a biological sample. Different tissue types may exhibit different patterns for the relative frequency of the sequence motif. Measurement of the relative frequency of a sequence motif of cell-free DNA may be used for analysis.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to and is a non-provisional application of U.S. Provisional Application No. 63 / 393,725, filed on July 29, 2022, entitled “Epigenetic Analysis of Cell-Free DNA,” the disclosure of which is incorporated by reference in its entirety for all purposes. Background Art

[0002] Cell-free DNA (cfDNA) is a rich source of information that can be used to diagnose and predict many physiological and pathological conditions, such as pregnancy and cancer (Chan, KCA et al. (2017), New England Journal of Medicine 377, 513-522; Chiu, RWK et al. (2008), Proceedings of the National Academy of Sciences of the United States of America 105, 20458-20463; Lo, YMD et al., (1997), The Lancet 350, 485-487). Cell-free DNA molecules in various body fluids (e.g., plasma, serum, urine, saliva, semen, peritoneal fluid, cerebrospinal fluid) may include a mixture of DNA molecules derived from various tissues. One mechanism for releasing such cfDNA molecules is through cell death (e.g., apoptosis or necrosis). Selected cell populations, such as lymphocytes and neutrophils, have also been shown to secrete DNA molecules into body fluids. cfDNA molecules consist of fragmented DNA molecules. The correlation between cfDNA fragmentation pattern and nucleosome structure has been demonstrated in many studies (Sun et al. Proc Natl Acad Sci USA. 2018; 115: E5106; Snyder et al. Cell. 2016; 164: 57-68). Although circulating cfDNA is now commonly used as a non-invasive biomarker and is known to circulate in the form of short fragments, the physiological factors that control the fragmentation and molecular distribution of cfDNA remain unclear.

[0003] Cell-free DNA can be analyzed to understand the epigenomic state. The epigenomic state of DNA can indicate the regulation of genes, tissue origin or disease. The amount of histone modification is an epigenomic factor. Conventional techniques for detecting histone modifications include the use of specific antibodies, relatively large amounts of samples and more complex sample processing. A simpler and more efficient technique is needed to determine the epigenomic state of DNA. These needs and other needs are addressed. Summary of the invention

[0004] The present disclosure describes various techniques, such as measuring the amount (e.g., relative frequency) of sequence motifs and the length of cell-free DNA fragments in a biological sample of an organism, for measuring a property of a sample (e.g., a concentration fraction of a tissue type or a characteristic of a tissue type), measuring the amount of histone modifications, determining the condition of an organism based on such measurements, and enriching a biological sample for clinically relevant DNA. Different tissue types exhibit different chromatin structural patterns. The present disclosure provides various uses for inferring chromatin structure based on, for example, measurements of the relative frequency of sequence motifs in a mixture of cell-free DNA from various tissues and / or the length of the cell-free DNA. DNA from one of the specific tissues may be referred to as clinically relevant DNA.

[0005] Various examples can quantitatively represent the amount of sequence motifs of the terminal sequences (i.e., terminal motifs) of DNA fragments. For example, embodiments can determine one or more relative frequencies of a set of one or more sequence motifs for the terminal sequences of DNA fragments. In various embodiments, a preferred set of terminal motifs can be determined by using another technique (e.g., cfChIP-seq [cell-free chromatin immunoprecipitation sequencing) to measure the epigenomic state (e.g., histone modification) of chromatin in a specific region of the object. A preferred set of terminal motifs can be selected based on more frequent occurrences in one or more regions with a specific epigenomic state compared to other terminal motifs. A specific epigenomic state can be related to a specific tissue type or clinically relevant DNA.

[0006] In various embodiments, the relative frequencies of the preferred set can be used to measure the classification of properties of new samples (e.g., the concentration fraction of clinically relevant DNA), the condition of an organism (e.g., the gestational age or pathology level of a fetus), or the measurement of epigenomic state (e.g., the amount of histone modification). Thus, embodiments can provide measurements to inform physiological changes, including cancer, autoimmune diseases, transplantation, and pregnancy.

[0007] As a further example, a set of preferred sequence end motifs can be used for physical enrichment and / or computer simulation enrichment of clinically relevant cell-free DNA fragments of biological samples. Enrichment can be based on preferred end motifs of one or more genomic regions with specific histone modifications. For certain clinically relevant tissues, such as fetuses, tumors or transplants, specific histone modifications in one or more genomic regions can be preferred. Physical enrichment can use one or more probe molecules that detect a specific set of sequence end motifs so that biological samples are enriched for clinically relevant DNA fragments. For computer simulation enrichment, a set of sequence readings of cell-free DNA fragments can be identified, and the cell-free DNA fragments have one of the preferred end sequences of clinically relevant DNA. Certain sequence readings can be stored based on the possibility corresponding to clinically relevant DNA, wherein the possibility explains the sequence readings including the preferred sequence end motifs. The stored sequence readings can be analyzed to determine the properties of the clinically relevant DNA of the biological sample.

[0008] In some embodiments, the amount of DNA fragments within a certain length range can be used to determine the amount of histone modifications in cell-free DNA. The amount of histone modifications inferred by the length information can be used to determine tissue scores, classification of disease grade, and status of tissue or organ transplantation.

[0009] In addition, although histone modifications in specific genomic regions can indicate DNA of a specific type of tissue, histone modifications in many genomic regions can be the result of several different tissues. Using histone modifications in genomic regions contributed by several different tissues can allow for more accurate analysis of biological samples than using only histone modifications in genomic regions produced by a single tissue. For example, using histone modifications contributed by several different tissues can lead to a more accurate analysis of tissue origin and disease grade.

[0010] These and other embodiments of the present disclosure are described in detail below.For example, other embodiments relate to systems, devices, and computer-readable media associated with the methods described herein.

[0011] Reference to the remainder of the specification, including the drawings and claims, will recognize other features and advantages of the present disclosure. Other features and advantages of the present disclosure, as well as the structure and operation of various embodiments of the present disclosure, will be described in detail below with reference to the drawings. In the drawings, the same reference numerals may represent the same or functionally similar components. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 An icon showing the structure of DNA.

[0013] Figure 2The use of immunoprecipitation to analyze plasma cfDNA molecules associated with histone modifications is shown.

[0014] Figure 3 A schematic representation of the terminal motifs of the fragments is shown.

[0015] Figure 4 is a diagram defining categories of H3K4me3 regions with different H3K4me3 ChIP signal levels according to an embodiment of the present invention.

[0016] Figure 5 is a table showing exemplary definitions of H3K4me3 region categories for H3K4me3 ChIP-seq analysis using pregnancy samples according to embodiments of the present invention.

[0017] Figure 6 A table showing exemplary definitions of H3K27ac region categories using H3K27ac ChIP-seq analysis of pregnancy samples according to embodiments of the present invention.

[0018] Figure 7 is a table showing exemplary definitions of H3K4me3 region categories using H3K4me3 ChIP-seq analysis of samples from non-pregnant healthy subjects according to embodiments of the present invention.

[0019] Figure 8 is a table showing exemplary definitions of H3K27ac region categories using H3K27ac ChIP-seq analysis of samples from non-pregnant healthy subjects according to embodiments of the present invention.

[0020] Fig. 9 Shown is a heat map of motif frequencies in regions with different levels of H3K4me3 ChIP signals for plasma DNA sequencing results with the immunoprecipitation step eliminated, according to an embodiment of the present invention.

[0021] Fig.10 is a comparative graph of terminal motif frequency rankings between plasma DNA sequencing results with or without H3K4me3-based immunoprecipitation according to an embodiment of the present invention.

[0022] Fig.11 Shown is a table of the 24 terminal motifs with the greatest ranking differences between conventional cfDNA sequencing and cfChIP-seq for H3K4me3 histone modification according to embodiments of the present invention.

[0023] Fig. 12A and Fig. 12BIt is described that according to an embodiment of the present invention, the end-motif pattern is used to infer plasma DNA histone modification signals for plasma DNA sequencing results in the absence of immunoprecipitation.

[0024] Fig.13 Shown is a graph of the correlation between the aggregate abundance of over-represented terminal motifs in H3K4me3-based immunoprecipitated plasma DNA and the H3K4me3 ChIP signal, according to an embodiment of the present invention.

[0025] Fig.14 is a graph showing the correlation between cfChIP signals and terminal motif frequencies for 11 peak groups according to an embodiment of the present invention.

[0026] Fig.15A and 15B is a graph showing the correlation between cfChIP signals and terminal motif frequencies for six and eight peak groups, according to an embodiment of the present invention.

[0027] Fig.16 is a graph of the correlation between H3K4me3 ChIP signals in placenta-specific H3K4me3 regions inferred by end motifs and fetal DNA fraction determined by a SNP-based approach, according to an embodiment of the present invention.

[0028] Fig.17 is a flow chart of an exemplary process associated with determining the fractional concentration of cell-free DNA fragments in a biological sample according to an embodiment of the present invention.

[0029] Fig.18 is a flow chart of an exemplary process associated with estimating a first value of a characteristic of a target tissue, according to an embodiment of the present invention.

[0030] Fig.19 is a flow chart of an exemplary process associated with using sequence motifs to determine the amount of histone modifications in one or more genomic regions according to embodiments of the present invention.

[0031] Fig. 20 is a flow chart of an exemplary process associated with determining the amount of histone modifications in one or more genomic regions using fragmentomic properties according to embodiments of the present invention.

[0032] Fig.21 It is shown that according to an embodiment of the present invention, the contribution from different tissues is determined using ChIP-seq (cell-free chromatin immunoprecipitation followed by sequencing).

[0033] Fig. 22] is a ROC curve for distinguishing patients with and without HCC using inferred H3K4me3 signals in liver-specific H3K4me3 regions using end motifs according to an embodiment of the present invention.

[0034] Fig.23 is a flow chart of an exemplary process associated with classifying a level of a condition according to an embodiment of the present invention.

[0035] Fig.24A , 24B and 24C show the percentage of cfDNA molecules of a particular length in categories of regions with different levels of H3K27ac signal, according to an embodiment of the present invention.

[0036] Fig.25A , 25B and 25C show that the correlation between length and ChIP signal of a histone modification can be generalized to other histone modifications according to embodiments of the present invention.

[0037] Fig.26A and Fig.26B It is described that according to an embodiment of the present invention, length information is used to infer plasma DNA histone modifications for plasma DNA sequencing results without immunoprecipitation.

[0038] Fig.27A , 27B and 27C show the correlation between the percentage of cfDNA molecules within a length range and the log-transformed H3K4me3 ChIP signal, according to an embodiment of the present invention.

[0039] Fig.28A Shown is an evaluation of the performance of inferred H3K4me3 ChIP signals in placenta-specific H3K4me3 regions for fetal DNA fraction inference according to embodiments of the present invention.

[0040] Fig.28B Shown is the performance of evaluating molecules within a certain length range in placenta-specific H3K4me3 regions for fetal DNA fraction inference, according to embodiments of the present invention.

[0041] Fig.29 is a graph evaluating the performance of inferred H3K27ac ChIP signals in placenta-specific H3K27ac regions for determining fetal DNA fraction, according to embodiments of the present invention.

[0042] Fig.30is a graph of the Pearson correlation coefficient of the amount of H3K27ac signal inferred from the DNA fragment length distribution in placenta-specific regions and the fragment length distribution itself with the fetal DNA fraction determined by a SNP-based method, according to an embodiment of the present invention.

[0043] Fig.31A and Fig.31B is a graph showing the use of inferred H3K4me3 ChIP signals based on liver-specific H3K4me3 regions for HCC detection according to an embodiment of the present invention.

[0044] Fig.32A and Fig.32B Shown is the use of inferred H3K27ac ChIP signals based on H3K27ac regions for HCC detection according to an embodiment of the present invention.

[0045] Fig.33 is a receiver operating characteristic (ROC) curve for distinguishing subjects with advanced hepatocellular carcinoma from healthy subjects according to an embodiment of the present invention.

[0046] Fig.34 is a graph showing the correlation between the inferred H3K27ac ChIP signal and the donor DNA fraction in the liver-specific H3K27ac region according to an embodiment of the present invention.

[0047] Fig.35 is a graph of the Pearson correlation coefficient of the amount of H3K27ac signal inferred from the DNA fragment length distribution in liver-specific regions and the fragment length distribution itself with the fetal DNA fraction determined by a SNP-based method, according to an embodiment of the present invention.

[0048] Fig.36 is a flow chart of an exemplary process associated with using fragment length to determine the amount of a histone modification in one or more genomic regions, according to an embodiment of the present invention.

[0049] Fig.37 A table showing tissue-specific histone modification regions according to embodiments of the present invention is shown.

[0050] Fig.38 is a graph showing plasma DNA tissue mapping of H3K4me3 histone modification based on cell-free DNA according to an embodiment of the present invention.

[0051] Fig.39 Shown is the correlation between the placental contribution inferred by the H3K4me3 ChIP signal according to an embodiment of the present invention and the fetal DNA fraction according to an embodiment of the present invention.

[0052] Fig.40is a graph of the percentage contribution of different tissues of pregnant and non-pregnant samples based on H3K27ac histone modification in cfDNA according to an embodiment of the present invention.

[0053] Fig.41 Shown is a heat map of tissue contributions inferred from H3K27ac ChIP signals in pregnant and non-pregnant subjects according to embodiments of the present invention.

[0054] Fig.42 is a box plot of the inferred H3K27ac ChIP signals at various tissue-specific regions according to an embodiment of the present invention.

[0055] Fig.43A Shown is the correlation between the placental contribution inferred from the H3K27ac ChIP signal and the fetal DNA fraction determined by the SNP-based method according to embodiments of the present invention.

[0056] Fig.43B Shown is the correlation between normalized reads / kb in placenta-specific regions and the fetal DNA fraction determined by a SNP-based method, according to embodiments of the present invention.

[0057] Fig.44 is a ROC curve for distinguishing between pregnant and non-pregnant subjects according to an embodiment of the present invention.

[0058] Fig.45 Shown is a receiver operating characteristic (ROC) curve for distinguishing between control subjects and colorectal cancer (CRC) subjects using inferred colon contribution, according to an embodiment of the present invention.

[0059] Fig.46A is a graph comparing erythrocyte blast contribution inferred from H3K27ac ChIP signals between subjects with β-thalassemia major and control subjects without β-thalassemia major, according to embodiments of the present invention.

[0060] Fig.46B is a ROC curve for distinguishing subjects with and without β-thalassemia major using inferred erythrocyte blast contribution according to an embodiment of the present invention.

[0061] Fig.47 is a heat map of tissue contributions inferred using H3K27ac ChIP signals in subjects with β-thalassemia major and control subjects, according to an embodiment of the present invention.

[0062] Fig.48A , 48Band 48C show the correlation between the percentage of erythroid DNA determined by ddPCR assay and the erythroid blast contribution determined by H3K27ac signal, according to an embodiment of the present invention.

[0063] Fig.49A and 49B is a graph of inferred H3K27ac signals in healthy subjects, subjects with colorectal cancer (CRC) but no liver metastasis, and subjects with CRC and liver metastasis, according to an embodiment of the present invention.

[0064] Fig.50 is a graph of tissue contributions in urine and plasma DNA samples using H3K27ac histone modification of cell-free DNA, according to an embodiment of the present invention.

[0065] Fig.51 is a flow chart of an exemplary process associated with determining a concentration score for a tissue type, according to an embodiment of the present invention.

[0066] Fig.52 is a flow chart of an exemplary process associated with determining a classification of a pregnancy or disease, according to an embodiment of the present invention.

[0067] Fig.53 Input features in a machine learning model for determining cancer classification according to an embodiment of the present invention are shown.

[0068] Fig.54A and 54B Results of a machine learning model for determining cancer classification according to an embodiment of the present invention are shown.

[0069] Fig.55 Shown are the area under the curve (AUC) results for distinguishing hepatocellular carcinoma (HCC) and non-HCC cases using machine learning models with different fragmentomic properties according to an embodiment of the present invention.

[0070] Fig.56 is a flow chart of an exemplary process associated with analyzing a biological sample of a subject to determine a classification of the subject's condition, according to an embodiment of the present invention.

[0071] Fig.57 is a flow chart of an exemplary process associated with enriching a biological sample for clinically relevant DNA, according to an embodiment of the present invention.

[0072] Fig.58 is a flow chart of an exemplary process associated with enriching a biological sample for clinically relevant DNA, according to an embodiment of the present invention.

[0073] Fig.59 A measurement system according to an embodiment of the invention is shown.

[0074] Fig.60 A block diagram of an exemplary computer system for use with systems and methods according to embodiments of the present invention is shown. the term

[0075] "Tissue" corresponds to a group of cells that come together as a functional unit. More than one type of cell can be found in a single tissue. Different types of tissues may consist of different types of cells (e.g. liver cells, alveolar cells or blood cells), but may also correspond to tissues from different organisms (mother vs. fetus) or to healthy cells vs. tumor cells.

[0075] A "biological sample" refers to any sample taken from a subject (e.g., a human (or other animal), such as a pregnant woman, a person with cancer, or a person suspected of having cancer, an organ transplant recipient, or a subject suspected of having a disease process involving an organ (e.g., the heart in myocardial infarction, or the brain in stroke, or the hematopoietic system in anemia), and containing one or more nucleic acid molecules of interest. The biological sample can be a body fluid, such as blood, plasma, serum, urine, vaginal fluid, fluid from the scrotum (e.g., ascites from the testicles), vaginal washings, pleural fluid, ascites, cerebrospinal fluid, saliva, sweat, tears, sputum, bronchoalveolar lavage fluid, discharge from the nipple, aspirated fluid from different parts of the body (e.g., thyroid, breast), intraocular fluid (e.g., aqueous humor), etc. Stool samples can also be used. In various In one embodiment, most of the DNA in a biological sample (e.g., a plasma sample obtained by a centrifugation protocol) that has been enriched for cell-free DNA can be cell-free, for example, greater than 50%, 60%, 70%, 80%, 90%, 95% or 99% of the DNA can be cell-free. The centrifugation protocol can include, for example, 3,000 g × 10 minutes, obtaining a liquid portion, and centrifuging again for 10 minutes at, for example, 16,000 g to remove residual cells. As part of the biological sample analysis, at least 1,000 cell-free DNA molecules can be analyzed. As other examples, at least 10,000 or 50,000 or 100,000 or 500,000 or 1,000,000 or 5,000,000 cell-free DNA molecules or more can be analyzed.

[0076] "Clinically relevant DNA" may refer to DNA of a specific tissue source to be measured, for example, to determine the concentration fraction of such DNA or the phenotype of a classified sample (e.g., plasma). Examples of clinically relevant DNA are fetal DNA in maternal plasma or tumor DNA in patient plasma or other samples with cell-free DNA. Another example includes measuring the amount of transplant-related DNA in the plasma, serum or urine of a transplant patient. Another example includes measuring the concentration fraction of hematopoietic and non-hematopoietic DNA in the plasma of the subject, or the concentration fraction of liver DNA fragments (or other tissues) in the sample, or the concentration fraction of brain DNA fragments in cerebrospinal fluid.

[0077] "Sequence read" refers to a string of nucleotides sequenced from any part or all of a nucleic acid molecule. For example, a sequence read can be a short string of nucleotides (e.g., 20-150 nucleotides) sequenced from a nucleic acid fragment, a short string of nucleotides at one or both ends of a nucleic acid fragment, or the sequencing of an entire nucleic acid fragment present in a biological sample. Sequence reads can be obtained in a variety of ways, such as using sequencing techniques or using probes, such as in hybridization arrays or capture probes, or amplification techniques, such as polymerase chain reaction (PCR) or linear amplification or isothermal amplification using a single primer. As part of the analysis of a biological sample, at least 1,000 sequence reads can be analyzed. As other examples, at least 10,000 or 50,000 or 100,000 or 500,000 or 1,000,000 or 5,000,000 sequence reads or more can be analyzed.

[0078] The sequence read may include an "end sequence" associated with the end of the fragment. The end sequence may correspond to the outermost N bases of the fragment, such as 2-30 bases at the end of the fragment. If the sequence read corresponds to the entire fragment, the sequence read may include two end sequences. When paired-end sequencing provides two sequence reads corresponding to the ends of the fragment, each sequence read may include one end sequence.

[0079] The "sequence motif" of a "sequence end tag" may refer to a short repeating pattern of bases in a nucleic acid fragment (e.g., a cell-free DNA fragment). The sequence motif may occur at the end of a fragment and thus be part of or include a terminal sequence. A "terminal motif" may refer to a sequence motif that preferentially occurs at the terminal sequence of a nucleic acid (e.g., a DNA fragment) that may be a particular type of tissue. The terminal motif may also occur before or after the end of a fragment, thereby still corresponding to a terminal sequence.

[0080] The term "fractional fetal DNA concentration" is used interchangeably with the terms "fetal DNA ratio" and "fetal DNA fraction" and refers to the proportion of fetal DNA molecules present in a biological sample derived from the fetus, such as a maternal plasma or serum sample (Lo et al, Am J Hum Genet. 1998; 62: 768-775; Lun et al, Clin Chem. 2008; 54: 1664-1672).

[0081] "Relative frequency" (also simply "frequency") can refer to a ratio (e.g., a percentage, a fraction, or a concentration). In particular, the relative frequency of a particular end motif pair (e.g., A<>A) can provide the ratio of cell-free DNA fragments having the particular end sequence pair.

[0082] An "aggregate value" can refer to, for example, a collective property of the relative frequencies of a set of terminal motifs. Examples include the mean, median, sum of relative frequencies, the variation between relative frequencies (e.g., entropy, standard deviation (SD), coefficient of variation (CV), interquartile range (IQR) or a percentile cutoff between different relative frequencies (e.g., 95th or 99th percentile), or the difference (e.g., distance) from a reference pattern of relative frequencies, as can be implemented in clustering. As another example, an aggregate value can contain an array / vector of relative frequencies, which can be compared to a reference vector (e.g., representing a multidimensional data point).

[0083] "Calibration sample" can correspond to a biological sample, the concentration fraction of which is clinically relevant nucleic acid (e.g., tissue-specific DNA portion) is known or measured by a calibration method, e.g., using tissue-specific alleles, e.g., in transplantation, whereby alleles present in the donor genome but not in the recipient genome can be used as markers for transplanted organs. As another example, a calibration sample can correspond to a sample from which an end motif can be determined. Calibration samples can be used for two purposes. Multiple calibration samples can be used. As an example, a first calibration sample can correspond to a biological sample, which has a measurable histone modification level in various target genomic regions. A second calibration sample can correspond to a biological sample, which has a measurable fragment omics property in various target genomic regions. The first and second calibration sampling can be used together to determine calibration values.

[0084] "Calibration data points" include "calibration values" and measured or known characteristic values ​​of the concentration fraction of target tissue types or clinically relevant nucleic acids (e.g., DNA of a specific tissue type). Calibration values ​​can be determined by various types of data measured from the nucleic acid molecules of the sample, such as the amount of terminal motifs or fragment lengths. Calibration values ​​correspond to parameters associated with desired properties, such as characteristic values ​​of target tissue types or concentration fractions of clinically relevant DNA. For example, calibration values ​​can be determined from the relative frequency (e.g., total value) of the terminal labels determined for the calibration sample, for which the desired properties are known. Calibration data points can be defined in a variety of ways, for example, as discrete points or as calibration functions (also referred to as calibration curves or calibration surfaces). Calibration functions can be derived from additional mathematical transformations of calibration data points. In some embodiments, "calibration data points" can include "calibration values" and a set of measured or known characteristic values ​​(e.g., fragment omics properties) of a genomic region of interest (e.g., characterized by certain levels of histone modifications).

[0085] A "separation value" corresponds to a difference or ratio involving two values ​​(e.g., two fractional contributions or two methylation levels). A separation value can be a simple difference or ratio. As an example, the direct ratio of x / y is a separation value, as is x / (x+y). A separation value can include other factors, such as a multiplication factor. As another example, a difference or ratio of a function of values ​​can be used, such as the difference or ratio of the natural logarithms (ln) of two values. A separation value can include a difference or ratio.

[0086] "Separation values" and "aggregate values" (e.g., of relative frequencies) are two examples of parameters (also called metrics) that provide a measure of the sample's variation between different classifications (states), and can therefore be used to determine the different classifications. Aggregate values ​​can be separation values, for example, when taking a difference between a set of relative frequencies for a sample and a relative frequency for a reference group, which can be done in clustering.

[0087] The term "classification" as used herein refers to any number or other character associated with a particular property of a sample. For example, a "+" symbol (or the word "positive") can indicate that a sample is classified as having a deletion or an amplification. The classification can be binary (e.g., positive or negative) or have more classification levels (e.g., a scale from 1 to 10 or 0 to 1). As a further example, the classification level can correspond to, for example, a concentration score or a characteristic value of a sample or target tissue type.

[0088] The term "parameter" as used herein refers to a numerical value that characterizes a quantitative data set and / or a numerical relationship between quantitative data sets. For example, the ratio (or a function of the ratio) between a first amount of a first nucleic acid sequence and a second amount of a second nucleic acid sequence is a parameter.

[0089] The terms "cutoff value" and "threshold value" refer to predetermined numbers used in operations. For example, a cutoff length may refer to a length above which fragments are excluded. A threshold value may be a value above or below which a specific classification is applied. Any of these terms may be used in any of these contexts. A cutoff value or threshold value may be a "reference value" or derived from a reference value representing a specific classification or distinguishing between two or more classifications. Such a reference value may be determined in various ways, as will be understood by those skilled in the art. For example, a metric (parameter) may be determined for two different groups of objects with different known classifications, and a reference value may be selected as a representative of a classification (e.g., an average value) or a value between two clusters of a metric (e.g., selected to obtain a desired sensitivity and specificity). As another example, a reference value may be determined based on a statistical simulation of a sample. Specific values ​​of a cutoff value, threshold value, reference value, etc. may be determined based on a desired accuracy (e.g., sensitivity and specificity). A parameter may be compared to a cutoff value, threshold value, reference value, or calibration value to determine a classification. Such a process for determining such a value may be performed as part of training a machine learning model, for example, the machine learning model receives a training vector of a set of one or more parameters. And comparison of the parameter to any of these values ​​can be achieved by inputting the parameter into a machine learning model, which is trained, for example, using parameter values ​​determined from other subjects, such as subjects with or without a condition, abnormality or pathology or subjects with known parameter values ​​(e.g., calibrated values).

