Techniques for detecting minimal residual disease

Abstract

The present disclosure describes techniques for determining an indication of minimal residual disease (MRD) in a subject. An indication of the MRD may be determined based on sequencing data from a biological sample of the subject. Partial implementations of these techniques are performed by using the same set of sequencing data to determine an indication of sequencing error and MRD from the same biological sample.

Description

[0001] Related patent applications

[0002] This patent application claims the benefit of UK (GB) patent application No. 2208273.9, filed on June 6, 2022, entitled “Techniques For Detecting Minimum Residual Disease” and assigned by Mewburn Ellis Agents as case number 008275836. The entire contents of the foregoing patent application are incorporated herein by reference, including all text, tables and drawings. Background Art

[0003] A central challenge in treating cancer is the early detection of minimal residual disease after cancer treatment. Minimal residual disease (MRD) can serve as an indicator of cancer recurrence, which often occurs before standard surveillance imaging. One strategy to identify MRD is to monitor circulating tumor DNA (ctDNA) from patient biospecimens, which can be shed by cancer. Summary of the invention

[0004] Some embodiments of the present disclosure provide a method for determining whether sequencing data of a biological sample of a subject provides an indication that the subject suffers from minimal residual disease. The method includes using at least one computer hardware processor to perform the following operations: (A) obtaining the sequencing data, the sequencing data being previously generated by sequencing the biological sample of the subject, the sequencing data including sequence reads covering the locations of mutations to be monitored; (B) using at least a first subset of the sequence reads to determine a first value indicating the expected number of mutations present in the sequencing data due to sequencing errors, the determination comprising: using the first subset of the sequence reads to determine the corresponding multiple trinucleotide fields (trinucleotide (c) determining a first value indicating the expected number of mutations present in the sequencing data using at least a second subset of the sequence reads; and (d) determining whether the sequencing data provides the indication that the subject suffers from minimal residual disease using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the position of the mutation to be monitored.

[0005] In some embodiments, the sequence reads cover at least 10 positions of mutations to be monitored. In some embodiments, the sequence reads cover 10-200 positions of mutations to be monitored. In some embodiments, the sequence reads cover 50-200 positions of mutations to be monitored.

[0006] In some embodiments, the method further comprises: obtaining the sequencing data by sequencing the biological sample. In some embodiments, the method further comprises: obtaining the biological sample from a body fluid of the subject. In some embodiments, the biological sample comprises circulating tumor DNA (ctDNA). In some embodiments, each of the sequence reads covers at least one position of the position of the mutation to be monitored. In some embodiments, the sequence reads are obtained using whole exome sequencing. In some embodiments, the sequence reads are obtained using a targeted gene sequencing kit. In some embodiments, the targeted gene sequencing kit targets a sequence covering the position of the mutation to be monitored. In some embodiments, primers are used to amplify the sequence covering the position of the mutation to be monitored. In some embodiments, the sequence targeted by the targeted gene sequencing kit is determined using sequence data from a primary tumor of the subject.

[0007] In some embodiments, the first subset of the sequence reads and the second subset of the sequence reads are the same. In some embodiments, (B) is performed using at least the first subset of the sequence reads and one or more sequence reads in the sequencing data that do not cover the position of the mutation to be monitored. In some embodiments, performing (B) further includes: using at least the first subset of the sequence reads to generate a consensus sequence read, wherein each of the consensus sequence reads is generated by a sequence read associated with a corresponding common unique molecular identifier (UMI) in at least the first subset of the sequence reads, wherein determining the multiple TNC error rates of the corresponding multiple trinucleotide field (TNC) error types is performed using the generated consensus sequence reads.

[0008] In some implementations, each of the consensus sequence reads is generated from at least a threshold number of sequence reads associated with a corresponding common UMI. In some implementations, the threshold number of sequence reads is between 2 and 20.

[0009] In some embodiments, the method further comprises: selecting a subset of the consensus sequence reads, wherein determining the plurality of TNC error rates for the corresponding plurality of trinucleotide field (TNC) error types is performed using only the selected subset of the consensus sequence reads. In some embodiments, the consensus sequence reads comprise a positive strand consensus sequence read and a negative strand consensus sequence read, and wherein selecting the subset is performed using a criterion for a measure of similarity between corresponding positive strand consensus sequence reads and negative strand consensus sequence reads. In some embodiments, the consensus sequence reads comprise a positive strand consensus sequence read and a negative strand consensus sequence read, and the subset of the consensus reads is selected using one or more criteria applicable to the positive strand consensus sequence reads and the negative strand consensus sequence reads.

[0010] In some embodiments, the method further comprises: determining the measure of similarity between corresponding positive strand consensus sequence reads and negative strand consensus sequence reads. In some embodiments, the consensus sequence reads comprise positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to the relative number of positive strand consensus sequence reads and corresponding negative strand consensus sequence reads.

[0011] In some embodiments, using the consensus sequence reads to determine the plurality of TNC error rates for the corresponding plurality of trinucleotide field (TNC) error types comprises: using a background region of the consensus sequence reads to determine the plurality of TNC error rates, wherein the position of the mutation to be monitored comprises a first position, wherein the consensus sequence reads comprise a first consensus sequence read covering the first position, and the background region comprises a first background region of the first consensus sequence reads, wherein the first background region comprises nucleotides in the first consensus sequence reads that are at least a first threshold distance from the first position. In some embodiments, the background region does not include the position of the mutation to be monitored.

[0012] In some embodiments, for the first position among the positions of the mutation to be monitored, the consensus sequence reads include a first group of positive chain consensus sequence reads associated with a positive chain primer binding sequence and a second group of negative chain consensus sequence reads associated with a negative chain primer binding sequence, the positive chain primer binding sequence starting at the 3' end of each positive chain primer in the first group, and the negative chain primer binding sequence starting at the 3' end of the negative chain primer in the second group, wherein using the consensus sequence reads to determine the multiple TNC error rates of the corresponding multiple trinucleotide field (TNC) error types includes: determining the multiple TNC error rates using: nucleotides within a second threshold distance range of the positive chain primer binding sequence in any sequence read in the first group of positive chain consensus sequence reads, and nucleotides within a third threshold distance range of the negative chain primer binding sequence in any sequence read in the second group of negative chain consensus sequence reads. In some embodiments, determining the plurality of trinucleotide field (TNC) error rates using the consensus sequence reads comprises determining a frequency of occurrence of each of the TNC error types in the consensus sequence reads.

[0013] In some embodiments, using the consensus sequence reads to determine the plurality of TNC error rates for the corresponding plurality of trinucleotide field (TNC) error types comprises: determining the plurality of TNC error rates based on a background region of the consensus sequence reads, wherein the consensus sequence reads comprise a first consensus sequence read, and the background region comprises a first background region of the first consensus sequence reads, wherein the TNC error rate is determined based on how frequently each of the TNC error types occurs in the first background region of the first consensus sequence. In some embodiments, the TNC error type corresponds to a mutation in any position of a given TNC.

[0014] In some embodiments, each of the TNC error types corresponds to a specific mutation of an intervening nucleotide in a given TNC.

[0015] In some embodiments, the method further comprises: after determining the plurality of trinucleotide field (TNC) error rates and before grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups, determining a confidence interval for the TNC error rates; and selecting the at least some of the plurality of TNC error rates for grouping based on the confidence interval for the TNC error rates. In some embodiments, grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups comprises clustering the plurality of TNC error rates. In some embodiments, grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups comprises grouping using a partition around medoids (PAM) clustering method. In some embodiments, grouping at least some of the plurality of TNC error rates comprises grouping into four TNC error rate groups.

[0016] In some embodiments, determining the first value indicative of the expected number of mutations present in the sequencing data is performed using at least some of the TNC group error rates and the number of times at least some of the positions of the mutations to be monitored are covered by sequence reads in the first subset of the sequence reads. In some embodiments, determining the first value indicative of the expected number of mutations present in the sequencing data comprises determining the first value as a weighted linear combination of the TNC error group rates, wherein each specific one of the TNC error group rates is weighted by the number of times the position to be monitored is covered by sequence reads in the first subset of the sequence reads corresponding to a TNC error type belonging to the specific TNC error group.

[0017] In some embodiments, performing (C) further includes: generating second consensus sequence reads using at least the second subset of the sequence reads, wherein each of the second consensus sequence reads is generated from sequence reads in at least the second subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI), wherein determining the second value indicating the actual number of mutations present at the position of the mutation to be monitored is performed using the second consensus sequence reads.