[0090] The term "cancer grade" may refer to whether cancer is present (i.e., present or absent), the stage of cancer, the length of a tumor, whether metastasis exists, the total tumor burden of the body, the response of cancer to treatment, and / or other measurements of the severity of cancer (e.g., recurrence of cancer). The cancer grade may be a number or other markers, such as symbols, letters, and colors. The grade may be zero. The cancer grade may also include a pre-malignant condition (status). The cancer grade may be used in a variety of ways. For example, screening may check whether cancer exists in a person who was previously not known to have cancer. An assessment may be conducted to investigate a person who has been diagnosed with cancer to monitor the progression of cancer over time, to study the effectiveness of treatment, or to determine a prognosis. In one embodiment, a prognosis may be expressed as the chance of a patient dying of cancer, or the chance of cancer progressing after a specific duration or time, or the chance or extent of cancer metastasis. Detection may refer to 'screening' or may refer to checking whether someone with suggestive features of cancer (e.g., symptoms or other positive tests) has cancer. "Disease grade" is similar to "cancer grade", but may refer to a disease rather than cancer.

[0091] "Abnormal grade" may refer to an abnormal amount, degree or severity associated with an organism, wherein the grade may be as described above for cancer. An abnormal example is a pathology associated with an organism. Another abnormal example is the rejection of a transplanted organ. Other abnormal examples may include autoimmune attacks (e.g., lupus nephritis or multiple sclerosis that damages the kidney), inflammatory diseases (e.g., hepatitis), fibrotic processes (e.g., cirrhosis), fatty infiltration (e.g., fatty liver disease), degenerative processes (e.g., Alzheimer's disease) and ischemic tissue damage (e.g., myocardial infarction or stroke). The health status of an object may be considered as a normal classification.

[0092] The term "gestational age" may refer to a measure of gestational age measured from the woman's last menstrual period (LMP), or to the corresponding gestational age as estimated by more precise methods, if available. Such methods include adding 14 days from fertilization (possibly in IVF) to a known duration or by obstetric ultrasound.

[0093] A "pregnancy-related disorder" includes any disorder characterized by abnormal relative expression levels of genes in maternal and / or fetal tissues or abnormal clinical features in the mother and / or fetus. These conditions include, but are not limited to, preeclampsia (Kaartokallio et al. Sci Rep. 2015; 5:14107; Medina-Bastidas et al. Int J Mol Sci. 2020; 21:3597), intrauterine growth restriction (Faxén et al. Am J Perinatol. 1998; 15:9-13; Medina-Bastidas et al. Int J Mol Sci. 2020; 21:3597), placenta accreta, preterm birth (Enquobahrie et al. BMC Pregnancy Childbirth. 2009; 9:56), hemolytic disease of the newborn, placental insufficiency (Kelly et al. Endocrinology. 2017; 158:743-755), fetal edema (Magor et al. Blood. 2015; 125:2405-17), fetal malformations (Slonim et al. al.Proc Natl Acad Sci USA.2009;106:9425-9), HELLP syndrome (Dijk et al.J Clin Invest.2012;122:4003-4011), systemic lupus erythematosus (Hong et al.J Exp Med.2019;216:1154-1169), and other immune diseases of the mother.

[0094] A "site" (also referred to as a "genomic site") corresponds to a single site, which can be a single base position or a group of related base positions, such as a CpG site or a larger group of related base positions. A "locus" can correspond to a region that includes multiple sites. A locus can include only one site, which would make the locus equivalent to a site in this context.

[0095] The term "about" or "approximately" can refer to within an acceptable error range of a particular value determined by a person of ordinary skill in the art, which will depend in part on how to measure or determine the value, i.e., the limitations of the measurement system. For example, "about" can refer to within 1 or more than 1 standard deviation according to the practice of the art. Alternatively, "about" can refer to a range of up to 20%, up to 10%, up to 5%, or up to 1% of a given value. Alternatively, particularly for biological systems or processes, the term "about" or "approximately" can refer to within an order of magnitude of a value, within 5 times, and in some forms within 2 times. When describing a particular value in an application and a claim, unless otherwise stated, it should be assumed that the term "about" refers to within an acceptable error range of the particular value. The term "about" can have a meaning commonly understood by a person of ordinary skill in the art. The term "about" can refer to ±10%. The term "about" can refer to ±5%.

[0096] In the case of providing a range of values, it should be understood that, unless the context clearly specifies otherwise, each intermediate value between the upper and lower limits of the range to one tenth of the unit of the lower limit is also specifically disclosed. It should also be understood that the endpoints of the provided range are included in the range. Each smaller range between any of the values ​​or intermediate values ​​in the range and any other values ​​or intermediate values ​​in the range is included in the embodiments of the present disclosure. The upper and lower limits of these smaller ranges can be independently included in the range or excluded from the range, and each range in which any one restriction, both restrictions are not or both restrictions are included in the smaller range is also included in the present disclosure, subject to any special exclusion restrictions in the range. In the case where the range includes one or two of the restrictions, the scope excluding one or two of the restrictions included is also included in the present disclosure.

[0097] Standard abbreviations may be used, eg, bp, base pair; kb, kilobase; pi, picoliter; s or sec, second; min, minute; h or hr, hour; aa, amino acid; nt, nucleotide, etc.

[0098] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of embodiments of the present disclosure, some potential and exemplary methods and materials can now be described. Detailed Description

[0099] The epigenomic state of different regions of chromatin (DNA and protein) can indicate the expression activity of genes, tissue origin or disease. Histone modification is an example of epigenomic factors, in which the measurement of the amount of histones with a specific epigenomic state can be used in various ways. The technology for detecting histone modifications includes cfChIP-seq (cell-free chromatin immunoprecipitation, followed by sequencing), which has some disadvantages. The cfChIP-seq technology requires 1-2ml or more samples, which is a large sample amount compared to the hundreds of microliters or less used when sequencing is performed alone. In addition, compared with the procedure of conventional plasma cfDNA-seq, cfChIP-seq uses a more complex and time-consuming sample technology. In the cfChIP-seq procedure, the target epigenome is associated with proteins (e.g., histone modifications). Compared with DNA, proteins are unstable. The effects of freezing, thawing and storage conditions on protein stability are greater than those of DNA.

[0100] The present disclosure shows that certain terminal motifs (i.e., sequences of naturally fragmented DNA ends), lengths and / or other fragmentomic properties of cell-free DNA are highly correlated with histone modifications. The amount of these terminal motifs can indicate the amount of histone modifications in a sample, and therefore in a subject. As a result, terminal motifs can be used to indicate the activity of a gene, tissue origin, or disease, avoiding the shortcomings of cfChIP-seq. Analysis of terminal motifs can use sequencing techniques that do not require additional steps of cfChIP-seq. Therefore, embodiments of the present invention can use less than 100 μl of biological samples, which can include about 500 μg of cell-free DNA. Sampling processing for sequencing is much simpler than using cfChIP-seq technology. Samples do not need to be frozen to temperatures below -80°C. Samples can be transported farther from the clinic to the laboratory. In addition, analysis of terminal motifs can be used to study a variety of different epigenomic types from a single measurement, rather than being limited to specific histone modifications associated with specific antibodies used in a specific cfChIP-seq assay.

[0101] Thus, measuring certain end motifs of cell-free DNA can provide an improved technique for determining the epigenomic state of a specific region of chromatin (e.g., corresponding to a specific region of a reference genome). Additionally, measuring certain end motifs can also determine different properties of a sample, where such properties are related to the epigenomic state of a specific region, such as the fractional concentration of a tissue type, the classification of a condition, gestational age, the nutritional state of an organ, the length of an organ, or other properties. These properties can be determined using the epigenomic state determined from the end motifs or directly from the end motifs.

[0102] The sample can be physically or in silico enriched for certain terminal motifs that are more frequently associated with certain epigenomic states, including histone modifications. Enrichment of the sample can allow for more accurate measurement of the properties of the sample, measurement of the amount of histone modifications, or determination of the condition of an organism. I. Epigenomic Status

[0103] Figure 1 An icon showing the structure of DNA. The DNA inside a cell is a large structure. In addition to the nucleotides in the DNA, the DNA is also assembled by several different proteins, including chromatin remodelers, transcription factors, nucleosomes, and histones. Histones are proteins around which the DNA is wrapped. DNA is usually wrapped around eight histones (e.g., histone octamer 104). The structural unit of DNA around histones is the nucleosome. Histones can carry modifications that can affect gene transcription. Histone modifications include methylation and acetylation. Histone modifications are part of the epigenome. The epigenomic state is different for different types of cells. The structure of DNA and proteins inside a cell is called chromatin. Within chromatin, the DNA itself is also methylated. Protein structures that physically open and close chromatin and other DNA modifications that contribute to chromatin structure are also part of the epigenome. Chromatin remodelers are general tools that catalyze a variety of chromatin-altering reactions, including sliding of the octamer on the DNA (nucleosome sliding), changing the conformation of nucleosomal DNA, and changing the composition of the octamer (histone variant exchange). In addition, chromatin remodelers can remove other chromatin proteins from the chromatin.

[0104] Histone modifications have various functions in cells. One function is to regulate gene expression. Gene expression can be promoted or inhibited. For example, the amount of H3K4me3 is related to transcriptional activity. In some cases, histone modifications can increase chromatin compaction and reduce transcription (e.g., H3K36me3). II. Measuring epigenomic status A. Histone modifications identified using cfChIP-seq

[0105] The plasma DNA pool is a mixture of DNA molecules released from various tissues, some of which will bind to histones accompanied by certain histone modifications. Histones include H1 (joint histones), H2A / B, H3 and H4 (core histones). DNA molecules will form nucleosome structures together with histones (Zhou et al.Nat Struct Mol Biol.2019; 26: 3-13). The curling of DNA around histones is mainly due to the electrostatic affinity between the positively charged histones and the negatively charged phosphate backbone of DNA. Histone modifications include but are not limited to histone methylation, acetylation, phosphorylation and ubiquitination, etc. (Barth et al.Trends Biochem.Sci.2010; 35: 618-626). Histone methylation can occur at different lysine residues of histones. The methylation of each lysine residue can involve one, two or three methyl groups, so that the lysine residues will be monomethylated, dimethylated or trimethylated, respectively. Examples of histone methylation include, but are not limited to, trimethylation of lysine (K) residue 4 at the N-terminus of histone H3 (H3K4me3), monomethylation of lysine (K) residue 4 at the N-terminus of histone H3 for transcriptional activation (H3K4me1), H3K27me3 and H3K9me3 for transcriptional inactivation, and H3K36me3 associated with transcribed regions in gene bodies. It is reported that H3K9me2 is a signal for the formation of heterochromatin in gene-poor chromosomal regions, telomeres, and near centromeres with tandem repeat structures (e.g., satellite repeats). Histone acetylation includes, but is not limited to, H3K27ac, H3K9ac, and H3K14ac, etc.

[0106] Plasma cfDNA molecules bound to histones with certain modifications can be isolated by chromatin immunoprecipitation. Those immunoprecipitated plasma cfDNA molecules can be analyzed using different techniques. In one embodiment, they can be analyzed by DNA sequencing.

[0107] Figure 2Plasma cfDNA molecules associated with histone modifications are shown to be analyzed using immunoprecipitation. Stage 204 shows the plasma portion of a blood sample. Plasma is separated. Stage 208 shows components of plasma, including DNA, DNA around histones, and DNA around histones with histone modifications. Plasma cfDNA molecules associated with histone modifications such as H3K27ac are precipitated by magnetic beads conjugated to H3K27ac antibodies. In stage 212, precipitated plasma cfDNA molecules are shown. In stage 216, a DNA library is prepared and the DNA molecules are connected to a barcode adapter. The precipitated cfDNA molecules are analyzed by next generation sequencing (e.g., Illumina NextSeq500). Sequencing reads can be aligned with the human reference genome GRCh37 (hg19), for example using Bowtie2 (Langmead et al. Nat Methods. 2012; 9: 357-359). In some embodiments, SOAP2 (Li et al. Bioinformatics. 2009; 25: 1966-67), Burrows-WheelerAligner (BWA) (Li et al. Bioinformatics. 2009; 25: 1754-60), BLAT (Kent. Genome Res. 2002: 12: 656-664), BLAST (Zhang et al. J Comput Biol. 2000; 7: 203-14), BFAST (Homer N et al. PLoS One. 2009; 4: e7767), MOSAIK (Lee et al. PLoS One. 2014; 9: e90581), etc. can be used but are not limited thereto. Stage 220 shows a graph of the relationship between histone modification signals (y-axis) and genomic positions (x-axis). The sequencing depth (or sequencing read density) in a specific genomic region represents the degree of H3K27ac modification present in the region of different cell types. The higher the sequencing depth in a specific region, the more H3K27ac modifications that may be identified. If this H3K27ac modification is specific to a specific cell type in a specific region, the sequencing depth in this region can be used to measure the amount of cfDNA molecules carrying H3K27ac from this cell type. In one embodiment, the sequencing depth can be standardized and corrected by sequencing deviation and / or the noise generated by non-specific binding. In some instances, the sequencing depth (i.e., ChIP-seq) associated with subsequent sequencing with chromatin immunoprecipitation assay can be used to define histone modification signals or ChIP signals. B. Selected terminal motifs indicate histone modifications

[0108] Using fragment omics properties, including but not limited to plasma DNA terminal motifs and length, we have developed a new method for analyzing histone modifications in plasma without the need for immunoprecipitation. Relatively rich regions of histone modifications will produce different fragment terminal motif patterns compared to regions lacking histone modifications. Therefore, the pattern of fragment terminal motifs can be used to infer histone modifications. The terminal motif can be defined as one or more nucleotides at one end of a cell-free DNA fragment. The number of nucleotides (nt) at the end of each fragment for analysis can be, for example, but not limited to, 1nt, 2nt, 3nt, 4nt, 5nt, 6nt, 7nt, 8nt, 9nt and 10nt or more. The length of plasma DNA fragments can be measured by various methods. In one embodiment, the length of plasma DNA fragments can be measured by the number of nucleotides present in plasma DNA molecules. In another embodiment, the length of plasma DNA fragments can be measured using paired end sequencing, aligning the sequence with the genome, and then inferring the length from the genomic coordinates of the aligned sequence. In the embodiment, tissue or disease-specific histone modification levels are inferred from cfDNA terminal motifs or length frequencies, etc., so that the physiology or pathology of one or more tissues can be monitored, or disease states can be detected.

[0109] Regions with histone modifications may include, but are not limited to, repeat regions, X-chromosome inactivation regions, chromatin structures [e.g., open and closed chromatin structures], pseudogenes, CTCF, DNase I hypersensitive sites [DHS], actively transcribed regions and passively transcribed regions, G-quadruplexes, etc. For example, terminal motifs selected in regions with DNase I hypersensitive sites can be used to inform the amount of histone modifications associated with the DNase I hypersensitive sites. As another example, the length of a DNA fragment in an X-chromosome inactivation region can inform the amount of histone modifications of an X-chromosome gene.

[0110] Specific regions can be associated with specific tissue types. In some cases, for specific tissue types, a certain property in the region may occur more often. As an example, for specific tissue types, the region of open chromatin (i.e., the large gap between the histones) may occur more often than other tissue types. Other properties may include the region as a repeat region, an X chromosome inactivation region, a closed chromatin structure, a pseudogene, CTCF, DHS, an active transcription region, a passive transcription region or a G quadruplex. Specific regions can be associated with specific specific tissue types, and not associated with other tissue types. In other embodiments, specific regions can be associated with several different tissue types. The popularity of this regional property may be related to the relative intensity of the specific tissue related to the contribution of the specific tissue type and the regional property. Deconvolution can be used to determine the contribution of tissues from these regions, similar to that described below for histone modification. 1. Identify terminal motifs associated with histone modifications

[0111] Different histone modifications can give different accessibility to DNA nucleases, resulting in characteristic fragmentation. Selective cutting of DNA by nucleases through cfDNA fragmentation occurs in TSS and CpG islands with specific epigenetic states (Han et al., Genome Res.2021:31:2008-2021). The fragmentation pattern of cell-free DNA can be used to infer histone modifications present in plasma DNA molecules. In the implementation scheme, we analyzed the nucleases with cutting preferences for cfDNA in the region of interest, which can be indicated by the pattern of cfDNA terminal motifs. The fragment terminal motif can be defined by one or more nucleotides at one end of the cell-free DNA fragment. For example, we determined the proportion of cfDNA molecules carrying specific 4-mer terminal motifs (a total of 256 types).

[0112] Figure 3 A diagram showing the terminal motifs of a fragment. Each nucleotide can be one of four nucleotides: A, C, G, T. For a terminal motif of four nucleotides, there are four 4 The 4-mer terminal motif is defined as the four nucleotides at the 5' end of the cfDNA molecule.

[0113] Regions involved in histone modifications can be grouped into different categories based on the length of the ChIP signal. Figure 4 is a graph defining categories of H3K4me3 regions with different levels of H3K4me3 ChIP signals. The y-axis is a graph with log 10 Scaled H3K4me3 signal. The x-axis shows the rank of genomic regions associated with H3K4me3. Higher ranks indicate higher signals. Regions were first sorted based on the length of ChIP signal and then empirically classified into 9 categories.

[0114] Figure 5 is a table showing exemplary definitions of H3K4me3 region categories using H3K4me3 ChIP-seq analysis of pregnancy samples. The first column shows the category identity. The second column shows the number of regions in the category. The third column shows the percentile range of the length of the ChIP signal in the regions of the category. The fourth column shows the average ChIP signal in the regions of the category. Figure 5As shown in the figure, we empirically classified the regions associated with H3K4me3 into 9 categories based on the percentile range of ChIP signal intensity. The intensity of the ChIP signal of a region can be the average FPKM of 12 pregnancy samples analyzed by H3K4me3 ChIP-seq. For example, the percentile range of ChIP signal from 0 to 70 is defined as category 1, with an average ChIP signal of 0.10; the percentile range of ChIP signal from 70 to 80 is defined as category 2, with an average ChIP signal of 0.81; the percentile range of ChIP signal from 80 to 90 is defined as category 3, with an average ChIP signal of 1.59; the percentile range of ChIP signal from 90 to 95 is defined as category 4, with an average ChIP signal of 3.27; the percentile range of ChIP signal from 95 to 97 is defined as category 5. Category 5, the average ChIP signal was 5.84; the percentile range of ChIP signal was 97 to 98 was defined as category 6, the average ChIP signal was 9.93; the percentile range of ChIP signal was 98-98.5 was defined as category 7, the average ChIP signal was 14.63; the percentile range of ChIP signal was 98.5 to 99 was defined as category 8, the average ChIP signal was 18.81; the percentile range of ChIP signal was 99 or above was defined as category 9, the average ChIP signal was 31.68.

[0115] Figure 6 A table showing exemplary definitions of categories of H3K27ac regions using H3K27ac ChIP-seq analysis of pregnancy samples is shown. Figure 6 The table in follows the Figure 5 The same format as the table in . Figure 6As shown in the figure, we empirically classified the regions associated with H3K27ac into 9 categories based on the percentile range of ChIP signal intensity. The intensity of the ChIP signal of a region can be the average FPKM of 19 pregnancy samples analyzed by H3K27ac ChIP-seq. For example, the percentile range of ChIP signal from 0 to 70 is defined as category 1, with an average ChIP signal of 0.45; the percentile range of ChIP signal from 70-80 is defined as category 2, with an average ChIP signal of 0.99; the percentile range of ChIP signal from 80 to 90 is defined as category 3, with an average ChIP signal of 1.31; the percentile range of ChIP signal from 90 to 95 is defined as category 4, with an average ChIP signal of 1.84; the percentile range of ChIP signal from 95 to 97 is defined as category 5 The percentile range of ChIP signals was defined as category 5, with an average ChIP signal of 2.43; the percentile range of ChIP signals was 97 to 98 was defined as category 6, with an average ChIP signal of 2.93; the percentile range of ChIP signals was 98-98.5 was defined as category 7, with an average ChIP signal of 3.34; the percentile range of ChIP signals was 98.5 to 99 was defined as category 8, with an average ChIP signal of 3.74; the percentile range of ChIP signals was 99 or above was defined as category 9, with an average ChIP signal of 5.33. We can also use other methods to define regional categories, including but not limited to k-means clustering analysis.

[0116] Figure 7 is a table showing exemplary definitions of categories of H3K4me3 regions using H3K4me3 ChIP-seq analysis of samples from non-pregnant healthy subjects. Figure 7 The table in follows the Figure 5 Same format as the table in . Figure 7 A non-pregnant healthy sample was used to establish a reference for ChIP-seq analysis. Figure 7As shown in the figure, we empirically classified the regions associated with H3K4me3 into 9 categories based on the percentile range of ChIP signal intensity. The intensity of the ChIP signal of a region can be the average FPKM of 4 healthy samples analyzed by H3K4me3 ChIP-seq. For example, the percentile range of ChIP signal from 0 to 70 is defined as category 1, and the average ChIP signal is 0.00; the percentile range of ChIP signal from 70-80 is defined as category 2, and the average ChIP signal is 0.15; the percentile range of ChIP signal from 80 to 90 is defined as category 3, and the average ChIP signal is 0.69; the percentile range of ChIP signal from 90 to 95 is defined as category 4, and the average ChIP signal is 2.71; the percentile range of ChIP signal from 95 to 97 is defined as category 5 Category 5, the average ChIP signal was 6.00; the percentile range of ChIP signal was 97-98 was defined as category 6, the average ChIP signal was 11.39; the percentile range of ChIP signal was 98-98.5 was defined as category 7, the average ChIP signal was 17.11; the percentile range of ChIP signal was 98.5-99 was defined as category 8, the average ChIP signal was 21.95; the percentile range of ChIP signal was 99 or above was defined as category 9, the average ChIP signal was 35.44.

[0117] Figure 8 is a table showing exemplary definitions of categories of H3K27ac regions using H3K27ac ChIP-seq analysis of samples from non-pregnant healthy subjects. Figure 8 The table in follows the Figure 5 The same format as the table in . Figure 8As shown in the figure, we empirically classified the regions associated with H3K27ac into 9 categories based on the percentile range of ChIP signal intensity. The intensity of the ChIP signal of a region can be the average FPKM of 6 healthy samples analyzed by H3K27ac ChIP-seq. For example, the percentile range of ChIP signal from 0 to 70 is defined as category 1, with an average ChIP signal of 0.23; the percentile range of ChIP signal from 70-80 is defined as category 2, with an average ChIP signal of 0.89; the percentile range of ChIP signal from 80 to 90 is defined as category 3, with an average ChIP signal of 1.49; the percentile range of ChIP signal from 90 to 95 is defined as category 4, with an average ChIP signal of 2.45; the percentile range of ChIP signal from 95 to 97 is defined as category 5, with an average ChIP signal of 2.65; The percentile range of ChIP signals was 97-98 as category 6, with an average ChIP signal of 4.07; the percentile range of ChIP signals was 98-98.5 as category 7, with an average ChIP signal of 4.56; the percentile range of ChIP signals was 98.5 to 99 as category 8, with an average ChIP signal of 5.01; and the percentile range of ChIP signals was 99 or above as category 9, with an average ChIP signal of 6.54.

[0118] For samples without immunoprecipitation, we analyzed the frequency of 4-mer terminal motifs in nine categories defined according to different levels of H3K4me3 signal.

[0119] Fig. 9 A heat map of motif frequencies in regions with different levels of H3K4me3 ChIP signals is shown for plasma DNA sequencing results. Figure 904 shows the average H3K4me3 ChIP signal. The y-axis shows the average H3K4me3 ChIP signal. The x-axis shows the 9 categories of H3K4me3 regions. The x-axis categories are aligned with the regions in heat map 908. The y-axis of heat map 908 corresponds to different 4-mer terminal motifs. The redder the point, the higher the frequency of the terminal motif in a region of a sample compared to other combinations of region category and sample. The bluer the point, the lower the frequency of the terminal motif in a region of a sample compared to other combinations of region category and sample. As shown in Fig. 9 As shown, the terminal motif frequency obtained from the plasma DNA sequencing data without immunoprecipitation changes according to the intensity of the ChIP signal obtained from the immunoprecipitated plasma DNA sequencing data, which indicates the probability of inferring plasma DNA histone modifications based on the terminal motifs of the plasma DNA molecules without immunoprecipitation. Point 912 is the point where four quadrants of unequal lengths intersect. The upper right quadrant is redder. The upper left quadrant is bluer. The lower right quadrant is bluer. The lower left quadrant is redder.

[0120] Fig.10 is a comparison of terminal motif frequency rankings between plasma DNA sequencing results with and without H3K4me3-based immunoprecipitation. The y-axis shows the ranking of terminal motifs from cfChIP-seq for H3K4me3 from 256 to 1, where 1 represents the most frequent terminal motif. The x-axis shows the ranking of terminal motifs from 256 to 1 generated by conventional cfDNA sequencing on plasma samples without the addition of H3K4me3 modification-specific antibodies. The shape of the data point represents the terminal nucleotide (A is a circle, C is a triangle, G is a square, and T is a plus sign).

[0121] Fig.10 A variety of 4-mer terminal motifs that appear to be over-represented in the sequencing results of plasma DNA from H3K4me3-mediated immunoprecipitation compared to plasma DNA without immunoprecipitation are shown, including but not limited to GCGG, GCGC, CGCG, CCGC, CCGA, TCCG, CCGT, GGCG, CCGG, TGCG, GCCG, CTCG, GCGA, TCGG, CGGC, TCGC, CGGG, CGCC, ACCG, AGCG, CGGA, GGGC, GCGT, CACG, etc. (i.e., such motifs where the rank on the y-axis is a lower number than the rank on the x-axis). Over-represented terminal motifs are considered to be above the diagonal y=x and xy > 100. Those terminal motifs that are over-represented may suggest the presence of a histone modification (H3K4me3). In another embodiment, under-represented motifs (i.e., motifs where the rank on the y-axis is a higher number than the rank on the x-axis) may be used.

[0122] Fig.11 A table of 24 terminal motifs with the largest ranking differences between conventional cfDNA sequencing and cfChIP-seq for H3K4me3 histone modification is shown. The first column shows the motif. The second column shows the nucleotide at the very end of the fragment (i.e., the first nucleotide listed in the first column). The third column shows the ranking of the motif in conventional cfDNA sequencing, where 1 is the most frequent and highest ranking and 256 is the least frequent and lowest ranking. The fourth column shows the ranking of the motif in cfChIP-seq modified by H3K4me3 histone. The fifth column shows the ranking difference when the cfChIP-seq ranking is adopted and the conventional cfDNA sequencing ranking is subtracted. The columns are arranged according to the length of the ranking difference. Data were obtained from multiple healthy subjects.

[0123] The results also showed that many terminal motifs with high rankings in cfChIP-seq had C and G nucleotides adjacent to each other. H3K4me3 sites appeared to be enriched in CG sequences.

[0124] Thus, terminal motifs with the greatest ranking differences appear at higher rates in regions associated with H3K4me3 than without cfChIP, genome-wide, or relative to a random set of DNA fragments.

[0125] Fig. 12A and Fig. 12B The results of plasma DNA sequencing without immunoprecipitation are described using end-terminal motif patterns to infer plasma DNA histone modification signals. Fig. 12A Shown is the establishment of a recalibration formula with the frequency of over-represented terminal motifs and the level of H3K4me3 ChIP signals in 9 categories. In stage 1204, the regions involving H3K4me3 are grouped into different categories according to the length of the ChIP signal. In one embodiment, the regions can be divided into 9 categories based on the length of the ChIP signal of each region. After we obtain the regional categories with different ChIP signals, the terminal motif pattern in each regional category (e.g., the total frequency of cfDNA molecules with over-represented terminal motifs from plasma DNA sequencing results without immunoprecipitation) can be used to correlate with the H3K4me3 ChIP signal. In stage 1208, based on the correlation between the fragment terminal motif and the ChIP signal, a recalibration formula can be determined. A linear formula is shown as an example of a recalibration formula, but a nonlinear formula can also be used.

[0126] Fig. 12B It is shown how to use the recalibration formula to infer ChIP signals in other regions (e.g., placenta-specific H3K4me3 regions) based on the corresponding end motif information (i.e., inferred ChIP signals) of those regions. In stage 1212, plasma DNA is sequenced without immunoprecipitation. In stage 1216, molecules derived from tissue-specific (e.g., placenta) H3K4me3 regions are identified. In stage 1220, the frequency of over-represented end motifs in H3K4me3-based immunoprecipitated plasma DNA is calculated. The end motif information is input into the recalibration formula, and in stage 1224, H3K4me3 ChIP signals are inferred in tissue-specific (e.g., placenta) H3K4me3 regions. 2. Testing for correlation with cfChIP-seq signals

[0127] Fig.13Figure 2 shows the correlation between the aggregate abundance of over-represented terminal motifs in H3K4me3-based immunoprecipitated plasma DNA and the H3K4me3 ChIP signal. The x-axis shows the frequency of over-represented terminal motifs as a percentage. The y-axis is log 10 Scaled H3K4me3 ChIP signal. Fig.13 The combined abundance of overrepresented terminal motifs was shown to be highly correlated with the H3K4me3 ChIP signal (Pearson's r: 0.99; p-value: <0.0001). This result shows that the use of plasma DNA terminal motifs can be used to infer the intensity of signals associated with certain histone modifications. Therefore, a linear regression model can be used to generate a recalibration formula to facilitate inferring H3K4me3 ChIP signals based on the terminal motifs of plasma DNA molecules without the need for immunoprecipitation assays. In addition, motif frequencies can be used to predict H3K4me3 histone modifications and any other properties that can use H3K4me3 histone modifications, such as the percentage of DNA from a specific tissue type or subject condition.

[0128] From Fig.11 A higher frequency of the 24 terminal motifs would be expected to correlate with a higher cfChIP-seq signal. To test this hypothesis, we divided the cfChIP-seq signal into different numbers of groups based on the height of the peaks.

[0129] Fig.14 is a graph showing the correlation between cfChIP signal and terminal motif frequency for 11 peak groups. Each dot (data point) corresponds to a different peak group among the 11 peak groups. Because peaks correspond to signal values, the signal increases with consecutive peak groups. The x-axis shows the total frequency of terminal motifs in the peak group relative to all motifs in the specific genomic region being analyzed. The terminal motif frequency of a peak group is the frequency of the specific genomic region associated with the peak group. The y-axis shows the average signal from cfChIP-seq for H3K4me3 histone modification for each peak group. As an example, each peak group can include multiple peaks, such as Figure 5 , 6 , 7 and 8 as shown.

[0130] Fig.15A is a graph showing the correlation between cfChIP signals and terminal motif frequencies for the six peak groups. The y-axis shows the average signal from cfChIP-seq for the H3K4me3 histone modification for each peak group. The x-axis shows the frequency of terminal motifs, similar to Fig.14 Each dot represents one of the six peak groups. The end motif frequency of a peak group is the frequency of a specific genomic region associated with the peak group. Fig.15A The terminal modules include Fig.11All 24 terminal motifs identified in the . The figure shows a high correlation with an R value of 0.98 and a p value of 0.00059. The figure shows that the frequency of using the top 24 terminal motifs is correlated with the cfChIP-seq signal of the H3K4me3 histone modification. The figure also shows that grouping the terminal motifs into six peak groups can maintain the correlation with the cfChIP-seq signal.

[0131] Fig. 15B is a graph showing the correlation between cfChIP signals and terminal motif frequencies for eight peak groups. The y-axis shows the average signal from cfChIP-seq for H3K4me3 histone modification for each peak group. The x-axis shows the frequency of terminal motifs, similar to Fig.14 Each dot represents one of the eight peak groups. The end motif frequency of a peak group is the frequency of a specific genomic region associated with the peak group. Fig.15A The terminal modules include Fig.11 All 24 terminal motifs identified in the . The figure shows a high correlation with an R value of 0.97 and a p value of 4.4e-05. The figure shows that the frequency of using the top 24 terminal motifs is correlated with the cfChIP-seq signal of the H3K4me3 histone modification. The figure also shows that grouping the terminal motifs into eight peak groups can maintain the correlation with the cfChIP-seq signal.

[0132] Fig.14 , 15A Figures 15 and 15B show that terminal motif frequencies are correlated with the signal from cfChIP-seq peaks within a group. The correlation remains high even when varying the number of peak groups. III. Analysis of epigenomic states using sequence motifs

[0133] Because terminal motif frequency can identify epigenomic states and different cells have different epigenomic states, terminal motif frequency can be used to identify tissue origin, determine the concentration fraction of a tissue in a sample, estimate the characteristics of a tissue, or determine the grade of a disease. Terminal motif frequency can also measure the amount of histone modification. A. Estimation of the concentration fraction of the tissue of origin

[0134] For placenta, genomic regions with high H3K4me3 signals are known ( Figure 4 In addition, the terminal motif frequencies of these genomic regions are known for different peak groups ( Fig.14 ). The total terminal motif frequency of 24 terminal motifs in different genomic regions corresponding to 11 peak groups was determined. Based on the terminal motif frequency, the H3K4me3 signal was predicted. Description Fig.14 The equation for the linear relationship in is log(average H3K4me3 signal)=a*(terminal motif frequency)+b. 1. Results

[0135] Fig.16 Graph showing the correlation between H3K4me3 ChIP signals in placenta-specific H3K4me3 regions inferred by terminal motifs and the fetal DNA fraction determined by a SNP-based approach. The x-axis is the percentage of fetal DNA fraction determined by a SNP-based approach. The y-axis is the H3K4me3 ChIP signal inferred using terminal motifs. The H3K4me3 ChIP signal inferred using terminal motifs correlated with the fetal DNA fraction in maternal plasma DNA (Pearson's r: 0.67; p-value: <0.001). 2. Exemplary Method for Determining Concentration Fractions

[0136] Fig.17 1700 is a flow chart of an exemplary process 1700 associated with determining the fractional concentration of cell-free DNA fragments in a biological sample. The biological sample may include cell-free DNA fragments. The biological sample may be any biological sample described herein, including plasma or serum. In some embodiments, Fig.17 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.17 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.17 One or more process blocks may be performed by one or more components of the measurement system 5900 , such as the determination 5908 , the determination device 5910 , the detector 5920 , the logic system 5930 , the local memory 5935 , the external memory 5940 , the storage device 5945 , and / or the processor 5950 .

[0137] A plurality of sequence reads of cell-free DNA fragments are received at block 1710. The plurality of sequence reads include end sequences corresponding to ends of the plurality of cell-free DNA fragments.

[0138] In some embodiments, method 1700 may include sequencing cell-free DNA fragments in a biological sample to obtain a plurality of sequence reads. In embodiments, the volume of the biological sample may be 100 μl or less, including 80 to 100 μl, 50 to 80 μl, or 30 to 50 μl. The biological sample may use a volume that is smaller than the volume used in cfChIP-seq.

[0139] In some embodiments, method 1700 may include probe-based techniques to measure the amount of a motif. Techniques may include qPCR, digital PCR, digital droplet PCR, and the like. As an example, cfDNA molecules may undergo a process of DNA end pairing, A-tailing, and common adapter ligation. The molecules connected to the adapters may be separated, for example, into different reactions, such as droplets. A pair of PCR primers may be designed so that one primer may bind to a common adapter region, while the other primer may bind to a specific region of interest. The DNA molecules will be amplified by a PCR primer pair within a reaction (e.g., a droplet). Fluorescent probes specific to a certain terminal motif may hydrolyze and emit fluorescent signals, thereby enabling detection of the presence of a specific motif and quantification of a specific motif. For digital PCR, the number of reactions positive for a specific terminal motif may be counted and used to determine the amount of DNA fragments having the terminal motif in the analysis region. For real-time PCR, the intensity of each signal may be used as a measure of the amount of DNA fragments terminated with a specific motif. The two intensities may be compared to each other.

[0140] At block 1720, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions has a histone modification associated with a target tissue type. The target tissue type may include placenta, liver, heart, neutrophils, monocytes, B cells, fat, NK cells, or any tissue type described herein. The histone modification may be H3K4me3, H3K4me1, H3K4me2, H3K27me3, H3K27ac, H3K36me3, H3K9me2, H3K9me3, H3S10P, H3R2me, H3T2P, H3K14ac, H3K9ac, H3K79me2, H3K79me3, H4K5ac, H4K8ac, H4K12ac, H4K16ac, H4K20me, H2BK120ub, H2AK119ub. One or more genomic regions may include a transcription start site, a promoter region, an enhancer region, a super enhancer region, a gene body, a repetitive sequence, a satellite repetitive sequence, a telomere, a pericentromeric region, a mitotic chromosome, a transcription end site, an exon, an intron, an insulator, etc. One or more genomic regions may have an amount of histone modification that is statistically significantly different from the amount of histone modification in other genomic regions, or the average amount of modification in other genomic regions or in all genomic regions. The sequence read can be aligned with a reference genome (e.g., a human reference genome) to determine whether the sequence read is located in one or more genomic regions.

[0141] At block 1730, for each sequence read of the set of sequence reads, one or more sequence motifs are determined that correspond to one or more end sequences of the corresponding cell-free DNA fragment. The one or more sequence motifs may correspond to a single nucleotide, a sequence of two nucleotides, a sequence of three nucleotides, a sequence of four nucleotides, a sequence of five nucleotides, a sequence of six nucleotides, a sequence of seven nucleotides, a sequence of eight nucleotides, or a sequence having more than eight nucleotides. The one or more sequence motifs may each have the same number of nucleotides. In some embodiments, the sequence motif includes nucleotides at the ends of the cell-free DNA fragments. The sequence motif may be at the 5' end of the cell-free DNA fragments. In some embodiments, the sequence motif may be at the 3' end. In embodiments, the one or more sequence motifs may include sequence motifs at the 3' end and the 5' end. If the entire fragment is sequenced, two sequence motifs may be determined.

[0142] At block 1740, one or more relative frequencies of a set of one or more sequence motifs are determined. The set of one or more sequence motifs appears at a higher rate in chromatin immunoprecipitation sequencing (cfChIP-seq) than in sequencing without chromatin immunoprecipitation for one or more genomic regions associated with histone modifications. Chromatin immunoprecipitation can be cell-free chromatin immunoprecipitation sequencing (cfChIP-seq), or can be cell-based chromatin immunoprecipitation sequencing. Sequencing without chromatin immunoprecipitation can include whole genome sequencing. The set of one or more sequence motifs corresponds to sequence motifs with similar relative frequencies, such as Fig.14 , 15A or the peak group in 15B. One or more sequence motifs may be, for example Fig.11 Any sequence motif in . The relative frequency can be Fig.14 , 15A Or the motif frequency in 15B. The set of one or more sequence motifs can include 1 to 5, 5 to 10, 11 to 15, 15 to 20 or 20 to 25 sequence motifs. The relative frequency of each sequence motif can be determined. In other embodiments, a relative frequency can be determined for multiple sequence motifs (including the set of one or more sequence motifs). The set of determining sequence motifs is described below.

[0143] In box 1750, determine one or more total values ​​of relative frequencies. Exemplary total values ​​are described throughout the disclosure, such as entropy values ​​(motif diversity scores or variances), sums of relative frequencies, and multidimensional data points corresponding to count vectors of sets of motifs (e.g., 256 counts of a vector of 256 motifs of possible 4-mers, or 64 counts of 64 motifs of possible 3-mers). When the set of one or more sequence motifs includes a variety of sequence motifs, the total value may include the sum of the relative frequencies of the set. In some embodiments, the total value may be an estimate of histone modifications. The level of histone modifications can be determined by various types of data, such as the amount of terminal motifs or fragment lengths.

[0144] At block 1760, the aggregate value is compared to one or more calibration values. The one or more calibration values ​​are determined from one or more calibration samples having a known fractional concentration of cell-free DNA fragments from the target tissue type.

[0145] One or more calibration values ​​can be determined by determining the total value of the sequence motifs of one or more calibration samples. For example, the total value determined from the biological sample can be a first total value determined from one or more first relative frequencies. One or more second relative frequencies of a set of one or more sequence motifs in one or more genomic regions can be determined for each calibration sample of one or more calibration samples. A second total value can be determined for one or more second relative frequencies of each calibration sample in one or more calibration samples. Therefore, each of the one or more second total values ​​can be associated with a known concentration of the calibration sample. The calibration value can include one or more second total values. For example, the calibration value can be a point along a straight line or curve that relates a known concentration to a second total value.

[0146] In some embodiments, one or more calibration values ​​can be determined based on a function that relates a known concentration to a second total value. A first total value can be input into a function to return a concentration fraction. The first total value is then used as a calibration value. The comparison of the total values ​​is to compare the total value with the calibration value used in the function and determine that the total value is the same as the calibration value.

[0147] At block 1770, the comparison is used to determine a fractional concentration of cell-free DNA fragments from the target tissue type. The fractional concentration may be a known fractional concentration associated with a calibration value that may have a value close to or equal to the first aggregate value. In some embodiments, the fractional concentration may be determined based on a function or straight line having one or more calibration values. The function or straight line may relate the known fractional concentration to the one or more calibration values. The fractional concentration of the target tissue type may be used to determine a tissue type and / or a characteristic of an object from which a biological sample is obtained.

[0148] The concentration score can be used to determine the classification of the condition or disease. For example, if the target tissue type is placenta, the method can further include using the concentration score to determine the classification of pregnancy-related conditions or gestational age. The concentration score can be compared with a cutoff value determined from a sample of a reference subject having a certain classification of pregnancy-related conditions or having a certain gestational age. Pregnancy-related conditions can include pre-eclampsia, intrauterine growth restriction, invasive placentation and premature birth, hemolytic disease of the newborn, placental insufficiency, fetal edema, fetal malformations, HELLP syndrome, systemic lupus erythematosus, and other immune diseases of the mother. Pregnancy-related conditions can be related to the fetus or the mother.

[0149] In some embodiments, the classification of cancer grade can be determined using the concentration score.The concentration score can be compared to a cutoff value determined from a sample of a reference subject having a certain classification of cancer grade. a) Concentration fraction of the second target tissue type

[0150] In some embodiments, the concentration fraction of multiple tissue types can be determined. Different tissues can show different histone modification amounts (e.g., as described in part 0) in different genomic regions. Biological samples, such as plasma samples, can have DNA fragments from different tissues. Therefore, DNA fragments can include fragments related to the histone modifications in different genomic regions. Each genomic region can have a sequence motif related to histone modifications. The sequence motifs in different genomic regions can be used to determine the concentration fractions of different tissues in the biological sample. The amount of the sequence motif is related to the concentration fraction of the tissue. The method can be repeated for a second target tissue to determine the concentration fraction of the second target tissue.

[0151] For example, the above steps can be used for a first target tissue type. The one or more genomic regions associated with the first target tissue type can be one or more first genomic regions. The group of sequence reads located in the one or more first genomic regions can be a first group of sequence reads. The histone modification in the one or more first genomic regions can be a first histone modification. The set of one or more sequence motifs can be a set of one or more first sequence motifs. The relative frequency can be a first relative frequency. The total value can be a first total value. The one or more calibration samples can be one or more first calibration samples. The concentration score can be a first concentration score.

[0152] The method may further include identifying a second set of sequence reads located in one or more second genomic regions in a manner similar to block 1720. Each of the one or more second genomic regions may have a second histone modification associated with a second target tissue type. The one or more second genomic regions may be the same or different from the one or more first genomic regions.

[0153] For each sequence read of the second set of sequence reads, one or more second sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments can be determined, similar to block 1730 .

[0154] Similar to block 1740, one or more second relative frequencies of the set of one or more second sequence motifs can be determined. The set of one or more second sequence motifs can appear at a higher rate in chromatin immunoprecipitation sequencing of one or more second genomic regions associated with a second histone modification than in sequencing without chromatin immunoprecipitation. Sequence motifs that appear more frequently in ChIP-sequencing can be used because those sequence motifs can be associated with a second histone modification (similar to Fig.10 ). The following describes the set of determined sequence motifs.

[0155] Similar to block 1750, a second aggregate value of the one or more second relative frequencies may be determined.

[0156] The one or more second aggregate values ​​may be compared to the one or more second calibration values ​​in a manner similar to block 1760 .

[0157] One or more second calibration values ​​can be determined from one or more second calibration samples whose fractional concentrations of DNA fragments from a second target tissue type are known. A second fractional concentration of cell-free DNA fragments from a second target tissue type can be determined using the comparison, similar to block 1770. b) Determine the sequence motif

[0158] A collection of one or more sequence motifs can be represented by Figure 3 , 10 The first ratio of each of the one or more sequence motifs relative to other sequence motifs in the cfChIP sequencing can be determined. The first ratio can be a ranking, such as Fig.10 As shown, or frequency. The frequency can be determined by the ratio of the raw counts of the sequence motifs in the set to the counts outside the set. In sequencing without chromatin immunoprecipitation, a second ratio of each of the one or more sequence motifs in the set relative to other sequence motifs. The second ratio can be of the same type as the first ratio (e.g., rank, frequency). Each of the one or more sequence motifs in the set can be identified as having a first ratio that is higher than the second ratio. The frequency can be determined by using a graphical representation (e.g., Fig.10 ) or by determining the difference between ranks or frequencies (e.g. Fig.11) for identification. Each set of one or more sequence motifs may have a difference above a threshold difference. Sequence motifs not in the set may have a difference below a threshold difference.

[0159] Process 1700 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described elsewhere herein.

[0160] although Fig.17 An exemplary block diagram of process 1700 is shown, but in some implementations, process 1700 may include Fig.17 The blocks shown in the process 1700 may be compared to additional blocks, fewer blocks, different blocks, or blocks in a different arrangement. Additionally or alternatively, two or more blocks of the process 1700 may be performed in parallel. B. Estimation of target tissue characteristic values

[0161] Sequence motifs associated with histone modifications can be used to estimate the values ​​of various features of the target tissue. These features can describe the health of the tissue, the age of the tissue, or the disease level in the tissue. For example, the determined features may include a specific gestational age or range (e.g., 8 weeks, 9-12 weeks). In another example, the determined features may be the length or nutritional status of an organ corresponding to a specific tissue type.

[0162] Fig.18 is a flow chart of an exemplary process 1800 associated with estimating a first value of a characteristic of a target tissue. In some embodiments, Fig.18 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.18 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.18 One or more process blocks of may be performed by one or more components of measurement system 5900, such as measurement 5908, measurement device 5910, detector 5920, logic system 5930, local memory 5935, external memory 5940, storage device 5945, and / or processor 5950. Process 1800 may include aspects described in conjunction with process 1700.

[0163] At block 1810, a plurality of sequence reads of cell-free DNA fragments are received. The plurality of sequence reads include end sequences corresponding to ends of the plurality of cell-free DNA fragments. Block 1810 can be performed in a manner similar to block 1710.

[0164] At block 1820, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions has a histone modification associated with a target tissue type. Block 1820 can be performed in a manner similar to block 1720.

[0165] At block 1830, one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments are determined for each sequence read of the set of sequence reads. Block 1830 can be performed in a manner similar to block 1730.

[0166] At block 1840, one or more relative frequencies of a set of one or more sequence motifs are determined. The set of one or more sequence motifs appears at a higher rate in chromatin immunoprecipitation sequencing (ChIP-seq) for histone modifications associated with one or more genomic regions than in sequencing without chromatin immunoprecipitation. Block 1840 can be performed in a manner similar to block 1740.

[0167] At block 1850 , an aggregate value of one or more relative frequencies is determined. Block 1850 may be performed in a similar manner as block 1750 .

[0168] At block 1860, the aggregate value is compared to one or more calibration values. The one or more calibration values ​​are determined from one or more calibration samples whose values ​​for the characteristic of the target tissue type are known. A machine learning model may be used to make the comparison, which may be any machine learning model described herein. A machine learning model may be used to determine the calibration values.

[0169] One or more calibration values ​​may be determined in the same manner as block 1760, but using calibration samples whose values ​​for the characteristic of the target tissue type are known. For example, the aggregate value determined from the biological sample may be a first aggregate value determined from one or more first relative frequencies. One or more second relative frequencies of a set of one or more sequence motifs in one or more genomic regions may be determined for each calibration sample of the one or more calibration samples. A second aggregate value may be determined for one or more second relative frequencies of each calibration sample in the one or more calibration samples. Thus, each of the one or more second aggregate values ​​may be associated with the value of the characteristic of the calibration sample. The calibration values ​​may include one or more second aggregate values. For example, the calibration value may be a point along a straight line or curve that associates a known value of a characteristic with a second aggregate value.

[0170] At block 1870, the comparison is used to estimate a first value of the characteristic of the target tissue type. The first value of the characteristic may be a known first value associated with a calibration value, which may have a value close to or equal to the aggregate value. In some embodiments, the first value of the characteristic may be determined according to a function or line having one or more calibration values. The function or line may relate the known first value to the one or more calibration values.

[0171] The target tissue type can be liver or hematopoietic cells. The target tissue type can be fetal tissue. In some embodiments, the biological sample can be obtained from a pregnant female subject, and the target tissue type can be placental tissue. In some embodiments, the target tissue type can be an organ suffering from cancer. The target tissue type can be any organ described herein. The feature can be the cancer grade or the nutritional status of the organ. For example, the nutritional status of the organ can be whether the organ is healthy, including any intermediate level of measuring organ health. As another example, the feature can be gestational age. In another example, the determined feature can be the concentration of a specific tissue type (e.g., hepatocytes) relative to the concentration of other tissue types (e.g., hematopoietic cells).

[0172] In some embodiments, process 1800 may include using length frequencies and relative frequencies of sequence motifs. Process 1800 may include measuring the length of cell-free DNA fragments using sequence reads. Process 1800 may also include determining one or more length frequencies of sequence reads of one or more length ranges, which may be any length ranges described herein. A total value of one or more length frequencies may be determined. The total value may be the sum of the length frequencies or any value similar to the total value of the relative frequencies of the sequence motifs. In some embodiments, the total value may be an estimate of histone modifications. The level of histone modifications may be determined by various types of data, such as the amount of end motifs or fragment lengths. The total value of one or more length frequencies may be compared to a calibration value determined using a calibration sample, the characteristic value of the target tissue type of the calibration sample being known. Estimating the first value of a feature may include using a comparison of the total value of the length frequencies, similar to the comparison of the total value of the relative frequencies of the sequence motifs.

[0173] Process 1800 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described herein. C. Measuring the amount of histone modifications

[0174] Sequence motifs can be used to determine the amount of histone modification. Fig.14 , 15AAs shown in Figures 15 and 15B, the motif frequency can be correlated with the cfChIP-seq signal associated with the H3K4me3 signal, which is proportional to the amount of H3K4me3. Therefore, the motif frequency can be correlated with the amount of histone modification. In addition, the amount of histone modification in different regions can be used to determine the concentration fraction of multiple tissues in the same sample. 1. Exemplary Methods for Determining the Amount of Histone Modifications Using Sequence Motifs

[0175] Fig.19 is a flow chart of an exemplary process 1900 associated with determining the amount of a histone modification in one or more genomic regions. In some embodiments, Fig.19 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.19 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.19 One or more process blocks of may be performed by one or more components of measurement system 5900, such as measurement 5908, measurement device 5910, detector 5920, logic system 5930, local memory 5935, external memory 5940, storage device 5945, and / or processor 5950. Process 1900 may include aspects described in conjunction with process 1700.

[0176] At block 1910, a plurality of sequence reads of cell-free DNA fragments are received. The plurality of sequence reads include end sequences corresponding to the ends of the plurality of cell-free DNA fragments. Block 1910 can be performed in a manner similar to block 1710.

[0177] At block 1920, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions has a histone modification associated with a target tissue type. Block 1920 can be performed in a manner similar to block 1720.

[0178] At block 1930, one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments are determined for each sequence read of the set of sequence reads. Block 1930 can be performed in a manner similar to block 1730.

[0179] In block 1940, one or more relative frequencies of a set of one or more sequence motifs are determined. The set of one or more sequence motifs is present at a higher rate in chromatin immunoprecipitation sequencing (ChIP-seq) than in sequencing without chromatin immunoprecipitation in one or more genomic regions associated with histone modifications. Block 1940 can be performed in a manner similar to block 1740.