[0018] In some embodiments, (D) is performed using a statistical hypothesis test with a null hypothesis by comparing the second value to a distribution associated with the null hypothesis, wherein the distribution has one or more parameters that depend on the first value. In some embodiments, the distribution is a Poisson distribution with a mean (λ), and the mean is set to the first value. In some embodiments, using the statistical hypothesis test includes determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the null hypothesis. In some embodiments, (D) is performed using a one-sided Poisson hypothesis test. In some embodiments, using the one-sided Poisson hypothesis test includes setting the mean (λ) of the Poisson distribution to the first value, and determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the Poisson distribution.

[0019] In some embodiments, the measure of likelihood is used to determine whether the sequencing data provides the indication that the subject has minimal residual disease. In some embodiments, where the second value indicates that the null hypothesis can be rejected, the subject may have minimal residual disease. In some embodiments, (D) further comprises: providing the indication that the subject has minimal residual disease.

[0020] In some embodiments, the method further includes using the at least one computer hardware processor to perform the following operations: obtaining one or more additional sequencing data previously generated by sequencing one or more additional biological samples of the subject, each of the one or more additional sequencing data including additional sequence reads covering the location of the mutation to be monitored, and for each of the additional sequence reads in the one or more additional sequencing data: using at least a first subset of the additional sequence reads to determine an additional first value indicating an expected number of mutations due to sequencing errors present in the corresponding sequencing data, the determination comprising: determining additional multiple trinucleotide field (TNC) error rates for corresponding multiple TNC error types; grouping at least some of the additional multiple TNC error rates into additional multiple TNC error rate groups; using additional TNC error rates among at least some of the additional multiple TNC error rates to determine additional TNC group error rates of the additional multiple TNC error rate groups; and using the additional TNC group error rates to determine the additional first value indicating the expected number of mutations present in the corresponding sequencing data; using at least a second subset of the additional sequence reads to determine an additional second value indicating the actual number of mutations present at the position; and using the additional first value indicating the expected number of mutations due to sequencing errors present in the corresponding sequencing data and the additional second value indicating the actual number of mutations present in the corresponding sequencing data at the position of the mutation to be monitored to determine whether the corresponding sequencing data provides the indication that the subject suffers from minimal residual disease.

[0021] The various aspects described above may be used interchangeably or in addition to various aspects of any system, method, and / or process described herein. Further, the system may be configured to operate according to a method having one or more of the aforementioned aspects. This system may include at least one computer hardware processor and at least one non-transitory computer-readable medium storing processor-executable instructions, which, when executed by the at least one computer hardware processor, causes the at least one computer hardware processor to perform this method. Further, the non-transitory computer-readable medium may include processor-executable instructions, which, when executed by at least one computer hardware processor of a data processing system, causes the at least one computer hardware processor to perform a method having one or more of the aforementioned aspects. Therefore, the foregoing is a non-limiting inventive summary of the present invention as defined by the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1is a diagram depicting an illustrative technique 100 for determining an indication of MRD in a subject using sequencing data from a biological sample of the subject and statistical analysis, according to some embodiments of the technology described herein.

[0023] Figure 2A is a flow chart of an illustrative process 200 for determining an indication of MRD in a subject using sequencing data from a biological sample according to some embodiments of the technology described herein.

[0024] Figure 2B is a flow chart of an illustrative process 210 for determining an expected number of mutations due to sequencing errors according to some embodiments of the technology described herein.

[0025] Figure 3 It is shown that according to some embodiments of the technology described herein, multiple TNC error rates are determined using background regions of consensus sequence reads.

[0026] Figure 4 It is shown that according to some implementations of the techniques described herein, TNC error types are selected based on their respective error rates.

[0027] Figure 5 Determining TNC group error rates for a corresponding plurality of TNC error rate groups is shown according to some implementations of the techniques described herein.

[0028] Figure 6

[00136] According to some embodiments of the technology described herein, aspects of determining the likelihood of the presence of MRD by statistical testing using a first value and a second value are shown.

[0029] Figure 7 Various aspects of determining an indication of MRD according to some embodiments of the technology described herein are shown.