[0180] At block 1950 , an aggregate value of one or more relative frequencies is determined. Block 1950 may be performed in a similar manner as block 1750 .

[0181] At block 1960, the total value is compared to one or more calibration values. The one or more calibration values ​​are determined from one or more calibration samples, the amount of histone modification of the calibration sample being known. The amount of histone modification in the one or more calibration samples can be known from performing ChIP-sequencing on each of the one or more calibration samples.

[0182] One or more calibration values ​​can be determined in the same manner as box 1760 or box 1860, but using a calibration sample whose amount of histone modification is known. For example, the total value determined from the biological sample can be a first total value determined from one or more first relative frequencies. One or more second relative frequencies of a set of one or more sequence motifs in one or more genomic regions can be determined for each calibration sample of one or more calibration samples. A second total value can be determined for one or more second relative frequencies of each calibration sample in one or more calibration samples. Therefore, each of the one or more second total values ​​can be associated with the amount of histone modification of the calibration sample. The calibration value can include one or more second total values. For example, the calibration value can be a point along a straight line or curve that associates a known value of a feature with a second total value.

[0183] In box 1970, the amount of histone modification in one or more genomic regions is determined using the comparison. The amount of histone modification can be a known amount with a calibration value, and the calibration value can have a value close to or equal to the total value. In some embodiments, the amount of histone modification can be determined by a function or straight line with one or more calibration values. The function or straight line can be related to the known amount of histone modification with one or more calibration values. The amount of histone modification can be in the target tissue type.

[0184] Process 1900 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described herein.

[0185] although Fig.19 An exemplary block diagram of process 1900 is shown, but in some implementations, process 1900 may include Fig.19 The blocks shown in the process 1900 may be fewer blocks, different blocks, or differently arranged blocks than other blocks. Additionally or alternatively, two or more blocks of the process 1900 may be performed in parallel. 2. Exemplary methods using fragmentomic properties

[0186] Fig. 20is a flow chart of an exemplary process 2000 associated with determining the amount of a histone modification in one or more genomic regions. In some embodiments, Fig. 20 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig. 20 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig. 20 One or more process blocks may be performed by one or more components of the measurement system 5900 , such as the determination 5908 , the determination device 5910 , the detector 5920 , the logic system 5930 , the local memory 5935 , the external memory 5940 , the storage device 5945 , and / or the processor 5950 .

[0187] At block 2010, a plurality of sequence reads of cell-free DNA fragments are received. Block 2010 can be performed in a similar manner to block 1710.

[0188] At block 2020, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions has a histone modification associated with a target tissue type. Block 2020 can be performed in a manner similar to block 1720.

[0189] At block 2030, a value of a fragmentomic property corresponding to each cell-free DNA fragment in each sequence read in the set of sequence reads is determined. The fragmentomic property may include fragment length, terminal motif, jagged ends (overhang of one strand relative to another strand), terminal nucleotides, topological form, and / or nucleosome imprint. The fragmentomic property may be any of the fragmentomic properties described herein.

[0190] For example, Fig.19 The fragmentomic property may be a sequence motif corresponding to an end sequence of ends of the cell-free DNA fragments, and the one or more value ranges may be the one or more sequence motifs.

[0191] As another example, the fragmentomic property can be length, and the one or more value ranges are one or more length ranges, as described in Section 0.

[0192] For example, the fragmentomic property can be a topological form, and one or more value ranges are one or more topological forms. The topological form can be circular or linear.

[0193] As an example, the fragmentomic property is a nucleosome footprint, and the one or more value ranges are one or more nucleosome footprints. The nucleosome footprint represents the binding pattern of nucleosomes to genomic DNA. The space between nucleosomes can be the value of the nucleosome footprint.

[0194] In box 2040, determine one or more relative frequencies of cell-free DNA fragments with values ​​of fragment omics properties in a set of one or more value ranges. The set of one or more value ranges appears in chromatin immunoprecipitation sequencing (ChIP-seq) and in sequencing without chromatin immunoprecipitation at different ratios for one or more genomic regions related to histone modification. Different ratios can be higher or lower, and can be statistically significant amounts. Box 2040 can be performed in a manner similar to box 1740, but one or more sequence motifs are replaced using one or more value ranges of fragment omics properties. In other embodiments, the set of one or more value ranges determined by sequencing a sample without cell-free chromatin immunoprecipitation is determined by focusing on a genomic region containing a difference rate, and the genomic region has a higher or lower histone modification signal predetermined from other reference samples or a database.

[0195] At block 2050, an aggregate value of the one or more relative frequencies is determined. The aggregate value may be a sum of the one or more relative frequencies or a statistical measure (eg, mean, median, mode, percentile) of the one or more relative frequencies.

[0196] At block 2060, the total value is compared to one or more calibration values. The one or more calibration values ​​are determined from one or more calibration samples, the amount of histone modification of the calibration samples being known. The amount of histone modification in the one or more calibration samples can be known by performing cfChIP sequencing on each of the one or more calibration samples. The one or more calibration values ​​can be determined in the same manner as block 1960, but the frequencies of one or more value ranges of the fragmentomic properties are used instead of one or more sequence motifs.

[0197] At block 2070, the amount of the histone modification in the biological sample is determined using the comparison. The amount of the histone modification can be in a target tissue type. Block 2070 can be performed in a similar manner to block 1970.

[0198] The amount of histone modification can be used to determine the concentration score of the target tissue, classify the disease stage, or classify the transplantation status of the target tissue type (eg, as described in process 2000).

[0199] although Fig. 20 An exemplary block diagram of process 2000 is shown, but in some implementations, process 2000 may include Fig. 20 The blocks shown in the process 2000 may be additional blocks, fewer blocks, different blocks, or differently arranged blocks. Additionally or alternatively, two or more blocks of the process 2000 may be performed in parallel. 3. Determine the concentration fraction using deconvolution

[0200] The concentration fractions of various tissue types can be determined through a deconvolution process. Fig.21 The application of ChIP-seq to determine the contribution from different tissues is shown. Figure 2104 is a graph of the histone modification signal from ChIP-seq on the y-axis and the genomic position on the x-axis. Figures 2108, 2112, and 2116 show tissue-specific regions of histone modification signals. Figure 2108 shows that region X carries neutrophil-specific histone modifications. Figure 2112 shows that region Y carries liver-specific histone modifications. Figure 2116 shows that region Z carries monocyte-specific histone modifications. By comparing the plasma DNA ChIP signals in these informative genomic regions with the patterns of ChIP signals from different tissues, the proportion of DNA contribution associated with H3K27ac that enters the plasma from different tissues is inferred. Figure 2120 shows the inferred proportion of DNA contribution from different tissues.

[0201] based on Fig.21 , biological samples containing DNA from multiple tissues can have H3K4me3 cfChIP-seq signals in the same region from multiple tissues. For example, the genomic region X shown in Figure 2108 has neutrophils with the highest H3K4me3 signals, but has lower signals in other tissues (e.g., liver and monocytes). Similarly, the genomic region Y shown in Figure 2112 also has different signals on different tissues, including neutrophils, liver and monocytes. The genomic region Z shown in Figure 2116 also has different signals on different tissues, including neutrophils, liver and monocytes. Overlapping H3K4me3 signals in the same region can allow the concentration fraction of the tissue to be determined.

[0202] A system of linear equations, one for each region, can be solved to determine the fractional concentration of each tissue in a cell-free mixture (eg, a plasma sample). H A =f1h 1,A +f2h 2,A +…+f n h n,A H B =f1h 1,B +f2h 2,B +…+f n h n,B H m =f1h 1,m +f2h 2,m +…+f n h n,m A collection of linear equations for m genomic regions and n tissues. A H represents the total amount of histone modification in genomic region A in the sample, which can be measured using one or more sequence motifs. B represents the total histone modification amount in genomic region B. H A and H B Can represent the same or different histone modifications. m represents the total histone modification amount in genomic region m. The concentration fraction of target tissue 1 is f1, the concentration fraction of target tissue 2 is f2, and the concentration fraction of target tissue n is f n . It is known that target tissue 1 has a quantity h in genomic region A 1,A , with quantity h in genomic region B 1,B , and has quantity h in genomic region m 1,m . It is known that target tissue 2 has a quantity h in genomic region A 2,A , with quantity h in genomic region B 2,B , and has quantity h in genomic region m 2,m . It is known that target tissue n has a quantity h in genomic region A n,A , with quantity h in genomic region B n,B , and has quantity h in genomic region m n,m In some embodiments, the matrix H may represent the amount of histone modification measured using one or more sequence motifs. If appropriate sequence motif quantities are used, it may not be necessary to directly calculate H and h to solve for the concentration fractions.

[0203] Certain genomic regions (e.g., h 1,A 、h 1,B The amount of histone modification in the target tissue (e.g., target tissue 1, target tissue 2, etc.) can be a relative amount. These amounts can be determined from calibration samples. For example, a calibration sample with half target tissue 1 and half target tissue 2 can show a certain ratio of the amount of histone modification, and this ratio can be used to calculate the amount of histone modification in the target tissue. 1,A and h 1,B .

[0204] The number of equations should be greater than or equal to the number of target tissues in order to solve for the concentration fraction. The number of equations can be equal to the number of genomic regions, and thus the number of genomic regions can be equal to the number of target tissues. If the sum of the concentration fractions is known (e.g., the sum is 1), then the number of genomic regions can be equal to the number of regions minus 1. Using the amount of histone modification in each genomic region measured by using the sequence motif, the concentration fraction can be determined by solving the system of equations.

[0205] Therefore, in some embodiments, multiple tissue types can have the same or similar sequence motifs associated with histone modifications in the same genomic region. The concentration fraction of each of these multiple tissue types can be determined by a deconvolution process. The deconvolution process can include solving a set of linear or nonlinear equations, such as those described herein.

[0206] The amount of the histone modification can be determined as described in process 1900. In process 1900, the set of sequence reads is a first set of sequence reads. The one or more genomic regions are one or more first genomic regions. The one or more sets of sequence motifs are one or more first sets of sequence motifs. The one or more relative frequencies are one or more first relative frequencies. The total value is a first total value. The one or more calibration values ​​are one or more first calibration values. The amount of the histone modification is a first amount of the histone modification. An example of the first amount is H in the above equation. A .

[0207] A second amount of histone modification in one or more second genomic regions can be determined for the system of linear equations. An example of a second amount is H B The histone modification may be associated with a first tissue type and a second tissue type in one or more first genomic regions.

[0208] The histone modification can be associated with the first tissue type and the second tissue type in the one or more second genomic regions. For example, the one or more first genomic regions can be associated with Fig.21 The one or more first genomic regions may be regions associated with region X in the genomic region, and the one or more second genomic regions may be regions associated with region Y. As another example, the one or more first genomic regions and the one or more second genomic regions may be regions within the same frame (e.g., region X or region Y).

[0209] Identify a second set of sequence reads located in one or more second genomic regions. Identify in a manner similar to that described in block 1920. Each of the one or more second genomic regions may have a histone modification associated with the first tissue type and the second tissue type. In some embodiments, the histone modification in the one or more second genomic regions may have a histone modification different from the histone modification in the one or more first genomic regions.

[0210] For each sequence read of the second set of sequence reads, one or more second sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments are determined. This determination can be performed in a manner similar to that described in block 1930.

[0211] Determine one or more second relative frequencies of a set of one or more second sequence motifs. The set of one or more second sequence motifs has a higher rate of occurrence in ChIP-seq than in sequencing without chromatin immunoprecipitation for one or more second genomic regions associated with histone modification. This determination can be performed in a manner similar to that described in block 1940.

[0212] Determine a second aggregate value of the one or more second relative frequencies. This determination may be performed in a manner similar to that described in block 1950.

[0213] The second aggregate value is compared to one or more second calibration values. The comparison may be performed in a manner similar to block 1960.

[0214] Using the comparison, a second amount of histone modification in one or more second genomic regions is determined. This determination can be made in a manner similar to block 1970.

[0215] The first concentration fraction of the first tissue type and the second concentration fraction of the second tissue type are determined by solving a linear or nonlinear system of equations. The linear system of equations can be a set of equations described herein. The linear system of equations can include a first amount of histone modification (e.g., H A ), a second amount of histone modification (e.g., H B ), and specifying the relative amounts of the corresponding histone modifications for each tissue type in the one or more first genomic regions and the one or more second genomic regions (e.g., h 1,A 、h 1,B 、h 2,A 、h 2,B The first concentration fraction may be f1, and the second concentration fraction may be f2.

[0216] The biological sample may include more than two target tissue types. The method for determining the concentration fraction of two target tissue types may be extended to three or more tissue types.

[0217] In an embodiment, the histone modification can be associated with a third tissue type in one or more first genomic regions and one or more second genomic regions. The histone modification can be associated with a first tissue type, a second tissue type, and a third tissue type in one or more third genomic regions. The process may include performing similar steps as described for the second tissue type. The process may include determining a third amount of histone modification in one or more third genomic regions (e.g., H m, where m is C). The third concentration fraction of the third tissue type can be determined by solving a linear or nonlinear system of equations. The linear system of equations can include parameters for the relative amount of each tissue type in the third amount of histone modification and one or more third genomic regions. D. Classification of disease levels

[0218] Sequence motifs can be used to classify disease levels. The disease can be specific to a particular tissue type or can be applied to an object. Sequence motifs can indicate the amount or presence of a histone modification, and the amount or presence of a histone modification can be associated with a specific level of disease. However, it may not be necessary to determine the amount or presence of a histone modification in order to use a sequence motif to classify the disease level.

[0219] Fig. 22 : is a ROC curve for distinguishing patients with and without hepatocellular carcinoma (HCC) using H3K4me3 signals inferred in liver-specific H3K4me3 regions using terminal motifs. Specificity is shown on the x-axis and sensitivity is shown on the y-axis. Using the plasma H3K4me3 ChIP signal inferred by the terminal motif, there was an AUC of 0.718 for distinguishing between patients with and without HCC using the cutoff value. These results suggest that ChIP signals of histone modifications inferred by terminal motifs will be clinically useful for non-invasive prenatal testing and cancer detection and monitoring.

[0220] Fig.23 is a flow chart of an exemplary process 2300 associated with classifying a disease level. In some embodiments, Fig.23 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.23 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.23 One or more process blocks of may be performed by one or more components of measurement system 5900, such as measurement 5908, measurement device 5910, detector 5920, logic system 5930, local memory 5935, external memory 5940, storage device 5945, and / or processor 5950. Process 2300 may include aspects described in conjunction with process 1700.

[0221] At block 2310, a plurality of sequence reads of cell-free DNA fragments are received. The plurality of sequence reads include end sequences corresponding to ends of the plurality of cell-free DNA fragments. Block 2310 can be performed in a manner similar to block 1710.

[0222] At block 2320, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions has a histone modification associated with one or more target tissue types. Block 2320 can be performed in a manner similar to block 1720.

[0223] At block 2330, one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments are determined for each sequence read of the set of sequence reads. Block 2330 can be performed in a manner similar to block 1730.

[0224] In block 2340, one or more relative frequencies of a set of one or more sequence motifs are determined. The set of one or more sequence motifs is present at a higher rate in chromatin immunoprecipitation sequencing (ChIP-seq) than in sequencing without chromatin immunoprecipitation in one or more genomic regions associated with histone modifications. Block 2340 can be performed in a manner similar to block 1740.

[0225] At block 2350 , an aggregate value of one or more relative frequencies is determined. Block 2350 may be performed in a similar manner as block 1750 .

[0226] At block 2360, the aggregate value is compared to one or more calibration values. The one or more calibration values ​​are determined from one or more calibration samples for which the classification of the disease level is known.

[0227] One or more calibration values ​​can be determined in the same manner as box 1760, box 1860, or box 1960, but using a calibration sample for which the classification of the disease level is known. For example, the total value determined from the biological sample can be a first total value determined from one or more first relative frequencies. One or more second relative frequencies of a set of one or more sequence motifs in one or more genomic regions can be determined for each calibration sample of the one or more calibration samples. A second total value can be determined for one or more second relative frequencies of each calibration sample in the one or more calibration samples. Therefore, each of the one or more second total values ​​can be associated with a classification of the disease level. The calibration value can include one or more second total values. For example, the calibration value can be a point along a straight line or curve that associates a known classification of the disease level with a second total value.

[0228] At block 2370, the comparison is used to determine a classification of the symptom level. The classification of the symptom level may be a known classification with a calibration value, which may have a value close to or equal to the aggregate value. In some embodiments, the classification of the symptom level may be determined based on a function or straight line with one or more calibration values. The function or straight line may relate a known classification to one or more calibration values. In some embodiments, the classification may be an abnormal level.

[0229] The condition may be in the target tissue type. The condition may be a cancer of the target tissue type. Cancer may include hepatocellular carcinoma (HCC), colorectal cancer (CRC), or any cancer described herein. In some embodiments, the condition is a pregnancy-related condition. The condition may be a blood condition. The condition may be any condition described herein.

[0230] In an embodiment, process 2300 may include using length frequency, as described in process 1800 .

[0231] Process 2300 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described herein.

[0232] although Fig.23 An exemplary block diagram of process 2300 is shown, but in some implementations, process 2300 may include Fig.23 The blocks shown in the process 2300 may be additional blocks, fewer blocks, different blocks, or differently arranged blocks than those shown in the process 2300. Additionally, or alternatively, two or more blocks of the process 2300 may be performed in parallel. IV. Using length information to infer histone modifications A. Length information used to infer ChIP signals

[0233] Plasma DNA length information can be used to detect and quantify histone modifications present in plasma DNA molecules. Similar to the relationship between cfDNA terminal motif information and histone modification levels, the length information of cfDNA molecules may be affected by histone modification levels (i.e., epigenetic status). We analyzed the length information of cfDNA molecules in the target region. Those regions involving histone modifications can be grouped into different categories based on the length of the ChIP signal. For example, regions were first sorted based on the length of the ChIP signal and then empirically classified into 9 categories (e.g., Figure 4 ). After obtaining the region categories with different H3K27ac ChIP signals, the DNA length information of the plasma DNA sequencing results without immunoprecipitation in different region categories can be compared.

[0234] Fig.24A ,24B 24C and 24C show the percentage of cfDNA molecules with a certain length in a regional category for different levels of H3K27ac signals. The x-axis is the length of base pairs from plasma DNA sequencing without H3K27ac-based precipitation. The y-axis is the percentage of cfDNA molecules with that length. The different colored lines in each figure show different regional categories using H3K27ac ChIP signals. Fig.24A The length range of 50-140 bp is shown. Fig. 24B A length range of 150-200 bp is shown. Fig.24C The length range of 250-350 bp is shown. Figures 24A-24C As shown, the length distribution varies according to the intensity of the ChIP signal. For example, the higher the ChIP signal, the more DNA molecules are observed within the range of 270-300bp. In addition, for different length ranges, the length difference shows different trends. For example, for a length of about 165-200bp, the higher the ChIP signal, the fewer DNA molecules. For a length of about 60-100bp, the higher the ChIP signal, the more DNA molecules. Therefore, it is feasible to infer plasma DNA histone modifications based on the length information of plasma DNA molecules that are not immunoprecipitated.

[0235] Fig.25A , 25B The correlation between length and ChIP signal of histone modification shown by 25C can be generalized to other histone modifications (e.g. H3K27ac). The x-axis is the cumulative length frequency of certain length fragments as a percentage of plasma DNA that was not immunoprecipitated. The y-axis is log 10 H3K27ac ChIP signal on scale. Fig.25A It is a fragment of 50-140bp. Fig.25B It is a fragment of 150-200bp. Fig.25C Among the nine categories, for plasma DNA without immunoprecipitation, the percentage of cfDNA molecules in the length range of 50-140 bp and 250-350 bp was positively correlated with the log-transformed ChIP signal obtained from ChIP-seq data, with Pearson's r of 0.99 (P value: <0.0001) ( Fig.25A ) and 0.99 (P value: <0.0001) ( Fig.25C The percentage of cfDNA molecules in the 150-200 bp length range was negatively correlated with the log-transformed ChIP signal (Pearson's r: -0.99; P value: <0.0001) ( Fig.25B ).

[0236] Fig.26A and Fig.26B Describes the use of length information to infer plasma DNA histone modifications in plasma DNA sequencing results without immunoprecipitation. Fig.26A The recalibration formula is shown using the percentage of cfDNA molecules of a certain length range and the level of H3K4me3 ChIP signal in 9 categories. Stage 2604 shows that regions involved in H3K4me3 are grouped into different categories based on the length of the ChIP signal. Fig.26A As shown, the length information derived from each region category (e.g., the percentage of cfDNA molecules in the length range of 250-350bp from plasma DNA sequencing results without immunoprecipitation) can be used to determine the correlation with the H3K4me3 ChIP signal. In stage 2608, based on the correlation between fragment length and ChIP signal (on a log scale), a recalibration formula can be determined. A linear formula is shown as an example of a recalibration formula, but a nonlinear formula can also be used.

[0237] Fig.26B Shown is how to use the recalibration formula to infer ChIP signals (i.e., inferred ChIP signals) in other regions (e.g., placenta-specific H3K4me3 regions) based on the corresponding length information of those regions. In stage 2612, plasma DNA is sequenced without immunoprecipitation. In stage 2616, molecules derived from tissue-specific (e.g., placenta) H3K4me3 regions are identified. In stage 2620, the percentage of molecules within a specific length range (e.g., 250-350bp) in immunoprecipitated plasma DNA based on H3K4me3 is calculated. The length information is input into the recalibration formula, and in stage 2624, H3K4me3 ChIP signals are inferred in tissue-specific (e.g., placenta) H3K4me3 regions.

[0238] Fig.27A , 27B Figure 27C shows the correlation between the percentage of cfDNA molecules in a range of lengths and the log transformed H3K4me3 ChIP signal. The x-axis is the cumulative length frequency of certain length fragments as a percentage of plasma DNA that was not immunoprecipitated. The y-axis is the H3K4me3 ChIP signal on a log10 scale. Fig.27A It is a fragment of 50-140bp. Fig.27B It is a fragment of 150-200bp. Fig.27CAmong these nine categories, for plasma DNA without immunoprecipitation, the percentage of cfDNA molecules in the length range of 50-140 bp and 250-350 bp was positively correlated with the log-transformed ChIP signal obtained from ChIP-seq data, with a Pearson's r of 0.99 (P value: <0.0001) ( Fig.27A ) and 0.99 (P value: <0.0001) ( Fig.27C The percentage of cfDNA molecules in the 150-200 bp length range was negatively correlated with the log-transformed ChIP signal (Pearson's r: -0.99; p-value: <0.0001) ( Fig.27B The results show that the fragment length pattern can be used to infer histone modifications in plasma DNA molecules (called inferred ChIP signal). B. Inferred ChIP signal and fetal fraction

[0239] We also use linear regression models to build models (i.e., recalibration formulas) for inferring H3K4me3ChIP signals in a region of interest or a set of regions of interest. As an example, we trained a model for inferring each sample of ChIP signals based on a length range of 250-350bp, i.e., Y=aX+b, where 'Y' represents a logarithmic transformed ChIP signal, and 'X' represents the percentage of cfDNA molecules in a length range of 250-350bp from a specific genomic region of interest or a set of regions of interest to be determined for histone modification. 'a' and 'b' are slope and intercept, respectively. In one embodiment, according to H3K4me3, we determined the percentage of cfDNA molecules in a length range of 250-350bp from those placenta-specific regions. We analyzed 30 plasma DNA samples from pregnant women. For illustration purposes, a length range of 250-350bp was selected. Other length ranges can also be used. A machine learning model can be used to select a length range.

[0240] Fig.28A and 28B Figure 2 shows an evaluation of the performance of inferred H3K4me3 ChIP signals in placenta-specific H3K4me3 regions for inferring fetal DNA fraction. The x-axis shows the fetal DNA fraction as a percentage, as determined by the SNP-based method. Fig.28A In the figure, the y-axis is the H3K4me3 ChIP signal inferred using the length range of 250-350 bp. Using the length metric, the inferred H3K4me3 ChIP signal correlated with the fetal DNA fraction (Pearson's r: 0.62; P value: <0.0001).

[0241] exist Fig.28BIn the y-axis, the cumulative length frequency of 250-350 bp fragments as a percentage. There was no significant correlation between the percentages of plasma DNA in the 250-350 bp length range (Pearson's r: -0.31, P value: 0.096). Fig.28A and 28B These results in suggest that, for plasma DNA samples without immunoprecipitation, the tissue of origin of plasma DNA molecules can be analyzed using putative ChIP signals.

[0242] We also use linear regression models to build models (i.e., recalibration formulas) for inferring H3K27ac ChIP signals in a region of interest or a set of regions of interest. As an example, we trained a model for inferring each sample of ChIP signals based on a length range of 250-350bp, i.e., Y=aX+b, where 'Y' represents a log-converted ChIP signal, and 'X' represents the percentage of cfDNA molecules in a length range of 250-350bp from a specific genomic region of interest or a set of regions of interest to be determined for histone modification. 'a' and 'b' represent slope and intercept, respectively. In one embodiment, according to H3K4me3, we determined the percentage of cfDNA molecules in a length range of 250-350bp from those placenta-specific regions. In one embodiment, according to H3K27ac, we determined the percentage of cfDNA molecules in a length range of 250-350bp from those placenta-specific regions. We analyzed 30 plasma DNA samples from pregnant women.

[0243] Fig.29 is a graph evaluating the performance of inferred H3K27ac ChIP signals in placenta-specific H3K27ac regions for determining fetal DNA fraction. The x-axis is the fetal DNA fraction as a percentage determined by the SNP-based method. The y-axis is the inferred H3K27ac ChIP signal using a length range of 250-350 bp. Based on this length metric, the inferred H3K27ac ChIP signal has a higher correlation with the fetal DNA fraction (Pearson's r: 0.95; P value: <0.0001) compared to H3K4me3-based analysis (Pearson's r: 0.62; P value: <0.0001) ( Fig.28A ). These results highlight that different types of histone modifications can be used to determine the tissue of origin of plasma DNA molecules through ChIP signals of histone modifications inferred from cfDNA length information.

[0244] We analyzed different length ranges for inferring H3K27ac ChIP signals and correlated the inferred H3K27ac ChIP signals with tissue DNA fractions determined by SNP-based methods. We analyzed 30 plasma DNA samples from pregnant women. The length ranges of 50-150 bp, 160-225 bp, and 230-350 bp are used for illustrative purposes. Other length ranges may also be used in some other embodiments.