[0030] Figure 8 Shown are box plots demonstrating the median error rate (%, y-axis) for each of 96 trinucleotide field (x-axis) conversion events from patient primary non-small cell lung cancer (NSCLC) tumor sequencing data.

[0031] Fig. 9 Shown are box plots demonstrating the median error rate (%, y-axis) for each of 96 trinucleotide field (x-axis) transversion events from patient primary non-small cell lung cancer (NSCLC) tumor sequencing data.

[0032] Fig. 10A Shown is the false positive detection rate of MRD in patients without clinical cancer recurrence when MRD analysis is performed according to some embodiments of the technology described herein (top graph).

[0033] Fig. 10B Shown is the false positive detection rate of MRD in patients with second primary tumors that would not be expected to be detectable by MRD when MRD analysis is performed according to some embodiments of the technology described herein (lower graph).

[0034] Fig. 10C It is shown that when MRD analysis is performed according to some embodiments of the technology described herein, the median time to detect MRD before clinical cancer recurrence is 151 days, and 37 of the 41 medical records were found to have MRD through MRD testing before clinical cancer recurrence. Fig. 10C It was also shown that when MRD analysis was performed according to some embodiments of the technology described herein, in "non-shedding" NSLC, the MRD detection rate was poor, with a median lead time to detection of 22 days.

[0035] Fig. 10D Overlap between indication of MRD based on ctDNA sequencing and determination of MRD using standard surveillance imaging is shown.

[0036] Fig.11A It was demonstrated that when MRD analysis was performed using some embodiments of the technology described herein, MRD was generally undetectable in non-relapsed patients.

[0037] Fig. 11B It was shown that when MRD analysis was performed using some embodiments of the technology described herein, MRD was almost always detected in patients with postoperative relapse.

[0038] Fig. 11C It was demonstrated that when MRD analysis was performed using some embodiments of the techniques described herein, MRD was almost always detected in patients prior to surgery.

[0039] Fig.11D : Simulated MRD-negative samples showed that when MRD analysis was performed using some embodiments of the technology described in this article, 121 of the 3157 simulated mimic sets were ctDNA positive at a threshold of P value <0.1 (in-silico specificity of 96.2%), and 22 of the 3157 simulated mimic sets were ctDNA positive at a P value <0.01 (in-silico specificity of 99.3%).

[0040] Fig.11E Shown is the relationship between the sensitivity (p-value) of the MRD analysis compared to the amount of ctDNA in a control sample spiked with background fragmented DNA when performing the MRD analysis using some embodiments of the technology described herein.

[0041] Fig.12An illustrative embodiment of a computer system is depicted that can be used in conjunction with some implementations of the techniques described herein. DETAILED DESCRIPTION

[0042] Early detection of cancer relapse (relapse / recurrence) is an important aspect of effective cancer treatment. A strategy for detecting cancer relapse is to look for minimal residual disease (MRD) in biological samples collected from subjects after cancer therapy. MRD can be determined by sequencing biological samples from subjects to identify circulating tumor DNA (ctDNA), which is an indicator of cancer relapse. In addition to cancer-related mutations, ctDNA largely contains the same sequence as the wild-type DNA of the subject.

[0043] Determining the presence of cancer-associated mutations in cfDNA or ctDNA, an indication of MRD, is challenging for many reasons, including sequencing errors in sequencing experiments. Sequencing errors can be introduced at multiple points in the process of obtaining the sample, preparing the sample, sequencing, and performing post-sequencing analysis on the data. For example, sequencing errors can lead to false positive identifications of MRD due to false positives for cancer-associated mutations. Therefore, methods are needed to correct for false positives of cancer-associated mutations when identifying MRD.