[0245] Fig.30 is a graph showing how well the length ranges correlate with the fetal DNA fraction with and without considering the histone modification levels. The y-axis shows the three length ranges tested. The x-axis shows the Pearson correlation coefficient. For each length range, two different bars are shown. The top bar in each pair (grey) shows the Pearson correlation coefficient for using the original length frequencies. The bottom bar in each pair (black) shows the Pearson correlation coefficient for the inferred H3K27ac signal levels in the placenta-specific H3K27ac regions.

[0246] like Fig.30 As shown, the fetal DNA fraction determined by the SNP-based method is highly correlated with the inferred H3K27ac signal level in the placenta-specific H3K27ac region, which ranges from 230 to 350 bp in length (Pearson's r: 0.96; P value: <0.0001). In contrast, no such correlation was observed with the raw cumulative length frequency itself (Pearson's r: -0.25; P value = 0.18). Other length ranges were also compared. For all length ranges tested, the inferred H3K27ac levels in the placenta-specific H3K27ac region showed a significantly higher correlation with the fetal DNA fraction (Pearson's r: 0.76 to 0.96) compared to the corresponding raw cumulative length frequency (Pearson's r: -0.25 to 0.53). Furthermore, the H3K27ac ChIP signals inferred based on molecules with a length range of 230-350 bp showed the best performance (Pearson's r = 0.96) compared to the other tested length ranges (Pearson's r: 0.76). C. Inferred ChIP Signaling and Cancer

[0247] In one embodiment, we explored whether ChIP signals inferred from histone modifications of plasma DNA without immunoprecipitation would be informative for cancer detection. We analyzed samples from 34 hepatocellular carcinoma (HCC) patients, 17 chronic hepatitis B virus (HBV) subjects, and 8 healthy controls.

[0248] Fig.31A and Fig.31Bis a graph showing HCC detection using inferred H3K4me3 ChIP signals based on liver-specific H3K4me3 regions. The H3K4me3 ChIP signals were inferred using the cumulative frequency of molecules ranging in length from 250 to 350 bp. Fig.31A Box plots showing inferred H3K4me3 ChIP signals (y-axis) versus subject type (x-axis). For liver-specific regions, inferred H3K4me3 ChIP signals were significantly higher in subjects with HCC (median: 0.21; range: 0-2.90) compared to subjects without HCC (median: 0.09; range: 0-5.36) (P value: 0.015, Mann-Whitney U test).

[0249] Fig.31B is a receiver operating characteristic (ROC) curve. ROC analysis showed that an AUC of 0.686 could be achieved in distinguishing subjects with and without HCC. These results suggest that the inferred ChIP signal can be used for cancer detection. This method will avoid the need for immunoprecipitation analysis prior to sequencing, thereby reducing costs and experimental time, and making it easy to combine with other techniques, such as whole genome random or targeted sequencing, or whole genome random or targeted bisulfite sequencing.

[0250] Fig.32A and Fig.32B Figure 2 shows the use of inferred H3K27ac ChIP signals based on H3K27ac regions for HCC detection. The H3K27ac ChIP signals were inferred using the cumulative frequency of molecules with a length range of 250-350 bp. In addition to using H3K27ac regions instead of H3K4me3 regions, Fig.32A and Fig.32B Respectively Fig.31A and Fig.31B When distinguishing patients with and without HCC, using the inferred ChIP signal associated with H3K27ac improved the classification ability, increasing the AUC from 0.686 ( Fig.31B ) increases to 0.738( Fig.32B ).

[0251] Fig.33 is a graph showing how length selection affects the performance of distinguishing cancer patients from healthy controls. Fig.33 is a ROC curve with sensitivity on the y-axis and specificity on the x-axis. The ROC curve is used to distinguish subjects with intermediate and advanced hepatocellular carcinoma (HCC) from subjects without HCC by inferring H3K27acChIP signals in liver-specific regions. The black line refers to molecules with a length range of 230-350 bp. The gray line refers to molecules with a length range of 50-150 bp.

[0252] ROC analysis showed that the H3K27ac ChIP signals inferred using the cumulative frequency of molecules in the length range of 230-350 bp in the liver-specific H3K27ac region achieved a significantly higher area under the receiver operating characteristic curve (AUC) of 0.934 for distinguishing patients with HCC in intermediate and advanced stages from those without HCC, compared with those in the length range of 50-150 bp (AUC: 0.586) (P = 0.001; Delong test). D. Inferred ChIP Signals and Transplantation

[0253] Fig.34 is a graph showing the correlation between the inferred H3K27ac ChIP signal and the donor DNA fraction in the liver-specific H3K27ac region. The H3K27ac ChIP signal was inferred using the cumulative frequency of molecules with a length range of 250-350bp. The y-axis shows the inferred H3K27ac ChIP signal. The x-axis shows the donor DNA fraction as a percentage. We used the liver-specific region to infer the H3K27ac ChIP signal in the plasma DNA of liver transplant patients. The figure shows a high correlation between the liver contribution determined by the inferred ChIP signal of histone modification in the liver-specific region according to an embodiment of the present disclosure and the donor DNA fraction by a SNP-based method (Pearson's r: 0.9; P value: <0.0001). The data show that the inferred H3K27ac ChIP signal in the liver-specific region can allow monitoring of subjects for organ transplantation.

[0254] We further analyzed the plasma DNA sequencing results without immunoprecipitation of the group of 14 liver transplant patients. The length ranges of 50-150bp, 160-225bp and 230-350bp are used for illustrative purposes. Other length ranges can also be used in some other embodiment schemes.

[0255] Fig.35 is a graph showing how well the length ranges are associated with the donor DNA fraction determined by the SNP-based method with and without considering the histone modification levels. The y-axis shows the three length ranges tested. The x-axis shows the Pearson correlation coefficient. For each length range, two different bars are shown. The top bar (grey) in each pair shows the Pearson correlation coefficient for using the original length frequency. The bottom bar (black) in each pair shows the Pearson correlation coefficient for the inferred H3K27ac signal level in the liver-specific H3K27ac region.

[0256] like Fig.35As shown, the highest correlation was observed between donor DNA fractions and H3K27ac values ​​inferred by those fractions in the length range of 230-350 bp in liver-specific H3K27ac regions (Pearson's r: 0.91; P value: <0.0001). E. Exemplary Methods for Using Length to Determine Histone Modifications

[0257] Fig.36 is a flow chart of an exemplary process 3600 associated with using fragment length to determine the amount of histone modifications in one or more genomic regions. In some embodiments, Fig.36 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.36 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.36 One or more process blocks may be performed by one or more components of the measurement system 5900 , such as the determination 5908 , the determination device 5910 , the detector 5920 , the logic system 5930 , the local memory 5935 , the external memory 5940 , the storage device 5945 , and / or the processor 5950 .

[0258] At block 3610, a plurality of sequence reads of the cell-free DNA fragments are received. The plurality of sequence reads can be obtained by random massively parallel sequencing. The plurality of sequence reads can be obtained using paired-end sequencing.

[0259] At block 3620, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions has a histone modification associated with one or more target tissue types. Block 3620 can be performed in a manner similar to block 1720.

[0260] At block 3630, the length of each cell-free DNA fragment corresponding to each sequence read in the set of sequence reads is measured. The length of the fragment can be measured using paired-end sequencing, aligning the sequence to the genome, and then inferring the length from the genomic coordinates of the aligned sequence. In some embodiments, the length of the fragment can be measured by sequencing the entire fragment and then determining the length from the sequence.

[0261] In box 3640, determine the one or more relative frequencies of cell-free DNA fragments of length in the set of one or more length ranges.The set of one or more length ranges appears in chromatin immunoprecipitation sequencing (ChIP-seq) and in the sequencing without chromatin immunoprecipitation with different ratios for one or more genomic regions related to histone modification.Different ratios can be higher or lower, and can be statistically significant amounts.One or more length ranges can include 50 to 100bp, 100 to 150bp, 150 to 200bp, 200 to 250bp, 250 to 300bp, 300 to 350bp, 350 to 400bp, 400 to 450bp, 450 to 500bp, more than 500bp, or any combination thereof.

[0262] At block 3650, an aggregate value of the one or more relative frequencies is determined. The aggregate value may be a sum of the one or more relative frequencies or a statistical measure (eg, mean, median, mode, percentile) of the one or more relative frequencies.

[0263] At block 3660, the total value is compared to one or more calibration values. The one or more calibration values ​​are determined from one or more calibration samples, the amount of histone modification of the calibration sample being known. The amount of histone modification in the one or more calibration samples can be known by performing cfChIP sequencing on each of the one or more calibration samples. The one or more calibration values ​​can be determined in the same manner as block 1960, but the frequencies of one or more length ranges are used instead of one or more sequence motifs.

[0264] At block 3670, the amount of the histone modification in the biological sample is determined using the comparison. The amount of the histone modification can be in a target tissue type. Block 3670 can be performed in a similar manner to block 1970.

[0265] The amount of histone modification can be used to determine the concentration fraction of the target tissue, the classification of the disease stage, or the classification of the transplantation status of the target tissue type. The amount of histone modification can be determined using sequence motifs, fragmentomic properties or any other technology (including length range).

[0266] In some embodiments, the amount of histone modification can be compared with one or more second calibration values. One or more second calibration values ​​can be determined from one or more second calibration samples, the concentration fraction of the target tissue type of the second calibration sample and the amount of histone modification are known. The concentration fraction of the target tissue type can be determined using a comparison of the amount of histone modification with one or more second calibration values.

[0267] In some embodiments, the amount of histone modification can be compared with one or more third calibration values. One or more third calibration values ​​can be determined from one or more third calibration samples, and the disease grade and the amount of histone modification of the third calibration sample are known. One or more third calibration values ​​are used to determine the classification of the disease grade. The disease can be any disease described herein.

[0268] In some embodiments, the amount of histone modification is compared to one or more fourth calibration values. One or more fourth calibration values ​​can be determined from one or more fourth calibration samples, the transplantation status of the fourth calibration sample and the amount of histone modification are known. One or more fourth calibration values ​​are used to determine the classification of the transplantation status of the target tissue type. The classification of transplantation status includes whether the transplanted organ is rejected by the object.

[0269] although Fig.36 An exemplary block diagram of process 3600 is shown, but in some implementations, process 3600 may include Fig.36 The blocks shown in the process 3600 may be fewer blocks, different blocks, or differently arranged blocks than other blocks. Additionally or alternatively, two or more blocks of the process 3600 may be performed in parallel. V. Organizational contributions inferred from histone modifications

[0270] The characteristic length distribution of cfDNA shows a pattern frequency at approximately 166 bp, with smaller molecules forming a series of peaks at a period of 10 bp (Lo et al. Sci Transl Med. 2010; 2: 61ra91). This length pattern of plasma DNA fragments suggests the presence of histones bound to cfDNA molecules. A recent study revealed the presence of histone modifications associated with cfDNA molecules in plasma using cell-free chromatin immunoprecipitation followed by sequencing (cfChIP-seq) (Sadeh et al. Nat Biotechnol. 2021; 39: 586-598). However, the study by Sadeh et al. did not provide any method to infer the percentage contribution of chromatin modifications from various tissues / organs.

[0271] Sadeh et al. analyzed the average number of reads per kilobase in genomic regions associated with tissue-specific histone modifications of a tissue as a signal to indicate the contribution from that tissue. When analyzing those signals, the tissue-specific regions inferred from the reference tissue were considered independent factors (Sadeh et al., 2021). One limitation of the method described by Sadeh et al. is that the DNA contribution from a tissue cannot be accurately inferred when the tissue lacks tissue-specific histone modifications or the number of regions in the tissue that show tissue-specific histone modifications is insufficient. Sadeh's method relies on the absolute signals of histone modifications associated with tissue-specific regions in plasma. However, Sadeh et al. did not take into account the relative intensity of the signals of histone modifications in each reference tissue in this method, which is likely to lead to inaccurate or impossible analysis.

[0272] For example, the number of reads per kilobase in a genomic region associated with a tissue's histone modification can be controlled by at least two factors: the first factor is the percentage of DNA (including DNA unrelated to histone modification) contributed by this tissue, and the second factor is the level of histone modification present in the tissue. Analysis regulated by the level of histone modification present in the tissue is important for tissue contribution analysis based on histone modification. Sadeh et al. attempted to use linear regression to analyze the percentage contribution from the liver. Plasma DNA from healthy subjects was considered to have a 0% liver contribution, and DNA from liver tissue was considered to have a 100% liver contribution. The difference in histone modifications between liver tissue and plasma DNA from healthy subjects was used to determine the liver contribution in other plasma DNA samples (Sadeh et al. Nat Biotechnol. 2021). This analysis does not use histone modification signals from two or more tissues. Plasma DNA includes contributions from various tissues, and the contribution of the liver to plasma can vary with healthy subjects. Therefore, the assumptions of linear regression analysis may not hold in this case.

[0273] Therefore, the contribution from two or more tissues to be analyzed cannot be accurately inferred in the method of Sadeh et al. The intensity of the histone modification signal from each tissue is important in the quantitative analysis of the signal present in plasma cfDNA. The intensity of the histone modification signal can refer to the percentage of cells in a tissue that carry the histone modification of interest, which can be measured by the depth of sequencing read coverage present in ChIP-seq. By not using the signal of histone modifications across different tissues, this method will greatly reduce the performance in determining the contribution of cfDNA with histone modifications to plasma from different tissues.

[0274] In the present disclosure, we have developed a method for comparing the relative signal of histone modification in plasma DNA with the signal from a reference tissue to infer the percentage contribution of each cell type or tissue, referred to herein as plasma DNA tissue imaging by histone modification. In one embodiment, this comparison will consider the signal of modified histones from various tissues as a covariate to deconvolute the percentage contribution of various tissues to plasma, such as but not limited to using quadratic programming, non-negative least squares (NNLS), etc. Sun et al. demonstrated that comparing the methylation signal of plasma DNA with the methylation signal of various tissues allows to infer the percentage contribution of DNA molecules across tissues to plasma by using quadratic programming (Sun et al., Proc Natl Acad Sci USA. 2018; 115: E5106). However, histone modification will occur at the amino acid sequence of histone protein, where the signal properties of the modified signal are different from DNA methylation. The signal processing process in DNA methylation analysis cannot be used for modified histones. Histone modification involves post-translational modification of histone proteins, which affects its interaction with DNA. In contrast, DNA methylation is a biochemical process in which a DNA base, typically cytosine, is enzymatically methylated at the 5-carbon position. Histone modifications and methylation involve different types of biochemical mechanisms. In some embodiments of the present disclosure, the contribution of histone modifications to plasma can be inferred by comparing the number of DNA immunoprecipitated via one or more antibodies of interest with corresponding measurements across various reference tissues. Compared to the methods used in the studies of Sadeh et al. (only tissue-specific histone modifications are informative), the methods present in the present disclosure can utilize both tissue-specific histone modifications and tissue-variable histone modifications. A. Imaging plasma DNA organization through histone modifications

[0275] In embodiments, the contribution percentage of DNA from various cell types to plasma can be determined by comparing the distribution of plasma DNA histone modifications with the distribution of histone modifications derived from various organs, tissues or cells. For example, H3K27ac ChIP-seq can be applied to various tissues, including but not limited to neutrophils, megakaryocytes, T cells, B cells, erythrocytes, monocytes, natural killer cells, or cells from liver, colon, adipose tissue, brain, pancreas, placenta, heart, lung, kidney, spleen, bladder, stomach, etc. The information type genomic region (e.g., H3K27ac) carrying tissue-specific histone modifications can be determined. The information type genomic region refers to a region that preferentially enriches for a specific histone modification (e.g., H3K27ac) in a specific tissue (e.g., liver) but is relatively lacking in such a modification in other tissues. Such a region can be referred to as a tissue-specific histone modification region (e.g., tissue-specific H3K27ac region). In some embodiments, the information type genomic region refers to a region that displays variable signals of certain histone modifications (e.g., H3K27ac) across a tissue of interest. The variable signal can be defined by a coefficient of variation (CV) of the histone signal exceeding, but not limited to, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%, 200%, etc., and the difference between the maximum and minimum values ​​of the modified histone signal exceeds a certain cutoff value, such as, but not limited to, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 50, 100, 500, 1000, 5000, 10,000 reads / kilobase, etc. Such regions can be defined as tissue variable histone modification regions (e.g., tissue variable H3K27ac regions). As previously described, Figure 3 The application of ChIP-seq to determine the contribution from different tissues is shown.

[0276] Since different pathological or physiological states will alter the chromatin state in certain cell types, we speculate that analysis of histone modifications on cfDNA molecules will allow non-invasive detection and monitoring of diseases such as fetal abnormalities in pregnant women, cancer, autoimmune diseases, presence of transplant rejection, hematological disorders, etc. B. Examples of plasma DNA organization imaging by histone modifications

[0277] The inferred histone modification signature can be used to determine the fetal DNA fraction, determine specific tissue contributions to a sample, classify subjects as pregnant or non-pregnant, and classify subjects as likely to have a condition such as cancer. 1. Biological samples from pregnant women

[0278] We recruited 19 pregnancy samples with a median gestational age of 38 weeks. Plasma was separated from whole blood by consecutive centrifugation steps within 6 hours of sample collection: centrifugation at 1,600g for 10 minutes, and then centrifugation of the plasma portion at 16,000g for another 10 minutes. Plasma can be stored at -80°C. We used two types of histone modifications (H3K27ac and H3K4me3) as examples. Antibody-conjugated magnetic beads were incubated with plasma by rotating overnight at 4°C and washing with wash buffer, and the immunoprecipitated DNA was connected to the barcode adapter on the beads. The DNA was eluted and then amplified by PCR. The DNA library was sequenced by Illumina platform (e.g., Nextseq500 or NovaSeq 6000) with several other libraries in multiplex sequencing, with a median of 4.3 million paired-end reads (range: 0.10-30.73). We performed H3K27ac ChIP-seq on 19 pregnant samples, 13 non-pregnant samples, and 12 samples with blood diseases (10 samples of β-thalassemia major, 1 sample of iron deficiency anemia, and 1 sample of aplastic anemia). In addition, we performed H3K4me3 ChIP-seq on 12 pregnant women, 4 non-pregnant healthy subjects, and 4 patients with blood diseases (2 samples of β-thalassemia major, 1 sample of iron deficiency anemia, and 1 sample of aplastic anemia).

[0279] The fetal DNA fraction in maternal plasma of each pregnant woman was calculated based on a method based on single nucleotide polymorphism (SNP) (Lo et al. Sci Transl Med. 2010; 2: 61ra91). Genotyping technology based on microarrays (Illumina Infinium Omni 2.5-8 array) was used to obtain the genotype of maternal buffy coat and placental tissue samples, and informative SNPs were identified (i.e., where the mother is homozygous (expressed as AA genotype) and the fetus is heterozygous (expressed as AB genotype)). Fetal-specific DNA fragments were identified based on DNA fragments carrying fetal-specific alleles at informative SNP sites. In this case, the B allele is fetal-specific, and it is inferred that the DNA fragments carrying the B allele are derived from fetal tissue. The number of fetal-specific molecules (p) carrying fetal-specific alleles (B) was determined. The number of molecules (q) carrying common alleles (A) was determined. The fetal DNA fraction of all cell-free DNA samples was calculated by 2p / (p+q)*100%.

[0280] For the purpose of illustration, ChIP-seq data of various tissues were obtained from public databases. The public databases used here include, but are not limited to, the Blueprint project (blueprint-epigenome.eu / ), the ENCODE project (encodeproject.org / ), and the Roadmap project (road_pepigenomics.org / ). In summary, we obtained H3K27ac ChIP-seq results from 18 tissue types, including but not limited to neutrophils, monocytes, B cells, T cells, natural killer cells, erythroblasts and megakaryocytes, liver, brain, pancreas, placenta, heart, colon, lung, fat, kidney, spleen and bladder), with a median of 22.5 million paired end / single end reads (range: 12-45 million). Additionally, we obtained H3K4me3 ChIP-seq data from 19 tissues including, but not limited to, neutrophils, monocytes, B cells, T cells, natural killer cells, erythroblasts, megakaryocytes, liver, brain, pancreas, placenta, heart, colon, lung, adipose, kidney, spleen, bladder, and stomach with a median of 25 million paired-end reads (range: 7–32 million).

[0281] Based on ChIP-seq data from various tissues, we determined the informative genomic regions carrying tissue-specific histone modifications. In one embodiment, many genomic regions known to be rich in specific types of histone modifications were analyzed. For example, it is known that H3K4me3 preferentially occurs in regions near the transcription start site (i.e., promoter regions). Therefore, the ChIP signal across the region near the transcription start site (TSS) can be determined. In one embodiment, the ChIP signal of the region of interest can be determined by the percentage of sequencing reads overlapping with such a region in the readings of all images. In another embodiment, the ChIP signal of the region of interest can be determined by the percentage of sequencing reads overlapping with such a region in the total image readings associated with all regions of interest. The ChIP signal will be adjusted by GC bias and mapping bias, expressed as the number of analysis fragments per million reads (i.e., FPKM) per kilobase.

[0282] In one embodiment, based on the ChIP signals identified from multiple tissues / organs, the human reference genome will be classified into regions where certain histone modifications (e.g., H3K27ac) are present (denoted as regions of interest [ROI]), and regions where such histone modifications are not present (denoted as background regions). ChIP-seq readings of plasma DNA present in background regions may be due to nonspecific antibody (Ab) binding during the experiment, which is considered background noise. The raw ChIP signal of the ROI is determined as the number of fragments whose 5' ends fall within the ROI. In some embodiments, the raw ChIP signal of the ROI is determined as the number of fragments in which at least one or more nucleotides in the molecule overlap with the ROI. The raw signal of the ROI can be deducted from the background noise on the background region surrounding the ROI being interrogated.

[0283] Taking H3K27ac as an example, we divide the genome into non-overlapping 5-Mb windows. For each 5-Mb window, we calculate the original signal in the ROI (N regions) bound by H3K27ac according to the ChIP results shown in the ENCODE and Blueprint projects. The remaining area (M region) is considered to be the background area for determining the noise. Poisson distribution can be used to estimate the average sequence depth of each thousand bases (kb) on the M background regions, which is called the estimated background noise. The original ChIP signal on the N ROIs deducting the estimated background noise (i.e., the ChIP signal deducted by noise) will be used for downstream analysis. In order to minimize the impact of sequencing depth on the comparison of ChIP signals between samples, we use sequencing reads to determine the proportional factor of sequencing depth between samples, and the sequencing reads are from those regions that are shown to be bound by H3K27ac on different samples. The ChIP signal deducted by noise will be adjusted by the proportional factor of the corresponding sequencing depth. In one embodiment, the above-mentioned ChIP signal can also be expressed as fragments per million (FPKM) per thousand bases. In some embodiments, for background noise estimation, multiple overlapping windows can be used. The window length can be, but is not limited to, 10kb, 50kb, 100kb, 500kb, 1Mb, 2Mb, 3Mb, 4Mb, 5Mb, 10Mb, etc.

[0284] Regions carrying tissue-specific histone modifications (i.e., tissue-specific regions) can be identified using the following criteria: 1. The first tissue of interest has the highest ChIP signal for each tissue-specific region of all tissues analyzed and its normalized ChIP signal is greater than 15. 2. The ratio of the ChIP signal in the log2 scale at the tissue-specific region between the first tissue and the second tissue with the second highest ChIP signal is greater than 3. As a result, we identified a total of 4,245 tissue-specific regions for H3K27ac, and 807 tissue-specific regions for H3K4me3.

[0285] Fig.37 Table showing tissue-specific histone modification regions. The first column lists the tissue / cell type. The second column lists the number of regions of that tissue showing H3K27ac modification. The third column lists the number of regions of that tissue showing H3K4me3 modification.

[0286] In one embodiment, the selected region need not be limited to tissue-specific regions. Regions showing high variability of histone modification signals on the tissue group of interest (tissue polymorphic regions) can be used for analysis. The following criteria can be used to determine these regions: 1. The first tissue of interest has the highest ChIP signal for each region analyzed across all tissues, where the normalized ChIP signal is greater than 15. As described above, normalization can take into account background noise, sequencing depth, GC bias, and length of the ROI. The normalized ChIP signal can be expressed as fragments per million (i.e., FPKM) per kilobase. 2. The relative percentage difference between the highest (denoted as H) and lowest (denoted as L) ChIP signals across all tissue types was required to be at least 20% (ie, (HL) / L*100%>20%). 3. The coefficient of variation (CV) of the ChIP signal across all tissue types was required to be at least 25%, where CV is defined as the ratio of the standard deviation to the mean multiplied by 100%. As a result, we identified a total of 27,941 tissue-variable regions for H3K27ac, and 17,321 tissue-variable regions for H3K4me3.

[0287] For the plasma ChIP-seq data for H3K27ac, the number of plasma DNA fragments whose 5' ends were derived from each tissue-specific region of H3K27ac was determined. The normalized ChIP signal in FPKM for each tissue-specific region was calculated accordingly. Similarly, for the plasma DNA ChIP data for H3K4me3, the number of plasma DNA fragments whose 5' ends were derived from each tissue-specific region of H3K4me3 was determined. The normalized ChIP signal for each tissue-specific region was calculated accordingly. Comparing the ChIP signal of plasma DNA with the ChIP signals from various tissues allows us to infer the DNA contribution in the plasma DNA pool associated with the histone modification of interest.

[0288] In one embodiment, the measured ChIP signal levels of DNA molecules are recorded in a vector (X) and the retrieved reference ChIP signal levels across different tissues are recorded in a matrix (M). Through quadratic programming, the proportional contribution of different tissues to the plasma DNA pool is inferred (P): in represents the ChIP signal level of tissue-specific region or tissue-variable region i in plasma DNA mixture; p k M represents the proportion of DNA associated with cell type k to the plasma DNA mixture; ik represents the ChIP signal level of tissue-specific region or tissue-variable region i in cell type k. When the number of regions is equal to or greater than the number of cell types, the individual p k The value of .

[0289] The contribution of aggregated DNA associated with a specific type of histone modification from all cell types will be constrained to 100%: ∑ k p k =100%, Furthermore, any contribution from a cell type will need to be non-negative: Therefore, p can be inferred by quadratic programming using, but not limited to, programs written in Python (python.org) or R (r-project.org). k In some other embodiments, linear or nonlinear regression, non-negative least squares, Bayesian frameworks, etc. may be used but are not limited to. In some embodiments, the regions used for tissue contribution inference may be only tissue-specific regions, or only tissue-variable regions, or a combination of tissue-specific regions and tissue-variable regions.

[0290] Fig.38A graph showing the percentage contribution of different tissues of pregnant and non-pregnant samples based on H3K4me3 histone modification of cfDNA is shown. The x-axis shows the tissue type. The y-axis represents the percentage contribution inferred by H3K4me3. Tissue types may include, but are not limited to, neutrophils, monocytes, B cells, T cells, natural killer cells, erythroblasts, megakaryocytes, liver, brain, pancreas, placenta, heart, colon, lung, fat, kidney, spleen, bladder, and stomach. It can be observed that for pregnant and non-pregnant subjects, the main contributors associated with H3K4me3 in plasma DNA are blood-related cell types (i.e., megakaryocytes, neutrophils, and erythroblasts), with median contributions of 61.74% and 82.13%, respectively. Notably, the placental contribution of H3K4me3 was significantly higher in pregnant subjects (median: 27%; range: 0%-36.67% (P value: 0.0081, Mann-Whitney U test) compared to non-pregnant subjects who had almost no contribution. These results suggest that it is feasible to use histone modifications to infer the proportional contribution of various tissues to plasma DNA.

[0291] Fig.39 Figure showing the placental contribution versus fetal DNA fraction determined by histone modification. The placental contribution as a percentage inferred by the H3K4me3 signal is on the y-axis. The fetal DNA fraction determined by the SNP-based method is on the x-axis. The placental contribution inferred by the ChIP signal of histone modification according to an embodiment of the present disclosure is highly correlated with the fetal DNA fraction inferred by the SNP-based method (Pearson's r: 0.68; P value: 0.031). These results indicate that the use of histone modifications can determine the proportional contribution of various tissues to plasma DNA.

[0292] Fig.40 is a graph of the percentage contribution of different tissues for pregnant and non-pregnant samples based on H3K27ac histone modification of cfDNA. The x-axis shows tissue. The y-axis shows the tissue contribution inferred by H3K27ac histone modification. Fig.40It was shown that the use of another type of histone modification, such as H3K27ac, can be used to infer the proportion of DNA contribution of histone modifications from various tissues to plasma DNA. It can be observed that for both pregnant and non-pregnant subjects, the major contributors associated with H3K27ac in plasma are blood-related cell types (i.e., megakaryocytes, neutrophils, and erythroblasts), with median contributions of 89.51% and 58.67%, respectively. Notably, the placental contribution of H3K27ac in pregnant subjects was significantly higher (median: 14.45%; range: 0.42-28.19%) compared to non-pregnant subjects (median: 0%; range: 0-4.01%) (P value: <0.0001, Mann-Whitney U test).

[0293] Fig.41 Heatmaps showing tissue contributions inferred from H3K27ac ChIP signals in pregnant and non-pregnant subjects. H3K27ac contribution to the placenta in pregnant subjects is significantly higher. However, tissues that are not usually associated with pregnant subjects have higher H3K27ac ChIP signal contributions in pregnant subjects than in non-pregnant subjects. Heatmaps and cluster analysis of tissue contributions reveal tissue clusters (e.g., placenta, lung, colon, spleen, pancreas, fat, heart, kidney) that present higher contributions in pregnant subjects compared to non-pregnant subjects. Fig.41 It was shown that tissues common to pregnant and non-pregnant subjects can have different contributions from histone modifications. Other tissue clusters composed of blood group cells (e.g., erythroblasts, megakaryocytes, neutrophils) showed relatively lower contributions in pregnant subjects. 2. Simultaneous Organizational Contribution Analysis

[0294] The contribution of a specific tissue of interest can be determined based on the inferred H3K27ac histone modification signal (ChIP signal in the present disclosure) associated with the tissue-specific histone modification region. In one embodiment, the amount of histone modification can be inferred by the fragment omics properties. In one embodiment, various tissue-specific histone modification regions can be used to simultaneously analyze the contribution from multiple tissue types. As an example, we analyzed plasma DNA samples from 8 healthy subjects. For each sample, we inferred the H3K27ac ChIP signal of the region carrying the histone modification of H3K27ac specific to different tissues. The H3K27ac ChIP signal was inferred using the cumulative frequency of molecules in the length range of 230-350bp.

[0295] Fig.42A plot of inferred H3K27ac signal relative to specific tissues is shown. The inferred H3K27ac histone modification signal (also called ChIP signal) is shown on the y-axis. Tissue-specific regions are shown on the x-axis. Each point represents a plasma DNA sample.

[0296] Comparing the inferred H3K27ac ChIP signals on various tissue-specific regions, neutrophil-specific regions showed the highest median levels compared to other tissues, indicating that neutrophils are the main contributors to plasma cfDNA. The contribution of each tissue can be correlated with the ChIP signal. For example, it can be determined that monocytes and megakaryocytes may be the next major contributors. The tissues with the smallest contribution may be the placenta and colon. These observations are consistent with previous studies on healthy individuals, which have demonstrated that neutrophils are the main contributors to plasma DNA (K. Sun, et al., Proc Natl Acad Sci USA. 2015; 112; E5503–E5512). 3. Classification of pregnant subjects

[0297] The ChIP signal can be used to determine the fetal DNA fraction or to distinguish between pregnant and non-pregnant subjects.

[0298] Fig.43A and Fig.43B is a graph showing the correlation between H3K27ac ChIP signal and fetal DNA fraction determined by SNP-based methods. The x-axis shows the percentage of fetal DNA fraction determined by SNP-based methods. Fig.43A As shown, the use of the H3K27ac signal allowed a higher correlation between the placental contribution inferred by histone modifications and the fetal DNA fraction inferred by the SNP-based method (Pearson's r: 0.96; P value: <0.0001). This result emphasizes that in some embodiments, the selective use of different types of histone modifications will improve the performance of plasma DNA deconvolution of tissue DNA contributions associated with histone modifications. Fig.43B As shown, there was a weak correlation between the fetal DNA fraction and the H3K27ac signal as reads / kb (on a scale of 1 million) in placenta-specific H3K27ac regions (Pearson's r: 0.64; P value: <0.046).

[0299] Fig.44is the ROC curve for distinguishing between pregnant and non-pregnant subjects. The x-axis shows specificity. The y-axis shows sensitivity. The solid line shows the use of placental contribution inferred from H3K27ac ChIP signal. The dotted line shows the use of reads (in millions) / kb in placenta-specific H3K27ac regions. The inferred placental contribution technique has an AUC of 0.984 for distinguishing between pregnant and non-pregnant subjects. The AUC of the reads / kb technique (i.e., the measure reported in the study of Sadeh et al.) is 0.785. The results show that the inferred tissue contribution obtained using quadratic programming has better classification performance than using reads / kb. 4. Samples from cancer subjects

[0300] In one embodiment, despite the absence of colon-specific H3K4me3 regions ( Fig.37 ), but other tissue-specific regions and tissue-variable regions can still be used to infer colon contribution. We analyzed the raw sequencing data from the study of Sadeh et al. according to embodiments of the present disclosure.

[0301] Fig.45 A receiver operating characteristic (ROC) curve using the inferred colon contribution for distinguishing control subjects from colorectal cancer (CRC) subjects is shown. The ROC curve shows an area under the curve (AUC) of 0.7. The colon contribution can be used as an indicator to distinguish subjects with CRC from control subjects. In some embodiments, only tissue polymorphic regions can be used. C. Detection and monitoring of diseases

[0302] The histone modification levels measured by the embodiments in the present disclosure can be used to determine the classification of the possibility of blood disorders and the classification of cancer grades, including whether the cancer has metastasized. The histone modification levels of biological samples from objects with severe beta-thalassemia are analyzed. Severe beta-thalassemia is an example of a blood disorder. Other blood disorders are expected to have similar abnormal results at least because the blood disorder may have abnormal contributions from cells in the blood. The histone modification levels of biological samples from objects with colorectal cancer (CRC) are analyzed. CRC is an example of cancer. When cancer is localized to a tissue or when cancer metastasizes to another tissue, other cancers are expected to have similar histone modification levels. 1. Blood disorders

[0303] To demonstrate the clinical application of using histone modification-based plasma DNA tissue deconvolution, we recruited patients with hematological diseases such as, but not limited to, β-thalassemia major, iron deficiency anemia, aplastic anemia, and idiopathic thrombocytopenic purpura. We applied H3K27ac-based immunoprecipitation analysis to those plasma DNA samples followed by massively parallel sequencing.

[0304] Fig.46A is a graph comparing erythroid blast contribution inferred by H3K27ac ChIP signal between subjects with β-thalassemia major and control subjects without β-thalassemia major. The x-axis shows the subject category. The y-axis shows the erythroid blast contribution inferred by H3K27ac ChIP signal as a percentage. Fig.46A Among them, those with β-thalassemia major showed abnormal contribution from erythroid blasts (median: 34.97%; range: 6.89-68.44%) compared with healthy control subjects (median: 7.54%; range: 0-12.85%) (P value: 0.00024, Mann-Whitney U test).

[0305] Fig.46B is a ROC curve for distinguishing subjects with and without major β-thalassemia using inferred erythrocyte blast contribution. The x-axis is specificity. The y-axis is sensitivity. ROC analysis revealed that an AUC of 0.923 can be achieved by inferred erythrocyte blast contribution in erythrocyte blast-specific regions, indicating that the use of plasma DNA tissue deconvolution based on histone modifications will be able to detect and / or monitor hematological disorders (e.g., major β-thalassemia). The inferred tissue contribution is superior to the regional signal measured by reads / kb, which has an AUC of 0.892.

[0306] Fig.47 is a heat map of tissue contributions inferred using H3K27ac ChIP signals in subjects with major β-thalassemia and control subjects. Tissues were clustered and separated by tissue contribution under different pathological conditions. Erythroblasts, monocytes, brain, and others showed higher contributions in major β-thalassemia subjects when compared to control subjects. T cells, neutrophils, and megakaryocytes showed lower contributions in major β-thalassemia subjects. In addition, we observed lower erythroblast contributions (1.62%) in subjects with aplastic anemia and higher erythroblast contributions (16.07%) in subjects with iron deficiency anemia, compared to the intermediate level of erythroblast contributions (7.54%) observed in control subjects. These results are consistent with previous findings, which also observed similar trends by droplet digital PCR (ddPCR) analysis using methylation markers (Lam, et al. Clin Chem. 2017; 63: 1614-1623). These results suggest possible clinical applications of plasma DNA tissue deconvolution based on histone modifications.

[0307] In addition, we measured erythroid DNA in those plasma DNA samples using a published ddPCR assay using differentially methylated regions that are hypomethylated in erythroid blasts and hypermethylated in other cell types (Lam et al. Clin Chem. 2017; 63: 1614-1623).

[0308] Fig.48A , 48B 48C shows the correlation between the percentage of erythroid DNA determined by ddPCR assay and the erythroid blast contribution determined by H3K27ac signal. The x-axis shows the erythroid blast contribution determined by H3K27ac signal. The y-axis shows the percentage of erythroid DNA determined by ddPCR assay. Fig.48A The use of FECH (chr18:55250563-55250585) marker is shown with a Pearson's r of 0.87 and a P value < 0.0001. Fig.48B The use of the Ery 1 (chr12:48227688-48227701) marker is shown with a Pearson's r of 0.90 and a P value < 0.0001. Fig.48C The use of the Ery 2 (chr12: 48228144-48228167) marker was shown, with a Pearson's r of 0.90 and a P value < 0.0001. The data in these figures further suggest that the use of histone modifications can accurately infer the contribution of various tissues to plasma DNA. 2. Cancer with metastasis

[0309] ChIP signals of histone modifications inferred from plasma DNA without immunoprecipitation can be used to distinguish localized cancer from metastatic cancer. We analyzed a cohort of 4 localized colorectal cancer (CRC) patients, 7 CRC patients with liver metastases, and 8 healthy control samples. For each sample, we inferred H3K27ac ChIP signals in colon- and liver-specific regions. H3K27ac ChIP signals were inferred using the cumulative frequency of molecules in the length range of 230-350 bp.

[0310] Fig.49AIt is a figure comparing the plasma DNA results in the colon-specific H3K27ac region of healthy controls and objects suffering from CRC. The figure is shown in the inferred H3K27ac signal on the y-axis and the object type on the x-axis (healthy, CRC without liver metastasis, and CRC with liver metastasis). Each point represents a plasma DNA sample. The H3K27ac signal inferred in healthy subjects (median: 0.54; range: 0.27-1.08) is lower than the level of local (i.e., no liver metastasis) CRC patients (median: 0.81; range: 0.47-1.09) and CRC patients with liver metastasis (median: 1.73; range: 0.93-22.28).

[0311] Fig.49B is a graph comparing plasma DNA results in liver-specific H3K27ac regions of healthy controls and subjects with CRC. The graph shows the inferred H3K27ac signal on the y-axis and the subject type on the x-axis (healthy, CRC without liver metastasis, and CRC with liver metastasis). Each point represents a plasma DNA sample. The inferred H3K27ac levels in the liver-specific H3K27ac region are shown to be increased only in CRC patients with liver metastasis, indicating an increased liver contribution to cfDNA caused by liver metastasis. Obtaining data from colon-specific and liver-specific regions, the inferred ChIP signal can be used to distinguish between localized and metastatic cancer patients, which can provide information for clinical management. D. Example of DNA tissue imaging in urine

[0312] We have shown that by comparing the distribution of plasma DNA histone modifications to that of histone modifications derived from a variety of organs, tissues or cells, the relative tissue contributions to the plasma DNA pool can be inferred. We further demonstrate that the methods presented in this disclosure can be extended to urine samples.

[0313] Fig.50 is a plot of tissue contributions in urine and plasma samples. The x-axis shows the type of tissue. The y-axis shows the percentage contribution of the tissue. Two box plots are included for each tissue. The first box plot (grey) represents the data from the plasma samples. The second box plot (black) represents the data from the urine samples. For urine samples, the contribution is inferred by comparing the distribution of urine DNA histone modifications (e.g., H3K27ac) to the distribution of histone modifications derived from a reference organ, tissue, or cell. For plasma samples, the contribution is similarly inferred by comparing the distribution of plasma DNA histone modifications to the distribution of histone modifications derived from a reference organ, tissue, or cell.

[0314] Urine DNA samples showed significantly higher percentage contributions from kidney (median: 10.66%) and bladder (median: 4.98%) compared to their counterparts in plasma DNA samples (median kidney: 0.00%, median bladder: 0.00%), which is expected from urine samples. These results suggest that urine samples can be used to determine tissue contributions using inferred histone modification levels. E. Exemplary Methods for Determining Fractional Concentration

[0315] Fig.51 is a flow chart of an exemplary process 5100 associated with determining a concentration score for a tissue type. In some embodiments, Fig.51 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.51 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.51 One or more process blocks may be performed by one or more components of the measurement system 5900 , such as the determination 5908 , the determination device 5910 , the detector 5920 , the logic system 5930 , the local memory 5935 , the external memory 5940 , the storage device 5945 , and / or the processor 5950 .

[0316] At block 5110, N genomic regions are identified. n is an integer greater than 1. The N genomic regions may be regions known to carry tissue-specific histone modifications. The regions may be determined by criteria described herein. For example, the region may have a histone modification level of a tissue greater than a cutoff. The cutoff may be a standardized ChIP signal, based on a relative percentage difference, and / or based on a coefficient of variation for all tissue types. The region may be any region of interest described herein.

[0317] At block 5120, for each of the M tissue types, N tissue-specific histone modification levels at N genomic regions are obtained. N is greater than or equal to M. The histone modification can be H3K27ac, H3K4me3, or any histone modification described herein. The tissue histone modification levels form a matrix A of length N×M. One of the M tissue types corresponds to a first tissue type. The first tissue type can be a fetus, an erythroid blast, Fig.37 or Fig.38 Any tissue listed in, or any tissue described herein. At least one genomic region in the N genomic regions comprises a non-zero histone modification level from at least two of the M tissue types. For example, at least one histone modification level may not be exclusive to a single tissue.

[0318] In box 5130, receive input data vector b. Input data vector b can include N kinds of mixture histone modification levels in N genomic regions. N kinds of mixture histone modification levels can be measured from multiple cell-free DNA molecules in the biological sample of the object. The biological sample can be any biological sample described herein. N kinds of mixture histone modification levels can be measured by cell-free chromatin immunoprecipitation sequencing (cfChIP-seq), by determining one or more relative frequencies of a set of one or more sequence motifs in multiple cell-free DNA molecules, or by determining one or more relative frequencies of one or more length ranges in multiple cell-free DNA molecules. The relative frequencies of fragmentation features other than sequence motifs and length ranges can also be used. The mixture histone modification level can be measured by any method described herein.

[0319] At block 5140, a concentration fraction of the first tissue type is determined using the computer system and using the matrix A and the input data vector b. The contribution fraction may be determined using quadratic programming.

[0320] Process 5100 can include using the concentration score to determine a classification. For example, the first tissue type can be fetal tissue, and process 5100 can also include using the concentration score of the first tissue type to determine a classification of pregnancy in the subject. The classification of pregnancy can be whether pregnancy is present, the gestational age of the fetus (e.g., second trimester), or the grade of a pregnancy-related condition (e.g., present).

[0321] As another example, process 5100 can include using the concentration score of the first tissue type to determine the classification of the disease. For example, the disease can be major beta-thalassemia, iron deficiency anemia, aplastic anemia, or idiopathic thrombocytopenic purpura. The first tissue type can be an erythroblast, a monocyte, a brain, a T cell, a neutrophil, a megakaryocyte, or any other tissue described herein. The disease grade can be whether the disease is present or the severity of the disease. The disease can be a disease of the first tissue type (e.g., cancer).

[0322] although Fig.51 An exemplary block diagram of process 5100 is shown, but in some implementations, process 5100 may include Fig.51 The blocks shown in the process 5100 may be additional blocks, fewer blocks, different blocks, or differently arranged blocks than those shown in the process 5100. Additionally or alternatively, two or more blocks of the process 5100 may be performed in parallel. F. Exemplary Methods for Determining Classification of Pregnancy or Disease

[0323] Fig.52 is a flow chart of an exemplary process 5200 associated with determining a concentration score for a tissue type. In some embodiments, Fig.52One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.52 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.52 One or more process blocks may be performed by one or more components of the measurement system 5900 , such as the determination 5908 , the determination device 5910 , the detector 5920 , the logic system 5930 , the local memory 5935 , the external memory 5940 , the storage device 5945 , and / or the processor 5950 .

[0324] At block 5210, N genomic regions are identified. Block 5210 may be performed in the same manner as block 5110.

[0325] At block 5220 , for each of the M tissue types, N tissue-specific histone modification levels at N genomic regions are obtained. N is greater than or equal to M. Block 5220 may be performed in the same manner as block 5120 .

[0326] At block 5230 , an input data vector b is received. Block 5230 may be performed in the same manner as block 5130 .

[0327] At block 5240, a computer system, matrix A, and input data vector b may be used to determine a classification of pregnancy in the subject or a classification of disease in the subject. The classification of pregnancy or the classification of disease may be any classification described with process 5100. Process 5200 may determine the classification without determining the concentration fraction of the tissue type.

[0328] Determining the classification of pregnancy or the classification of disease may include inputting the matrix A and the input data vector b into a model (e.g., a machine learning model). The model may be trained by receiving the matrix A and a plurality of training input data vectors b obtained from a plurality of biological samples of a plurality of training subjects. Each training subject may have a known classification of the condition of the training subject. The condition may be a state of pregnancy or a known classification of a disease or any condition described herein. A plurality of training samples may be stored. Each training sample may include one of the plurality of training input data vectors b and a first label indicating a known classification of the condition. When the matrix A and the plurality of training input data vectors b are input into the model, the parameters of the model may be optimized using the plurality of training samples based on the output of the model of the corresponding labels that match or do not match the first label. The output of the model may specify the classification of the condition. The model may be used to determine the classification of the condition.

[0329] The model may include a convolutional neural network (CNN). The CNN may include a set of convolutional filters configured to filter multiple input data vectors b. The filter may be any filter described herein. The number of filters for each layer may be 10 to 20, 20 to 30, 30 to 40, 40 to 50, 50 to 60, 60 to 70, 70 to 80, 80 to 90, 90 to 100, 100 to 150, 150 to 200 or more. The kernel length of the filter may be 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 15 to 20, 20 to 30, 30 to 40 or more. The CNN may include an input layer configured to receive multiple filtered input data vectors b. The CNN may also include multiple hidden layers, which include multiple nodes. The first layer of the multiple hidden layers is coupled to the input layer. The CNN may also include an output layer coupled to a last layer in the plurality of hidden layers and configured to output an output data structure. The output data structure may include properties.

[0330] The model may include a supervised learning model. The supervised learning model may include different methods and algorithms, including analytical learning, artificial neural networks, back propagation, boosting (meta-algorithms), Bayesian statistics, case-based reasoning, decision tree learning, inductive logic programming, Gaussian process regression, genetic programming, grouping methods for data processing, kernel estimators, learning automata, learning classifier systems, minimum message length (decision trees, decision diagrams, etc.), multilinear subspace learning, naive Bayes classifiers, maximum entropy classifiers, conditional random fields, nearest neighbor algorithms, probably approximately correct learning (PAC) learning, pulsating rules, knowledge acquisition methods, symbolic machine learning algorithms, sub-symbolic machine learning algorithms, support vector machines, minimum complexity machines (MCM), random forests, ensembles of classifiers, sequential classification, data preprocessing, handling unbalanced data sets, statistical relational learning, or Proaftn (a multi-criteria classification algorithm). The model can be linear regression, logistic regression, a deep recurrent neural network (e.g., long short-term memory, LSTM), a Bayesian classifier, a hidden Markov model (HMM), linear discriminant analysis (LDA), k-means clustering, density-based clustering with noise (DBSCAN), a random forest algorithm, a support vector machine (SVM), or any model described herein.

[0331] As part of training a machine learning model, the parameters of the machine learning model (e.g., weights, thresholds, such as activation functions that can be used in a neural network, etc.) can be optimized based on the training samples (training set) to provide optimized accuracy in classifying the modification of the nucleotide at the target position. Various forms of optimization can be performed, such as back propagation, empirical risk minimization, and structural risk minimization. The accuracy of the model can be verified using a validation set (data structure and label) of the sample. Cross-validation can be performed using different parts of the training set for training and verification. The model can include multiple sub-models, thereby providing an integrated model. The sub-models can be weaker models that, once combined, provide a more accurate final model.

[0332] although Fig.52 An exemplary block diagram of process 5200 is shown, but in some implementations, process 5200 may include Fig.52 The blocks shown in the process 5200 may be additional blocks, fewer blocks, different blocks, or blocks arranged differently. Additionally or alternatively, two or more blocks of the process 5200 may be performed in parallel. VI. Disease Detection Using Sequence Motifs and Fragment Length

[0333] In embodiments, one or both of the fragment length and sequence motifs can be used to classify a pregnancy or condition. For example, the terminal motifs can be as described elsewhere in this application, including Fig.23 0 and process 2300. The length of the fragment can be used, which may not be limited to certain terminal motifs. The machine learning model can use the terminal motifs and / or the length of the fragment to classify the pregnancy or condition. A. Exemplary Results

[0334] Fig.53 The input features included in the machine learning model to distinguish hepatocellular carcinoma (HCC) and non-HCC cases are described. Arrays 5304, 5308, 5312 and 5316 each include data from tissue-specific regions. Tissue-specific regions include liver-specific regions, neutrophil-specific regions, megakaryocyte-specific regions and erythrocyte-specific regions. Each array includes fragment length and fragment end motif information. The frequency of all molecules within 230-350nt (when considering length, the molecules are not limited to any specific fragment end motif) is in each array. For example, in array 5304, the fragment aligned with a liver-specific region of length 230 has a frequency of 0.1 relative to other lengths in the liver-specific region. Other length ranges are also possible.

[0335] The array also includes the frequencies of all molecules with 9 H3K27ac-related terminal motifs (when considering terminal motifs, the molecules are not limited to any fragment length). H3K27ac-related terminal motifs include, but are not limited to, CCGG, CCGC, GCGG, TCGG, TCGC, CCGA, CCCG, GCGC and / or CCGT. H3K27ac-related terminal motifs can be defined by terminal motifs that are over-represented in regions with high H3K27ac signals compared to regions with low H3K27ac signals in sequencing results of plasma DNA samples without immunoprecipitation. For example, when comparing the results of plasma DNA samples in regions with high and low H3K27ac signals, over-representation can be a multiple change in terminal motif frequencies of 1x, 2x, 3x, 4x, 5x, 6x, 7x, 8x, 9x, 10x, 20x, 30x, 50x, etc. In some embodiments, H3K27ac-associated terminal motifs can be defined by those motifs that are over-represented in sequencing results for plasma DNA samples with immunoprecipitation compared to results for plasma DNA samples without immunoprecipitation. For example, over-representation can be a fold change in terminal motif frequency of 1x, 2x, 3x, 4x, 5x, 6x, 7x, 8x, 9x, 10x, 20x, 30x, 50x, etc. when comparing results for plasma DNA samples with and without immunoprecipitation.

[0336] Data from all arrays (i.e., a larger array or matrix) can be input into a machine learning model to distinguish non-HCC objects from HCC objects. The machine learning model can include, but is not limited to, a support vector machine, a random forest, a convolutional neural network, or any model described herein. In this example, there are a total of 130 properties for one type of tissue-specific H3K27ac-associated region. For four different tissue-specific regions, there are 520 properties.

[0337] Fig.54A and Fig.54B Shows the results from using Fig.53 Results of a machine learning model of the nature shown. Fig.54A The probability of HCC determined by the machine learning model for control subjects, subjects with chronic hepatitis B virus (HBV), and subjects with HCC is shown. The y-axis is the HCC probability. The x-axis shows the type of subject. Fig.54A It was shown that the HCC probability determined by the above machine learning model was significantly higher in HCC patients compared with patients without HCC.