[0044] Conventional methods correct sequencing errors by sequencing negative control samples from healthy subjects (see, for example, Abbosh, Christopher et al., Nature 545.7655 (2017): 446-451). These methods assume that the sequencing errors associated with healthy subjects are roughly the same as the sequencing errors associated with cancer subjects, so when the MRD signal in the cancer subject significantly exceeds the MRD signal in the healthy subject, a positive MRD determination is made. However, the assumption that the sequencing errors between healthy subjects and cancer subjects are similar is not accurate in all cases. The sequencing errors between sequencing data collected from different biological samples may depend on the collection of biological samples, the preparation of biological samples for sequencing, the sequencing instrument (including the type of instrument and the maintenance of the instrument), and the analysis of sequencing data. Therefore, conventional techniques that determine how many sequencing errors exist in the sequencing data of a sample based on the errors existing in the sequencing data of other samples may result in inaccurate estimates of the errors. For example, a recent experiment compared genome sequencing data from the same cell line sequenced under different circumstances and found that the probability of false negatives (absence of mutations when expected) was 42-51%, and the probability of false positives (presence of mutations when expected) was 5-8% (see Kim, Young-Ho et al., PloS one 14.9 (2019): e0222535). In contrast to conventional methods, the inventors have developed a technique for determining the amount of sequencing errors present in a sample using sequencing data from the same sample (e.g., sequencing errors and indications of MRD can be determined using the same sequencing data from the same biological sample).

[0045] This technology involves determining the sequencing error rate (e.g., a value representing the ratio of incorrect nucleotides identified at a certain position) by monitoring the error rate in a nucleotide or a group of nucleotides (i.e., a nucleotide field (NC)); an incorrect nucleotide can be identified at a certain position due to an event occurring in sample collection, preparation, sequencing, post-sequence analysis, or any other situation in which the sample or data is manipulated). A nucleotide field (NC) refers to a series of continuous nucleic acids having a specific base in a nucleic acid sequence or sequence read. In some embodiments, the error rate in a single nucleotide (single nucleotide field) is monitored. In some embodiments, the error rate in a group of two nucleotides (dinucleotide fields), three nucleotides (trinucleotide fields), four nucleotides (tetranucleotide fields), five nucleotides (pentanucleotide fields), six nucleotides (hexanucleotide fields) or more nucleotides is monitored. In some embodiments, the error rate in a group of trinucleotide fields is monitored as described herein. Furthermore, the estimated sequencing error rate can be compared with the actual number of mutations observed in the position of the mutation to be monitored to determine an indication of MRD. In some embodiments, this technique involves estimating sequencing errors based on sequencing results at positions that do not target cancer-associated mutations to be monitored (the set of such sequence read positions may be referred to herein as "background regions").

[0046] Thus, some embodiments provide computer-implemented methods for determining whether sequencing data of a biological sample (eg, plasma) of a subject (eg, a human subject) provides an indication that the subject has minimal residual disease. In some embodiments, the method includes: (A) obtaining the sequencing data, which was previously generated by sequencing a biological sample of a subject, and the sequencing data includes sequence reads covering the positions of mutations to be monitored (for example, the positions can be determined by sequencing results of the subject's primary tumor to identify positions that provide information for subsequent monitoring of MRD); (B) using at least a first subset of the sequence reads (for example, a first subset that can be selected based on data quality) to determine a first value indicating the expected number of mutations due to sequencing errors present in the sequencing data, the determination including: using the first subset of the sequence reads to determine multiple NC error rates for corresponding multiple nucleotide field (NC) error types (for example, different types of mutations that can be observed in a nucleotide field), the NC error rates selected from single nucleotide fields, dinucleotide fields, trinucleotide fields, tetranucleotide fields, pentanucleotide fields and hexanucleotide fields (for example, the rate of nucleotide repeat mutations due to sequencing errors); grouping at least some of the multiple NC error rates into multiple NC error rate groups (for example, using PAM clustering, k-nearest neighbor clustering or hierarchical aggregation Clustering method; grouping can reduce statistical noise by increasing the number of sequence reads used to determine the NC group error rate, which is in turn used to determine the first value); using the NC error rates of at least some of the multiple NC error rates to determine the NC group error rate of the multiple NC error rate groups (for example, the population-weighted average of the NC error rate can be the NC group error rate); and using the NC group error rate to determine the first value indicating the expected number of mutations present in the sequencing data; (C) using at least a second subset of the sequence reads to determine a second value indicating the actual number of mutations present at the location of the mutation to be monitored (for example, the number of mutations present at the location of the mutation to be monitored); and (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored (for example, using a one-sided Poisson test, in which case the first value is λ) to determine whether the sequencing data provides the indication that the subject has minimal residual disease (for example, the probability that the subject has MRD).

[0047] Although the embodiments, examples, claims and drawings of this document often refer to trinucleotide fields (TNCs), it should be understood that the methods described in the embodiments, examples, claims and drawings of this document can be performed using any suitable nucleotide field (e.g., single nucleotide field (SNC), dinucleotide field (DNC), trinucleotide field (TNC), tetranucleotide field (4NC), pentanucleotide field (5NC), hexanucleotide field (6NC), etc.).

[0048] Thus, some embodiments provide computer-implemented methods for determining whether sequencing data of a biological sample (eg, plasma) of a subject (eg, a human subject) provides an indication that the subject has minimal residual disease. In some embodiments, the method includes: (A) obtaining the sequencing data, which was previously generated by sequencing a biological sample of a subject, and the sequencing data includes sequence reads covering the positions of mutations to be monitored (for example, the positions can be determined by sequencing results of the subject's primary tumor to identify positions that provide information for subsequent monitoring of MRD); (B) using at least a first subset of the sequence reads (for example, a first subset that can be selected based on data quality) to determine a first value indicating the expected number of mutations due to sequencing errors present in the sequencing data, the determination including: using the first subset of the sequence reads to determine multiple nucleotide field (NC) error types (for example, different types of mutations that can be observed in NC) for multiple NC error rates (for example, the rate of nucleotides or groups of consecutive nucleotides in a sequence that may have mutated due to sequencing errors); grouping at least some of the multiple NC error rates into multiple NC error rate groups (for example, using PAM clustering; the grouping can be performed by increasing the number of sequence reads used to determine the NC group error rate). to reduce statistical noise, and the error rate is then used to determine the first value); using the NC error rates of at least some of the multiple NC error rates to determine the NC group error rate of the multiple NC error rate groups (for example, the population-weighted average of the NC error rate can be the NC group error rate); and using the NC group error rate to determine the first value indicating the expected number of mutations present in the sequencing data; (C) using at least a second subset of the sequence reads to determine a second value indicating the actual number of mutations present at the location of the mutation to be monitored (for example, the number of mutations present at the location of the mutation to be monitored); and (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored (for example, using a one-sided Poisson test, in which case the first value is λ) to determine whether the sequencing data provides the indication that the subject has minimal residual disease (for example, the probability or possibility that the subject has MRD).

[0049] Thus, some embodiments provide a computer-implemented method for determining whether sequencing data of a biological sample (eg, plasma) of a subject (eg, a human subject) provides an indication that the subject has minimal residual disease. In some embodiments, the method includes: (A) obtaining the sequencing data, which was previously generated by sequencing a biological sample of a subject, and the sequencing data includes sequence reads covering the positions of mutations to be monitored (for example, the positions can be determined by analyzing the sequencing results of the subject's primary tumor to identify positions that provide information for subsequent monitoring of MRD); (B) using at least a first subset of the sequence reads (for example, a first subset that can be selected based on data quality) to determine a first value indicating the expected number of mutations due to sequencing errors present in the sequencing data, the determination including: using the first subset of the sequence reads to determine multiple TNC error rates corresponding to multiple trinucleotide field (TNC) error types (for example, different types of mutations that can be observed in TNC) (for example, the rate of trinucleotide repeat sequences that may be mutated due to sequencing errors); grouping at least some of the multiple TNC error rates into multiple TNC error rate groups (for example, using PAM clustering; the grouping can be performed by increasing the number of sequence reads used to determine the error rates of the TNC groups); (c) using at least a second subset of the sequence reads to determine a second value indicating the actual number of mutations present at the location of the mutation to be monitored (e.g., the number of mutations present at the location of the mutation to be monitored); and (d) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored (e.g., using a one-sided Poisson test, in which case the first value is λ) to determine whether the sequencing data provides the indication that the subject has minimal residual disease (e.g., the probability or likelihood that the subject has MRD).