[0338] Fig.54Bis the receiver operating characteristic (ROC) curve. Sensitivity is on the y-axis. Sensitivity is on the x-axis. ROC analysis showed that an area under the curve (AUC) of 0.96 for distinguishing non-HCC and HCC cases could be achieved by HCC probability.

[0339] Fig.55 is a graph showing the AUC values ​​measured using different fragmentomic properties used to distinguish between non-HCC and HCC cases. The y-axis shows the AUC value. The x-axis shows the different fragmentomic properties used in the machine learning model to distinguish between non-HCC and HCC cases. The first column shows that the AUC using the frequency of molecule lengths within 230-350bp is 0.93. The model includes 484 properties (121 different lengths and 4 tissue-specific regions). The second column shows that the AUC using the frequency of molecules with H3K27ac-related motifs is 0.95. The model uses 36 features (9 motifs and 4 tissue-specific regions). The third column uses the frequency of molecule lengths within 230-350bp and the frequency of H3K27ac-related motifs, with an AUC of 0.96. The model uses 520 properties and Fig.53 and described in 54. Fig.55 It was shown that combining length frequency and motif frequency improved the accuracy of determining HCC status. Fig.55 It was also shown that for different tissue-specific regions alone, length frequency and motif frequency can be used to distinguish HCC cases from non-HCC cases. B. Exemplary Methods

[0340] Fig.56 is a flow chart of an exemplary process 5600 for analyzing a biological sample of a subject to determine a classification of the subject's condition. The biological sample includes cell-free DNA fragments. In some embodiments, Fig.56 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.56 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.56 One or more process blocks of may be performed by one or more components of measurement system 5900, such as measurement 5908, measurement device 5910, detector 5920, logic system 5930, local memory 5935, external memory 5940, storage device 5945, and / or processor 5950. Process 5600 may include aspects described in conjunction with process 1700.

[0341] A plurality of sequence reads of the cell-free DNA fragments are received at block 5610. The plurality of sequence reads may include end sequences corresponding to ends of the cell-free DNA fragments.

[0342] At block 5620, a set of sequence reads located in one or more genomic regions is identified. Each of the one or more genomic regions may have a histone modification associated with one or more target tissue types. One or more target tissue types may include organs with cancer or fetal tissue. In some embodiments, one or more target tissue types may include liver, neutrophils, megakaryocytes, or erythroblasts. The histone modification may be H3K4me1, H3K4me2, H3K27me3, H3K27ac, H3K36me3, H3K9me2, H3K9me3, H3S10P, H3R2me, H3T2P, H3K14ac, H3K9ac, H3K79me2, H3K79me3, H4K5ac, H4K8ac, H4K12ac, H4K16ac, H4K20me, H2BK120ub, or H2AK119ub.

[0343] At block 5630, one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments are determined for each sequence read of the set of sequence reads. The set of one or more sequence motifs may include 1 to 5, 5 to 10, 11 to 15, 15 to 20, or 20 to 25 sequence motifs. The cell-free DNA fragments may consist of fragments having sequence motifs of the set of one or more sequence motifs.

[0344] At block 5640, the length of the cell-free DNA fragments is measured using the sequence reads. The cell-free DNA fragments can have a length in a predetermined length range. The predetermined length range can be any length range described herein, including 230-350 nt.

[0345] At block 5650, one or more sequence motif frequencies are determined for a set of one or more sequence motifs for each of the one or more target tissue types. The set of one or more sequence motifs appears at a higher rate in chromatin immunoprecipitation sequencing (ChIP-seq) than in sequencing without chromatin immunoprecipitation in one or more genomic regions associated with a histone modification.

[0346] At block 5660, one or more length frequencies of sequence reads of one or more length ranges are determined for each of the one or more target tissue types.

[0347] At block 5670, one or more sequence motif frequencies and one or more length frequencies for each of the one or more target tissue types are input into a machine learning model. The machine learning model may include a support vector machine, a random forest, or a convolutional neural network. The machine learning model may be any machine learning model disclosed herein, including models similar to the model described in process 5200.

[0348] The machine learning model can be trained by receiving a training data set. For each of the one or more target tissue types, the training data set can include training sequence motif frequencies of a set of one or more sequence motifs and training length frequencies of cell-free DNA fragments from a plurality of biological samples of a plurality of training subjects. Each training subject can have a known condition classification.

[0349] The machine learning model can also be trained by storing multiple training samples. For each of the one or more target tissue types, each training sample can include one or more training sequence motif frequencies of a set of one or more sequence motifs that appear in the cell-free DNA fragments in the training sample. For each of the one or more target tissue types, each training sample can include training sequence length frequencies of the cell-free DNA fragments in the training sample. Each training sample can also include a first label indicating a known classification of a condition.

[0350] When the sequence motif frequency and length frequency of each of the one or more target tissue types are input to the machine learning model, the machine learning model can be trained by optimizing the parameters of the machine learning model based on the output of the machine learning model of the corresponding label that matches or does not match the first label using a plurality of training samples. The output of the machine learning model can specify the classification of the condition.

[0351] In some embodiments, process 5600 may include, for each sequence motif in a set of one or more sequence motifs, determining a length parameter of a segment having a corresponding sequence motif. The length parameter may be a statistical value (e.g., mean, median, mode, percentile) of the segment having the corresponding sequence motif. Process 5600 may also include inputting one or more length parameters into a machine learning model. The machine learning model in these embodiments may be trained with training samples including the determined length parameters.

[0352] At block 5680, a classification of a condition of the subject is determined using a machine learning model. The condition may be pregnancy. For example, the classification of pregnancy may provide gestational age or the presence or severity of a pregnancy-related condition, including any pregnancy-related condition described herein. The condition may be a disease. The classification of a disease may be the presence or severity of a disease. The disease may be cancer, including hepatocellular carcinoma (HCC) or any cancer described herein.

[0353] In some embodiments, process 5600 can be modified to use sequence motif frequencies or length frequencies. For example, process 5600 can include using only the length frequencies of molecules within a certain length range (e.g., Fig.55In this case, blocks 5630 and 5650 are optional. Block 5670 can be modified so that one or more length frequencies are input to the machine learning model instead of one or more sequence motif frequencies. As another example, process 5600 can include using only the motif frequencies of the molecule (e.g., Fig.55 ). In this case, box 5640 and box 5660 are optional. Box 5670 can be modified so that one or more sequence motif frequencies are input into the machine learning model instead of one or more length frequencies.

[0354] Process 5600 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described elsewhere herein.

[0355] although Fig.56 An exemplary block diagram of process 5600 is shown, but in some implementations, process 5600 may include Fig.56 The blocks shown in the process 5600 may be additional blocks, fewer blocks, different blocks, or differently arranged blocks. Additionally or alternatively, two or more blocks of process 5600 may be performed in parallel. VII. Enriched Regions

[0356] The preference of DNA fragments associated with a specific epigenomic state for a specific set of terminal motifs can be used to enrich samples of DNA with that specific epigenomic state. Therefore, embodiments can enrich samples of clinically relevant DNA (including DNA from specific tissues). For example, only DNA fragments with specific terminal sequences can be sequenced, amplified and / or captured using assays. As another example, filtering of sequence reads can be performed. A. Physical enrichment

[0357] Physical enrichment can be performed in various ways, such as by targeted sequencing or PCR, which can be performed using specific primers or adapters. If a specific end motif of the end sequence is detected, adapters can be added to the ends of the fragments. Then, when sequencing is performed, only DNA fragments with adapters will be sequenced (or at least predominantly sequenced), thereby providing targeted sequencing.

[0358] As another example, primers hybridized with a specific set of terminal motifs can be used. Then, these primers can be used to order-check or amplify. Capture probes corresponding to the specific terminal motifs can also be used to capture DNA molecules with those terminal motifs for further analysis. Some embodiments can connect short oligonucleotides to the ends of plasma DNA molecules. Then, probes can be designed so that it will only identify the sequence of the oligonucleotides that are partly terminal motifs and partly connected.

[0359] Some embodiments may use CRISPR-based diagnostic techniques, such as using guide RNAs to locate sites corresponding to preferred terminal motifs of clinically relevant DNA, and then using nucleases to cleave DNA fragments, which can be performed using Cas-9 or Cas- 12. For example, adapters can be used to recognize terminal motifs, and then CRISPR / Cas9 or Cas-12 can be used to cleave terminal motif / adapter hybrids and generate universal recognizable ends for further enrichment of molecules with desired ends.

[0360] Fig.57 5700 is a flow chart of an exemplary process 5700 for enriching a biological sample for clinically relevant DNA. The biological sample may include clinically relevant DNA and other DNA that is cell-free. In some embodiments, Fig.57 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.57 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.57 One or more process blocks of may be performed by one or more components of measurement system 5900, such as measurement 5908, measurement device 5910, detector 5920, logic system 5930, local memory 5935, external memory 5940, storage device 5945, and / or processor 5950. Process 5700 may include aspects described in conjunction with process 1700.

[0361] At block 5710, a plurality of sequence reads of cell-free DNA fragments are received. The plurality of sequence reads include end sequences corresponding to the ends of the plurality of cell-free DNA fragments. One or more sequence motifs may correspond to one or more end sequences of each cell-free DNA fragment. Block 5710 may be performed in a manner similar to block 1710.

[0362] At block 5720, a set of one or more sequence motifs is identified. The set of one or more sequence motifs is present at a higher rate in chromatin immunoprecipitation sequencing (ChIP-seq) of histone modifications on clinically relevant DNA than in sequencing without chromatin immunoprecipitation. Identifying sequence motifs can be combined with process 1700 and Figure 3 , 10 , 11 and 51 are similar to the procedures described.

[0363] At block 5730, a plurality of cell-free DNA fragments may be subjected to one or more probe molecules that detect a set of one or more sequence motifs in the terminal sequences of the plurality of cell-free DNA fragments, thereby obtaining detected DNA fragments. Such use of probe molecules may result in obtaining detected DNA fragments. In one example, the one or more probe molecules may include one or more enzymes that search for a plurality of cell-free DNA fragments and append new sequences for amplifying the detected DNA fragments. In another example, the one or more probe molecules may be attached to a surface for detecting sequence motifs in the terminal sequences by hybridization.

[0364] At block 5740, the detected DNA fragments are used to enrich the biological sample for clinically relevant DNA fragments. In some embodiments, using the detected DNA fragments to enrich the biological sample may include amplifying the detected DNA fragments. In some embodiments, using the detected DNA fragments to enrich the biological sample for clinically relevant DNA fragments may include capturing the detected DNA fragments and discarding the undetected DNA fragments.

[0365] Process 5700 can also include analyzing the enriched biological sample to determine a classification of tissue of origin or disease grade. Analyzing the enriched biological sample can include sequencing DNA fragments in the enriched biological sample.

[0366] Process 5700 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described herein.

[0367] although Fig.57 An exemplary block diagram of process 5700 is shown, but in some implementations, process 5700 may include Fig.57 The boxes shown in 5700 may be more boxes, fewer boxes, different boxes, or boxes arranged differently than those shown in 5700. Additionally or alternatively, two or more boxes of process 5700 may be executed in parallel. B. Computer simulation enrichment

[0368] Computer simulation enrichment can use various criteria to select or discard certain DNA fragments. Such criteria may include terminal motifs, open chromatin regions, length, sequence variation, methylation and other epigenetic features. Epigenetic features include all modifications of the genome that do not involve changes in the DNA sequence. The standard may specify a cutoff value, such as requiring certain properties, such as a specific length range, a methylation metric above or below a specific amount, a combination of methylation states of more than one CpG site (e.g., methylation haplotypes (Guo et al, Nat Genet. 2017; 49: 635-42), etc., or having a combined probability above a threshold. This enrichment may also include weighting the DNA fragments based on this probability.

[0369] As an example, the enriched sample can be used for classification pathology (as described above), as well as for identification of tumor or fetal mutations or for tag counting for amplification / deletion detection of chromosomes or chromosome regions. For example, if a particular end motif or set of end motifs is associated with liver cancer (i.e., a higher relative frequency than non-cancer or other cancers), an embodiment for performing cancer screening can weight such DNA fragments higher than DNA fragments that do not have the preferred end motif or set of preferred end motifs.

[0370] Fig.58 58 is a flow chart of an exemplary process 5800 associated with enriching a biological sample for clinically relevant DNA. The biological sample may include clinically relevant DNA and other DNA that is cell-free. Clinically relevant DNA is DNA from the tissue of origin or DNA from diseased tissue. In some embodiments, Fig.58 One or more process blocks of can be performed by a system (e.g., measurement system 5900). In some embodiments, Fig.58 One or more process blocks of may be performed by another device or group of devices separate from or including the system. Additionally or alternatively, Fig.58 One or more process blocks of may be performed by one or more components of measurement system 5900, such as measurement 5908, measurement device 5910, detector 5920, logic system 5930, local memory 5935, external memory 5940, storage device 5945, and / or processor 5950. Process 5800 may include aspects described in conjunction with process 1700.

[0371] At block 5810, a plurality of sequence reads of cell-free DNA fragments are received. The plurality of sequence reads include end sequences corresponding to the ends of the plurality of cell-free DNA fragments. One or more sequence motifs may correspond to one or more end sequences of each cell-free DNA fragment. Block 5810 may be performed in a manner similar to block 1710.

[0372] Multiple sequence reads can be located in one or more predetermined genomic regions, wherein each of the one or more predetermined genomic regions has a histone modification associated with one or more target tissue types. The sequence reads can be aligned with a reference genome to determine their locations. Identifying sequence reads in these locations can be performed in a manner similar to block 1720.

[0373] At block 5820, one or more sequence motifs corresponding to one or more end sequences of the cell-free DNA fragments are determined for each sequence read of the set of sequence reads. Block 5820 can be performed in a manner similar to block 1730.

[0374] At block 5830, a set of one or more sequence motifs is identified. The set of one or more sequence motifs appears at a higher rate in chromatin immunoprecipitation sequencing (ChIP-seq) for histone modifications of clinically relevant DNA than in sequencing without chromatin immunoprecipitation. Identifying sequence motifs can be similar to combining process 1700 and Figure 3 , 10 , 11 and 51 described procedures.

[0375] At block 5840, a set of sequence reads having a collection of one or more sequence motifs in the end sequence is identified. This can be viewed as the first stage of filtering.

[0376] At block 5850, a likelihood that the sequence read corresponds to a clinically relevant DNA is determined for each sequence read of the set of sequence reads based on an end sequence of the sequence read that includes a sequence motif of the one or more set of sequence motifs. For example, for each sequence read of the set of sequence reads, a likelihood that the sequence read corresponds to a clinically relevant DNA can be determined based on an end sequence of the sequence read that includes a sequence motif of the one or more set of sequence motifs.

[0377] At block 5860, the likelihood is compared to a threshold for each sequence read of the group of sequence reads. As an example, the threshold can be determined empirically. For example, various thresholds of a sample can be tested, i.e., the concentration of a clinically relevant DNA of a group of sequence reads can be measured. The optimal threshold can maximize the concentration while maintaining a certain percentage of the total number of sequence reads. The threshold can be determined by one or more given percentiles (5th, 10th, 90th, or 95th) of the concentration of one or more terminal motifs present in a healthy control or a control group exposed to similar etiological risk factors but without the disease. The threshold can be a regression or probability score.

[0378] At block 5870, the sequence reads are stored when the likelihood exceeds a threshold for each sequence read of the set of sequence reads. The sequence reads may be stored in memory (e.g., in a file, table, or other data structure) to obtain stored sequence reads. Sequence reads having a probability below the threshold may be discarded or not stored in the memory location where the reads are stored, or a field of the database may include a flag indicating that the read has a lower threshold so that subsequent analysis may exclude such reads. As an example, various techniques may be used to determine likelihood, such as odds ratios, z scores, or probability distributions.

[0379] In box 5880, the sequence readings stored are analyzed to determine the properties of the clinically relevant DNA of the biological sample. For example, the property can be any property described herein, including in conjunction with other flow charts. For example, the property of the clinically relevant DNA of the biological sample can be the concentration fraction of the clinically relevant DNA. As another example, the property can be the pathological level of the object from which the biological sample is obtained, wherein the pathological level is related to the clinically relevant DNA. As another example, the property can be the gestational age of the fetus of the pregnant woman from which the biological sample is obtained.

[0380] Other criteria can be used to determine the likelihood. The lengths of multiple cell-free DNA fragments can be measured using sequence reads. The likelihood that a particular sequence read corresponds to clinically relevant DNA can be further based on the lengths of the cell-free DNA fragments corresponding to the particular sequence read.

[0381] Methylation can also be used. Thus, embodiments can measure one or more methylation states at one or more sites of cell-free DNA fragments corresponding to a particular sequence read. The likelihood that a particular sequence read corresponds to clinically relevant DNA can be further based on one or more methylation states. As another example, whether a read is within a set of identified open chromatin regions can be used as a filter.

[0382] Process 5800 may include additional implementations, such as any single implementation or any combination of implementations described herein and / or in conjunction with one or more other processes described herein.

[0383] although Fig.58 An exemplary block diagram of process 5800 is shown, but in some implementations, process 5800 may include Fig.58 The blocks shown in the process 1700 may be compared to additional blocks, fewer blocks, different blocks, or blocks in a different arrangement. Additionally or alternatively, two or more blocks of the process 1700 may be performed in parallel. VIII. Exemplary Systems

[0384] Fig.59A measurement system 5900 according to an embodiment of the present disclosure is shown. The system shown includes a sample 5905, such as a cell-free nucleic acid molecule (e.g., DNA and / or RNA) within an assay device 5910, wherein an assay 5908 may be performed on the sample 5905. For example, the sample 5905 may be contacted with a reagent for assay 5908 to provide a signal of a physical characteristic 5915 (e.g., sequence information of a cell-free nucleic acid molecule). An example of an assay device may be a flow cell or a tube through which a droplet moves including a probe and / or primer for the assay (the droplet includes the assay). The physical characteristic 5915 (e.g., fluorescence intensity, voltage, or current) from the sample is detected by a detector 5920. The detector 5920 may be measured at intervals (e.g., periodic intervals) to obtain data points constituting a data signal. In one embodiment, an analog-to-digital converter converts an analog signal from the detector into digital form at multiple times.

[0385] The assay device 5910 and the detector 5920 can form an assay system, for example, a sequencing system for sequencing according to the embodiments described herein. The data signal 5925 is sent from the detector 5920 to the logic system 5930. As an example, the data signal 5925 can be used to determine the sequence and / or position in the reference genome of a nucleic acid molecule (e.g., DNA and / or RNA). The data signal 5925 can include multiple measurements performed simultaneously, such as different electrical signals of samples 5905 of different colors of fluorescent dyes or different molecules, so the data signal 5925 can correspond to multiple signals. The data signal 5925 can be stored in a local memory 5935, an external memory 5940, or a storage device 5945. The assay system can be composed of multiple assay devices and detectors.

[0386] The logic system 5930 may be or may include a computer system, an ASIC, a microprocessor, a graphics processing unit (GPU), etc. It may also include a display (e.g., a monitor, an LED display, etc.) and a user input device (e.g., a mouse, a keyboard, a button, etc.) or be coupled to a display (e.g., a monitor, an LED display, etc.) and a user input device (e.g., a mouse, a keyboard, a button, etc.). The logic system 5930 and other components may be part of a stand-alone or network-connected computer system, or they may be directly connected to or incorporated into a device (e.g., a sequencing device) including a detector 5920 and / or an assay device 5910. The logic system 5930 may also include software executed in the processor 5950. The logic system 5930 may include a computer-readable medium storing instructions for controlling the measurement system 5900 to perform any of the methods described herein. For example, the logic system 5930 may provide commands to a system including an assay device 5910 to perform sequencing or other physical operations. Such physical operations may be performed in a specific order, for example, adding and removing reagents in a specific order. Such physical manipulations can be performed by a robotic system, for example, including a robotic arm, which can be used to obtain samples and perform assays.

[0387] The measurement system 5900 may also include a treatment device 5960 that can provide treatment to the subject. The treatment device 5960 can determine the treatment and / or be used to perform the treatment. Examples of such treatments may include surgery, radiotherapy, chemotherapy, immunotherapy, targeted therapy, hormone therapy, and stem cell transplantation. The logic system 5930 can be connected to the treatment device 5960, for example, to provide the results of the methods described herein. The treatment device can receive input from other devices, such as imaging devices and user input (e.g., to control the treatment, such as controls on a robotic system).

[0388] Any computer system described herein may utilize any suitable number of subsystems. Examples of such subsystems are Fig.60 10 is shown in FIG. 10. In some embodiments, the computer system includes a single computer device, wherein the subsystem can be a component of the computer device. In other embodiments, the computer system can include multiple computer devices, each of which is a subsystem with internal components. The computer system can include desktop and laptop computers, tablet computers, mobile phones, and other mobile devices.

[0389] Fig.60The subsystems shown are interconnected via a system bus 75. Additional subsystems, such as a printer 74, a keyboard 78, a storage device 79, a monitor 76 (e.g., a display screen, such as an LED), coupled to a display adapter 82, and other subsystems are shown. Peripheral devices and input / output (I / O) devices coupled to the I / O controller 71 can be connected to the computer system by any number of means known in the art, such as an input / output (I / O) port 77 (e.g., USB, Lightning, Thunderbolt). For example, the I / O port 77 or external interface 81 (e.g., Ethernet, Wi-Fi, etc.) can be used to connect the computer system 10 to a wide area network, such as the Internet, a mouse input device, or a scanner. The interconnection via the system bus 75 allows the central processor 73 to communicate with each subsystem and control the execution of multiple instructions from the system memory 72 or the storage device 79 (e.g., a fixed disk such as a hard drive or optical disk), as well as the exchange of information between the subsystems. The system memory 72 and / or the storage device 79 can be embodied as a computer-readable medium. Another subsystem is a data acquisition device 85, such as a camera, microphone, accelerometer, etc. Any data mentioned here can be output from one component to another and can be output to the user.

[0390] The computer system may include multiple identical components or subsystems, for example, connected together through an external interface 81, through an internal interface, or via a removable storage device that can be connected from one component to another component and removed from another component. In some embodiments, the computer system, subsystem or device can communicate through a network. In this case, one computer can be considered a client and another computer can be considered a server, where each computer can be part of the same computer system. The client and the server can each include multiple systems, subsystems or components.

[0391] Aspects of the embodiments may be implemented using hardware circuits (e.g., application specific integrated circuits or field programmable gate arrays) and / or using computer software stored in memory in a modular or integrated manner and a generally programmable processor in the form of control logic, and thus the processor may include a memory storing software instructions for configuring the hardware circuits, and an FPGA with configuration instructions or an ASIC. As used herein, the processor may include a single-core processor, a multi-core processor on the same integrated chip, or multiple processing units on a single circuit board or networked processors, as well as dedicated hardware. Based on the present disclosure and the teachings provided herein, those skilled in the art will know and understand other ways and / or methods of implementing the embodiments of the present disclosure using hardware and a combination of hardware and software.

[0392] Any of the software components or functions described in this application can be implemented as software code executed by a processor using any suitable computer language (e.g., Java, C, C++, C#, Objective-C, Swift, or a scripting language such as Perl or Python) using, for example, conventional or object-oriented techniques. The software code can be stored as a series of instructions or commands for storage and / or transmission on a computer-readable medium. Suitable non-transitory computer-readable media can include random access memory (RAM), read-only memory (ROM), magnetic media such as a hard drive or floppy disk, or optical media such as a compact disk (CD) or DVD (digital versatile disc) or Blu-ray disc, flash memory, etc. The computer-readable medium can be any combination of such devices. In addition, the order of operations can be rearranged. When the operation of a process is completed, the process can be terminated, but it can also have additional steps not included in the figure. A process can correspond to a method, a function, a program, a subroutine, a subprogram, etc. When a process corresponds to a function, its termination can correspond to the return of the function to the calling function or the main function.

[0393] The carrier signal that can be used to be suitable for transmitting via wired, optical and / or wireless networks (comprising the Internet) that meet various protocols is encoded and transmitted. Therefore, computer-readable media can be produced using data signals encoded with such programs. The computer-readable media encoded with program code can be packaged together with compatible devices or provided separately with other devices (for example, downloaded via the Internet). Any such computer-readable media can reside on or in a single computer product (for example, hard disk drive, CD or entire computer system), and can be present on or in different computer products in a system or network. The computer system can include a monitor, a printer or other suitable displays for providing any result in the result mentioned herein to the user.

[0394] Any method in the method described herein can be performed completely or partially with a computer system including one or more processors, and the computer system can be configured to perform steps. Any operation (e.g., alignment, determination, comparison, computing, calculation) performed with a processor can be performed in real time. The term "real time" can refer to a computing operation or process completed within a certain time constraint. The time constraint can be 1 minute, 1 hour, 1 day or 7 days. Therefore, the embodiment can be directed to a computer system configured to perform the steps of any method in the method described herein, and the computer system may have different components for performing corresponding steps or corresponding step groups. Although presented as numbered steps, the method steps herein can be performed simultaneously or at different times or in different orders. In addition, each part of these steps can be used together with each part of other steps from other methods. Moreover, all or part of the steps can be optional. In addition, any step in the steps of any method in the method can be performed with other devices of a module, unit, circuit or system for performing these steps.

[0395] Without departing from the spirit and scope of the embodiments of the present disclosure, the specific details of the particular embodiments may be combined in any suitable manner. However, other embodiments of the present disclosure may be directed to specific embodiments involving each individual aspect, or specific combinations of these individual aspects.

[0396] The above description of exemplary embodiments of the present disclosure has been presented for purposes of illustration and description. The above description is not intended to be exhaustive or to limit the disclosure to the precise form described, and many modifications and variations are possible in light of the above teachings.

[0397] Unless specifically indicated to the contrary, the reference to "a / an" or "the" is intended to mean "one or more". Unless specifically indicated to the contrary, the use of "or" is intended to mean "inclusive or" rather than "exclusive or". Reference to a "first" component does not necessarily require the provision of a second component. Furthermore, reference to a "first" or "second" component does not limit the referenced components to specific locations unless explicitly stated. The term "based on" is intended to mean "based, at least in part, on".