[0050] In some embodiments, coverage and / or resolution play an important role in determining the optimal field size (e.g., NC) to determine the error rate, for example, in such cases, coverage refers to the maximum number of observations of the error rate field under the sequencing depth of a given sample, and resolution refers to the total number of error rate fields of a given size. Sometimes, a larger field size will produce more fields, for example, following the formula (N=3*4^k), where "k" is the field size. More fields (i.e., higher resolution) generally allow for more accurate estimates of the error rate driven by sequence deviations around variants. This sometimes directly and proportionally affects potential coverage (depth / N). For example, in the case of a sample minimum sequencing depth of 10,000 reads, a trinucleotide field can theoretically detect an error rate as low as 1 / 52 (1.9%) on average, while improving the overall resolution relative to a dinucleotide or mononucleotide field. The inventors herein have found that a trinucleotide field is generally the maximum field size that produces an acceptable detectable error rate at many sequencing depths, and therefore generally provides a technical advantage over larger or smaller nucleotide fields.

[0051] In some embodiments, the biological sample can be plasma and can include cell-free DNA and / or ctDNA. Various aspects of biological samples are described herein, including what is referred to below as a "biological sample."

[0052] In some embodiments, the subject can be a human subject who has been previously treated for cancer (eg, lung cancer).Various subjects and cancer types are described herein, including those referred to below as "subjects."

[0053] In some embodiments, a statistical test (e.g., a statistical hypothesis test, which can be a one-sided or two-sided test, and can be, for example, a Poisson test) can be used to determine an indication of minimal residual disease. Various aspects of statistical tests are described herein. In some embodiments, minimal residual disease (MRD) can be used as an indicator of cancer recurrence, typically occurring before standard monitoring imaging detects cancer recurrence. Various aspects of minimal residual disease (MRD) are described herein, including portions referred to below as "minimal residual disease (MRD)".

[0054] In some embodiments, obtaining the sequencing data may include sequencing the nucleic acid in the biological sample to obtain the sequencing data. Various aspects of sequencing data are described herein, including the portion referred to below as "sequencing data". In some embodiments, the sequence reads may cover regions (e.g., background regions) of non-mutations to be monitored in addition to the sequences of mutations to be monitored. Thus, sequencing data from a biological sample of a subject may include sequence reads that can be used to monitor the positions of mutations to be monitored and to determine sequencing errors. In some embodiments, the positions of mutations to be monitored may be determined by previously sequencing the primary tumor of the subject. Various aspects of the positions of mutations to be monitored are described herein, including the portion referred to below as "positions of mutations to be monitored". In some embodiments, mutations at at least 10, 10-200, or 50-200 positions are monitored. Sequencing data may be obtained using a suitable method. In some embodiments, the sequencing data may be obtained using whole genome sequencing. In some embodiments, the sequencing data may be obtained using whole exome sequencing. In some embodiments, the sequencing data may be obtained using a targeted gene sequencing kit or method. Various aspects of the targeted gene sequencing kit are described herein, including the portion referred to below as "sequencing data".

[0055] As described above, in some embodiments, the method includes (B) using at least a first subset of the sequence reads to determine a first value indicating the expected number of mutations due to sequencing errors present in the sequencing data. In some embodiments, the first subset of sequence reads includes any number of sequence reads or a combination of sequence reads in the sequencing data. For example, the first subset may include a consensus sequence read. In some embodiments, the consensus sequence reads may be identified or determined using a suitable alignment method (e.g., Pileup, Bowtie, BarraCUDA, BFAST, CUSHAW, ELAND, FASTA, SOAP, etc. or variants or combinations thereof). The consensus sequence reads may be determined using multiple sequence reads with the same unique molecular identifier (e.g., barcode). Specifically, the first subset may include a deep consensus sequence read. In some embodiments, a deep consensus sequence read can be a consensus read determined by using at least 2 sequence reads, at least 3 sequence reads, at least 4 sequence reads, at least 5 sequence reads, at least 6 sequence reads, at least 7 sequence reads, at least 8 sequence reads, at least 9 sequence reads, at least 10 sequence reads, at least 15 sequence reads, or at least 20 sequence reads having the same unique molecular identifier.

[0056] In some embodiments, the first value may be determined using sequence reads covering one or more positions of multiple mutations to be monitored (e.g., cancer-associated mutations). However, in other embodiments, the first value may be determined using one or more sequence reads each covering one or more positions of multiple mutations to be monitored (e.g., cancer-associated mutations) and one or more sequence reads each not covering any position of multiple mutations to be monitored. However, in other embodiments, the first value may be determined using only sequence reads that do not cover any position of cancer-associated mutations to be monitored.