[0398] The claims may be drafted to exclude any element that may be optional. Thus, this statement is intended to serve as antecedent basis for use of exclusive terminology such as "solely," "only," and the like in connection with the recitation of claim elements, or use of a "negative" limitation.

[0399] All patents, patent applications, publications and descriptions mentioned herein are incorporated herein by reference in their entirety for all purposes. None is acknowledged to be prior art. In the event of a conflict between the present application and the references provided herein, the present application will prevail.

Claims

1. A method for analyzing a biological sample, wherein the biological sample comprises cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification associated with a target tissue type; determining a value of a fragmentomic property for each cell-free DNA fragment corresponding to each sequence read in the set of sequence reads; determining one or more relative frequencies of cell-free DNA fragments having values ​​of a fragmentomic property within a set of one or more value ranges, wherein the set of one or more value ranges occurs at different ratios in chromatin immunoprecipitation sequencing (ChIP-seq) and sequencing without chromatin immunoprecipitation for one or more genomic regions associated with a histone modification; determining an aggregate value of the one or more relative frequencies; comparing the aggregate value to one or more calibration values; and The comparison is used to determine the amount of the histone modification in the biological sample.

2. The method according to claim 1, wherein: The fragmentomic property is a sequence motif corresponding to a terminal sequence at the ends of the cell-free DNA fragments, and The one or more value ranges are one or more sequence motifs.

3. The method according to claim 1, wherein: The fragmentomic property is length, and The one or more value ranges are one or more length ranges.

4. The method according to claim 1, wherein: The fragmentomic properties are topological in nature, and The one or more value ranges are one or more topological forms.

5. The method according to claim 1, wherein: The fragmentomic property is nucleosome imprinting, and The one or more value ranges are one or more nucleosome footprints.

6. The method according to claim 1, further comprising: comparing the amount of the histone modification to one or more second calibration values, and: Using a comparison of the amount of the histone modification to the one or more second calibration values: determining the concentration fraction of the target tissue type, Determine the classification of the disease level, or A classification of the engraftment status of the target tissue type is determined.

7. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments, wherein the plurality of sequence reads include end sequences corresponding to ends of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification associated with a target tissue type; for each sequence read of the set of sequence reads, determining one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragment; Determining one or more relative frequencies of the set of one or more sequence motifs, wherein the set of one or more sequence motifs occurs at a higher rate in chromatin immunoprecipitation sequencing than in sequencing without chromatin immunoprecipitation for one or more genomic regions associated with a histone modification; determining an aggregate value of the one or more relative frequencies; comparing the aggregate value to one or more calibration values; and The fractional concentration of cell-free DNA fragments from the target tissue type is determined using the comparison, wherein the one or more calibration values ​​are determined from one or more calibration samples whose fractional concentration of cell-free DNA fragments from the target tissue type is known.

8. The method according to claim 7, wherein: the one or more relative frequencies are one or more first relative frequencies, The aggregate value is a first aggregate value, and The one or more calibration values ​​are determined by: For each calibration sample of one or more calibration samples: determining one or more second relative frequencies of the set of one or more sequence motifs in the one or more genomic regions determining a second aggregate value of the one or more second relative frequencies, Each of the one or more second aggregate values ​​is thereby associated with a known concentration fraction, wherein the one or more calibration values ​​include the one or more second aggregate values.

9. A method according to claim 7, wherein the aggregate value is a value selected from the following: (i) an entropy value; (ii) a sum of relative frequencies; (iii) a ratio of relative frequencies; and (iv) a multidimensional data point corresponding to a count vector of a set of one or more sequence motifs.

10. The method of claim 7, wherein the sequence motif of the set of one or more sequence motifs corresponds to a single nucleotide, a sequence of two nucleotides, a sequence of three nucleotides, a sequence of four nucleotides, a sequence of five nucleotides, a sequence of six nucleotides, or a sequence of seven nucleotides.

11. The method of claim 10, wherein the sequence motif comprises nucleotides at the ends of the cell-free DNA fragments.

12. The method of claim 10, wherein the sequence motif is at the 5' end.

13. The method of claim 7, wherein the target tissue type comprises placenta, liver, heart, neutrophils, monocytes, B cells, fat, or NK cells.

14. The method of claim 7, wherein the target tissue type is placenta. The method further comprises: The concentration fraction is used to determine a classification of a pregnancy-related condition or gestational age.

15. The method of claim 7, further comprising using the concentration score to determine a classification of cancer grade.

16. The method according to claim 7, wherein: The set of sequence reads is a first set of sequence reads, The one or more genomic regions are one or more first genomic regions. The histone modification is a first histone modification, the target tissue type is a first target tissue type, The set of one or more sequence motifs is a set of one or more first sequence motifs, the one or more relative frequencies are one or more first relative frequencies, The total value is a first total value, The one or more calibration samples are one or more first calibration samples, and The concentration fraction is a first concentration fraction, The method further comprises: identifying a second set of sequence reads located in one or more second genomic regions, wherein each of the one or more second genomic regions has a second histone modification associated with a second target tissue type, for each sequence read of the second set of sequence reads, determining one or more second sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments, and determining one or more second relative frequencies of the one or more second sequence motifs, wherein the set of the one or more second sequence motifs is present at a higher rate in chromatin immunoprecipitation sequencing than in sequencing without chromatin immunoprecipitation for one or more second genomic regions associated with a second histone modification, determining a second aggregate value of the one or more second relative frequencies, comparing the second aggregate value to one or more second calibration values, and A second fractional concentration of cell-free DNA fragments from the second target tissue type is determined using the comparison, wherein the one or more second calibration values ​​are determined from one or more second calibration samples whose fractional concentrations of DNA fragments from the second target tissue type are known.

17. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments, wherein the plurality of sequence reads include end sequences corresponding to ends of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification associated with a target tissue type; for each sequence read of the set of sequence reads, determining one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragment; Determining one or more relative frequencies of the set of one or more sequence motifs, wherein the set of one or more sequence motifs occurs at a higher rate in chromatin immunoprecipitation sequencing than in sequencing without chromatin immunoprecipitation for one or more genomic regions associated with a histone modification; determining an aggregate value of the one or more relative frequencies; comparing the aggregate value to one or more calibration values; and The comparison is used to estimate a first value for a characteristic of the target tissue type, wherein the one or more calibration values ​​are determined from one or more calibration samples whose values ​​of the characteristic of the target tissue type are known.

18. The method of claim 17, wherein: the one or more relative frequencies are one or more first relative frequencies, The total value is a first total value, The one or more calibration values ​​are determined by: For each calibration sample of one or more calibration samples: determining a second relative frequency of the set of one or more sequence motifs in the one or more genomic regions, and determining a second aggregate value of the one or more second relative frequencies, Each of the one or more second aggregate values ​​is thereby associated with a known value of the characteristic, wherein the one or more calibration values ​​include the one or more second aggregate values.

19. The method of claim 17, wherein the target tissue type is liver or hematopoietic cells.

20. The method of claim 17, wherein the target tissue type is an organ suffering from cancer.

21. The method of any one of claims 17 to 20, wherein the feature is cancer grade.

22. The method of claim 17, wherein the characteristic is the nutritional status of the organ.

23. The method of claim 17, wherein the target tissue type is fetal tissue.

24. The method of claim 17, wherein the biological sample is obtained from a pregnant woman, and wherein the target tissue type is placental tissue.

25. The method of claim 17, wherein the target tissue type is placental tissue, and wherein the characteristic of the placental tissue includes the gestational age of the pregnant subject.

26. The method of claim 17, wherein: The aggregate value is a first aggregate value, and The one or more calibration values ​​are one or more first calibration values, The method further comprises: measuring the length of the cell-free DNA fragments using the sequence reads, determining one or more length frequencies of sequence reads of one or more length ranges, determining a second aggregate value of the one or more length frequencies, and comparing the second aggregate value to one or more second calibration values, Wherein estimating the first value of the characteristic comprises using a comparison of the second aggregate value with the one or more second calibration values, wherein the one or more second calibration values ​​are determined from the one or more calibration samples.

27. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments, wherein the plurality of sequence reads include end sequences corresponding to ends of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification; for each sequence read of the set of sequence reads, determining one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragment; Determining one or more relative frequencies of the set of one or more sequence motifs, wherein the set of one or more sequence motifs occurs at a higher rate in chromatin immunoprecipitation sequencing than in sequencing without chromatin immunoprecipitation for one or more genomic regions associated with a histone modification; determining an aggregate value of the one or more relative frequencies; comparing the aggregate value to one or more calibration values; and The amount of the histone modification in the one or more genomic regions is determined using the comparison, wherein the one or more calibration values ​​are determined from one or more calibration samples whose amount of the histone modification is known.

28. The method of claim 27, wherein: the one or more relative frequencies are one or more first relative frequencies, The total value is a first total value, The one or more calibration values ​​are determined by: For each calibration sample of one or more calibration samples: determining one or more second relative frequencies of the set of one or more sequence motifs in the one or more genomic regions, determining a second aggregate value of the one or more second relative frequencies, Each of the one or more second aggregate values ​​is thereby associated with a known amount of a histone modification, wherein the one or more calibration values ​​include the one or more second aggregate values.

29. The method of claim 27 or 28, wherein the amount of histone modification of the one or more calibration samples is known from performing chromatin immunoprecipitation sequencing on each of the one or more calibration samples.

30. The method of claim 27, wherein: The set of sequence reads is a first set of sequence reads, The one or more genomic regions are one or more first genomic regions, The set of one or more sequence motifs is a set of one or more first sequence motifs, the one or more relative frequencies are one or more first relative frequencies, The total value is a first total value, The one or more calibration values ​​are one or more first calibration values, The amount of the histone modification is a first amount of the histone modification, and The histone modification is associated with a first tissue type and a second tissue type in the one or more first genomic regions; The method further comprises: identifying a second set of sequence reads located in one or more second genomic regions, wherein each of the one or more second genomic regions has a histone modification associated with the first tissue type and the second tissue type; for each sequence read of the second set of sequence reads, determining one or more second sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragments; determining one or more second relative frequencies of the set of the one or more second sequence motifs, wherein the set of the one or more second sequence motifs is present at a higher rate in chromatin immunoprecipitation sequencing than in sequencing without chromatin immunoprecipitation for one or more second genomic regions associated with a histone modification; determining a second aggregate value of the one or more second relative frequencies; comparing the second aggregate value to one or more second calibration values; determining a second amount of the histone modification in the one or more second genomic regions using the comparison, wherein the one or more second calibration values ​​are determined from one or more second calibration samples whose amount of the histone modification is known; and A first concentration fraction of the first tissue type and a second concentration fraction of the second tissue type are determined by solving a linear or nonlinear system of equations including a first amount of a histone modification, a second amount of a histone modification, and parameters specifying relative amounts of the corresponding histone modification for each tissue type in one or more first genomic regions and one or more second genomic regions.

31. The method of claim 30, wherein: The histone modification is associated with a third tissue type in the one or more first genomic regions and the one or more second genomic regions. The histone modification is associated with the first tissue type, the second tissue type, and the third tissue type in one or more third genomic regions, The method further comprises: determining a third amount of the histone modification in the one or more third genomic regions in the same manner as the second amount of the histone modification was determined, The third concentration fraction of the third tissue type is determined by solving a linear or nonlinear system of equations including parameters for the third amount of histone modification and the relative amount of each tissue type in the one or more third genomic regions.

32. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments, wherein the plurality of sequence reads include end sequences corresponding to ends of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification associated with one or more target tissue types; for each sequence read of the set of sequence reads, determining one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragment; Determining one or more relative frequencies of each sequence motif in the set of one or more sequence motifs, wherein the set of one or more sequence motifs occurs at a higher rate in chromatin immunoprecipitation sequencing than in sequencing without chromatin immunoprecipitation for one or more genomic regions associated with a histone modification; determining an aggregate value of the one or more relative frequencies; comparing the aggregate value to one or more calibration values; and The classification of the disorder level is determined using the comparison, wherein the one or more calibration values ​​are determined from one or more calibration samples for which the classification of the disorder level is known.

33. The method of claim 32, wherein: the one or more relative frequencies are one or more first relative frequencies, The aggregate value is a first aggregate value, and The one or more calibration values ​​are determined by: For each calibration sample of one or more calibration samples: determining a second relative frequency of the set of one or more sequence motifs in the one or more genomic regions determining a second aggregate value of the one or more second relative frequencies, Each of the one or more second aggregate values ​​is thereby associated with a known classification of the disease stage, wherein the one or more calibration values ​​include the one or more second aggregate values.

34. The method of claim 32, wherein each of the one or more genomic regions has a histone modification associated with a target tissue type.

35. The method of claim 34, wherein the disorder is in the one target tissue type.

36. The method of claim 34, wherein the condition is cancer of the one target tissue type.

37. The method of claim 32, wherein the disorder is a pregnancy-related disorder.

38. The method of claim 32, wherein: The aggregate value is a first aggregate value, and The one or more calibration values ​​are one or more first calibration values, The method further comprises: measuring the length of the cell-free DNA fragments using the sequence reads, determining one or more length frequencies of sequence reads of one or more length ranges, determining a second aggregate value of the one or more length frequencies, and comparing the second aggregate value to one or more second calibration values, Wherein determining the classification of the disease level comprises using a comparison of the second aggregate value with the one or more second calibration values, wherein the one or more second calibration values ​​are determined from the one or more calibration samples.

39. The method according to any one of claims 7 to 37, wherein the set of one or more sequence motifs comprises 1-5, 5-10, 11-15, 15-20 or 20-25 sequence motifs.

40. The method according to any one of claims 7 to 39, wherein the set of one or more sequence motifs is determined by: determining a first ratio of each of the one or more sequence motifs relative to other sequence motifs in chromatin immunoprecipitation sequencing, determining a second ratio of each of the set of one or more sequence motifs relative to other sequence motifs in sequencing without chromatin immunoprecipitation, Each of the set of one or more sequence motifs is identified as having a first ratio that is higher than the second ratio.

41. The method of any one of claims 7 to 40, wherein the histone modifications are H3K4me1, H3K4me2, H3K27me3, H3K27ac, H3K36me3, H3K9me2, H3K9me3, H3S10P, H3R2me, H3T2P, H3K14ac, H3K9ac, H3K79me2, H3K79me3, H4K5ac, H4K8ac, H4K12ac, H4K16ac, H4K20me, H2BK120ub and H2AK119ub.

42. The method of any one of claims 7 to 41, further comprising sequencing the cell-free DNA fragments in the biological sample to obtain the plurality of sequence reads.

43. The method of claim 42, wherein the volume of the biological sample is 100 μl or less.

44. The method of any one of claims 7 to 43, wherein the biological sample is plasma or serum.

45. A method for enriching a biological sample for clinically relevant DNA, the biological sample comprising cell-free said clinically relevant DNA and other DNA, the method comprising: receiving a plurality of sequence reads of cell-free DNA fragments from the biological sample, the sequence reads comprising end sequences corresponding to ends of the cell-free DNA fragments; determining, for each sequence read of the plurality of sequence reads, one or more sequence motifs corresponding to one or more end sequences of the cell-free DNA fragments; identifying a set of one or more sequence motifs that occur at a higher rate in chromatin immunoprecipitation sequencing of histone modifications of the clinically relevant DNA than in sequencing without chromatin immunoprecipitation; identifying a set of sequence reads having the set of one or more sequence motifs in the end sequence; For each sequence read of the set of sequence reads: determining a likelihood that the sequence read corresponds to the clinically relevant DNA based on an end sequence of the sequence read that includes a sequence motif from the set of one or more sequence motifs; comparing the likelihood to a threshold value; and When the likelihood exceeds the threshold, storing the sequence reading, thereby obtaining stored sequence reading; and The stored sequence reads are analyzed to determine the nature of the clinically relevant DNA of the biological sample.

46. ​​The method of claim 45, wherein the plurality of sequence reads are located in one or more predetermined genomic regions, wherein each of the one or more predetermined genomic regions has a histone modification associated with a target tissue type.

47. The method of claim 46, wherein the clinically relevant DNA is DNA from the target tissue type or DNA from diseased tissue.

48. A method according to claim 45, wherein the plurality of sequence reads are located in one or more predetermined genomic regions, wherein each of the one or more predetermined genomic regions has a histone modification associated with any target tissue type of a plurality of target tissue types.

49. The method according to any one of claims 45 to 48, further comprising: The lengths of the cell-free DNA fragments are measured using the sequence reads, and wherein determining the likelihood that a particular sequence read corresponds to the clinically relevant DNA is further based on the lengths of the cell-free DNA fragments corresponding to the particular sequence read.

50. The method according to any one of claims 45 to 49, further comprising: One or more methylation states are measured at one or more sites of the cell-free DNA fragments corresponding to specific sequence reads, wherein determining the likelihood that the specific sequence read corresponds to the clinically relevant DNA is further based on the one or more methylation states.

51. A method for enriching a biological sample for clinically relevant DNA, the biological sample comprising cell-free clinically relevant DNA and other DNA, the method comprising: receiving a plurality of cell-free DNA fragments from the biological sample, wherein: The ends of the plurality of cell-free DNA fragments have terminal sequences, and the one or more sequence motifs corresponding to one or more end sequences of each of the plurality of cell-free DNA fragments; identifying a set of the one or more sequence motifs that occur at a higher rate in chromatin immunoprecipitation sequencing of histone modifications of the clinically relevant DNA than in sequencing without chromatin immunoprecipitation; subjecting the plurality of cell-free DNA fragments to one or more probe molecules that detect the set of one or more sequence motifs in the terminal sequences of the plurality of cell-free DNA fragments, thereby obtaining detected DNA fragments; and The detected DNA fragments are used to enrich clinically relevant DNA fragments of the biological sample.

52. The method of claim 51, further comprising analyzing the enriched biological sample to determine a classification of tissue of origin or disease stage.

53. The method of claim 51, wherein using the detected DNA fragments to enrich the clinically relevant DNA fragments of the biological sample comprises: The detected DNA fragment is amplified.

54. The method of claim 51, wherein the one or more probe molecules include one or more enzymes that search for the plurality of cell-free DNA fragments and append new sequences for amplifying the detected DNA fragments.

55. The method of claim 51, wherein using the detected DNA fragments to enrich the clinically relevant DNA fragments of the biological sample comprises: capturing the detected DNA fragments; and Undetected DNA fragments were discarded.

56. The method of claim 55, wherein one or more probe molecules are attached to a surface and the sequence motif in the terminal sequence is detected by hybridization.

57. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: identifying N genomic regions, where N is an integer greater than 1; For each of the M tissue types: Obtaining N tissue histone modification levels in the N genomic regions, N is greater than or equal to M, wherein: The tissue histone modification levels form a matrix A of N×M dimensions, One of the M tissue types corresponds to a first tissue type, and At least one genomic region of the N genomic regions comprises non-zero histone modification levels from at least two of the M tissue types; receiving an input data vector b comprising N mixture histone modification levels at the N genomic regions, the N mixture histone modification levels being measured from a plurality of cell-free DNA molecules in a biological sample of a subject; and The concentration fraction of the first tissue type is determined using a computer system using the matrix A and the input data vector b.

58. A method according to claim 57, wherein the histone modification levels of the N mixtures are measured by chromatin immunoprecipitation sequencing, by determining one or more relative frequencies of a set of one or more sequence motifs in the plurality of cell-free DNA molecules, or by determining one or more relative frequencies of one or more length ranges in the plurality of cell-free DNA molecules.

59. The method of claim 57, wherein the histone modification is H3K27ac or H3K4me3.

60. The method of claim 57, wherein the first tissue type is fetal or erythroblast tissue.

61. The method of claim 57, wherein the first tissue type is fetal tissue, The method further comprises: The fractional concentration of the first tissue type is used to determine a classification of pregnancy in the subject.

62. The method of claim 57 further comprising using the concentration fraction of the first tissue type to determine a disease classification.

63. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: identifying N genomic regions, where N is an integer greater than 1; For each of the M tissue types: Obtaining N tissue-specific histone modification levels in the N genomic regions, N being greater than or equal to M, wherein: The tissue-specific histone modification levels form a matrix A of N×M dimensions, one of the M tissue types corresponds to a first tissue type, and At least one genomic region of the N genomic regions comprises non-zero histone modification levels from at least two of the M tissue types; receiving an input data vector b comprising N mixture histone modification levels at the N genomic regions, the N mixture histone modification levels being measured from a plurality of cell-free DNA molecules in a biological sample of a subject; and Use a computer system to use matrix A and input data vector b to determine: the classification of pregnancy in the subject, or Classification of the disease in the subject.

64. The method of claim 63, wherein determining the classification of the pregnancy or the classification of the disease comprises: The matrix A and the input data vector b are input into a machine learning model, which is trained by: Store multiple training samples, each training sample includes: one of a plurality of training input data vectors b, wherein the plurality of training input data vectors b are obtained from a plurality of biological samples of a plurality of training subjects, and a first label indicating a known classification of the condition of the training object, and When the matrix A and a plurality of training input data vectors b are input to the machine learning model, the parameters of the machine learning model are optimized based on the output of the machine learning model of the corresponding labels that match or do not match the first label using the plurality of training samples, wherein the output of the machine learning model specifies the classification of the condition, and The machine learning model is used to determine a classification of pregnancy or a classification of disease.

65. A method for analyzing a biological sample, the biological sample comprising cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification associated with one or more target tissue types; measuring the length of each cell-free DNA fragment corresponding to each sequence in the set of sequence reads; determining one or more relative frequencies of cell-free DNA fragments of lengths in a set of one or more length ranges, wherein the set of one or more length ranges occurs at different ratios in chromatin immunoprecipitation sequencing and sequencing without chromatin immunoprecipitation for one or more genomic regions associated with a histone modification; determining an aggregate value of the one or more relative frequencies; comparing the aggregate value to one or more calibration values; and The comparison is used to determine the amount of the histone modification in the biological sample.

66. The method according to claim 65, further comprising: comparing the amount of the histone modification to one or more second calibration values, and Comparison of the amount of the histone modification to the one or more second calibration values ​​is used to determine the fractional concentration of the target tissue type.

67. The method according to claim 65, further comprising: comparing the amount of the histone modification to one or more second calibration values, and A classification of the disorder level is determined using the one or more second calibration values.

68. The method according to claim 65, further comprising: comparing the amount of the histone modification to one or more second calibration values, and A classification of the engraftment status of the target tissue type is determined using the one or more second calibration values.

69. A method of analyzing a biological sample of a subject, the biological sample comprising cell-free DNA fragments, the method comprising: receiving a plurality of sequence reads of the cell-free DNA fragments, wherein the plurality of sequence reads include end sequences corresponding to ends of the cell-free DNA fragments; identifying a set of sequence reads located in one or more genomic regions, wherein each of the one or more genomic regions has a histone modification associated with one or more target tissue types; for each sequence read of the set of sequence reads, determining one or more sequence motifs corresponding to one or more end sequences of the corresponding cell-free DNA fragment; measuring the length of the cell-free DNA fragments using the sequence reads; For each of the one or more target tissue types: determining one or more sequence motif frequencies of the set of one or more sequence motifs, wherein the set of one or more sequence motifs occurs at a higher rate in chromatin immunoprecipitation sequencing of histone modifications associated with the one or more genomic regions than in sequencing without chromatin immunoprecipitation; determining one or more length frequencies of sequence reads of one or more length ranges; for each of the one or more target tissue types, inputting the one or more sequence motif frequencies and the one or more length frequencies into a machine learning model, and A classification of the condition of the object is determined using the machine learning model.

70. The method of claim 69, wherein the condition is pregnancy.

71. The method of claim 69, wherein the condition is a disease.

72. The method of claim 69, wherein the condition is cancer.

73. The method of claim 69, wherein the machine learning model is trained by: Store multiple training samples, each training sample includes: For each of the one or more target tissue types: one or more training sequence motif frequencies in the set of the one or more sequence motifs that occur in the cell-free DNA fragments of the training sample, and The training length frequencies of cell-free DNA fragments in the training samples, and a first label indicating a known classification of the condition, and When the sequence motif frequency and the length frequency are input into the machine learning model, the parameters of the machine learning model are optimized based on the output of the machine learning model of the corresponding label that matches or does not match the first label using the multiple training samples, wherein the output of the machine learning model specifies the classification of the condition.

74. The method of claim 69, wherein the cell-free DNA fragments have a length within a predetermined length range.

75. The method of claim 74, wherein the predetermined length range is 230 to 350 nt.

76. The method of claim 69, wherein the cell-free DNA fragments consist of fragments having sequence motifs of the set of one or more sequence motifs.

77. The method according to claim 69, further comprising: For each sequence motif in the set of one or more sequence motifs, determining a length parameter of a segment having the corresponding sequence motif, and The one or more length parameters are input into the machine learning model.

78. The method of claim 69, wherein the one or more target tissue types include an organ having cancer or fetal tissue.

79. The method of claim 69, wherein the one or more target tissue types include liver, neutrophils, megakaryocytes, or erythroblasts.

80. The method of claim 69, wherein the set of one or more sequence motifs comprises 1-5, 5-10, 11-15, 15-20 or 20-25 sequence motifs.

81. The method of claim 69, wherein the histone modification is H3K4me1, H3K4me2, H3K27me3, H3K27ac, H3K36me3, H3K9me2, H3K9me3, H3S10P, H3R2me, H3T2P, H3K14ac, H3K9ac, H3K79me2, H3K79me3, H4K5ac, H4K8ac, H4K12ac, H4K16ac, H4K20me, H2BK120ub or H2AK119ub.

82. A method according to claim 69, wherein the machine learning model includes linear regression, logistic regression, deep recurrent neural network, Bayesian classifier, convolutional neural network (CNN), hidden Markov model (HMM), linear discriminant analysis (LDA), k-means clustering, density-based clustering algorithm with noise (DBSCAN), random forest algorithm or support vector machine (SVM).

83. A computer product comprising a computer readable medium storing a plurality of instructions for controlling a computer system to perform the method of any one of the above claims.

84. Systems, including: The computer product of claim 83; and One or more processors for executing instructions stored on a computer-readable medium.

85. A system comprising means for performing the method of any of the above claims.

86. A system comprising one or more processors configured to perform the method of any of the above claims.

87. A system comprising modules for respectively performing the steps of the method described in any one of the above claims.

Images