[0057] In some embodiments, performing (B) includes generating a consensus sequence read using at least the first subset of the sequence reads, wherein each of the consensus sequence reads can be generated from a plurality of sequence reads in at least the first subset of the sequence reads, the plurality of sequence reads being associated with a corresponding common unique molecular identifier (UMI). In some embodiments, using UMIs to generate consensus reads can mitigate polymerase chain reaction amplification bias that may occur during preparation of samples for sequencing. Various aspects of consensus sequence reads are described herein. In some embodiments, each of the consensus sequence reads can be generated from at least a threshold number (e.g., 2-20) of sequence reads associated with a corresponding common UMI. In some embodiments, the method includes selecting a subset of consensus sequence reads, wherein determining the plurality of NC error rates (e.g., TNC error rates) of the corresponding plurality of NC or trinucleotide field (TNC) error types is performed using only the selected subset of consensus sequence reads. In some embodiments, the consensus sequence reads include a positive strand consensus sequence read and a negative strand consensus sequence read, and wherein selecting the subset can be performed based on a measure of similarity between the corresponding positive strand consensus sequence read and the negative strand consensus sequence read. In some embodiments, the consensus sequence reads include a positive strand consensus sequence read and a negative strand consensus sequence read, and a subset of the consensus reads is selected using one or more criteria applicable to the positive strand consensus sequence read and the negative strand consensus sequence read. In some embodiments, the method includes using a suitable statistical test to determine a measure of similarity between the corresponding positive strand consensus sequence read and the negative strand consensus sequence read. In some embodiments, the consensus sequence reads include a positive strand consensus sequence read and a negative strand consensus sequence read, and wherein selecting the subset can be performed based on the relative number of positive strand consensus sequence reads and corresponding negative strand consensus sequence reads. Therefore, by including a threshold and making these selections, in some embodiments, the consensus reads are more likely to represent the actual DNA sequence from the subject.

[0058] In some embodiments, a nucleotide field (NC) refers to a specific base in a nucleic acid sequence or a sequence read. In some embodiments, a nucleotide field (NC) refers to a series of continuous nucleic acids (e.g., 2, 3, 4, 5, 6, 7, 8 or more continuous nucleic acids) with a specific base in a nucleic acid sequence or a sequence read. NC error type refers to a specific mutation in any given NC (with reference to wild type and / or reference genome). For example, for any single NC, there are three possible error types (e.g., for A nucleotides, there are A>T, A>C or A>G). In another example, for a dinucleotide field with a wild-type sequence AT, NC error types may include but are not limited to AA, AG, AC, GT, GA, GC, GG, CT, CA, CC, CG, TT, TA, TC and TG. NC error rate may refer to the frequency of each NC error type occurring in NC. In some embodiments, only one NC error type or multiple NC error types may be considered for selection.

[0059] A trinucleotide field (TNC) refers to a series of three consecutive nucleic acids (e.g., TAC) with specific bases in a nucleic acid sequence or sequence read. Various aspects of a trinucleotide field are described herein, including the portion referred to below as a "trinucleotide field (TNC)". A TNC error type refers to a specific mutation in any given TNC (with reference to wild type and / or reference genome). For example, where the expected (wild type) TNC is TAC, then the TNC error type may include, but is not limited to, TTC, TCC, AAC, TAG, CAC, GAC, TAT, TAA, and TGC. TNC error types are further described herein, including with reference to Figure 3 The TNC error rate may refer to the frequency of each TNC error type appearing in the field of TNC. In some embodiments, only a corresponding plurality of TNC error types may be considered. For example, in some embodiments, a corresponding plurality of TNC error types may refer to TNC error types in which a sudden change occurs at the middle position of the TNC and a sudden change may not occur at the first (3') position and the third (5') position.

[0060] In some embodiments, using sequence reads (e.g., consensus sequence reads) to determine the multiple TNC error rates for corresponding multiple trinucleotide field (TNC) error types includes: using a background region of the consensus sequence reads to determine the multiple TNC error rates, wherein the position of the mutation to be monitored includes a first position, wherein the consensus sequence reads include a first consensus sequence read covering the first position, and the background region includes a first background region of the first consensus sequence read; wherein the first background region includes nucleotides in the first consensus sequence read that are at least a first threshold distance away from the first position. Therefore, the TNC error rate can be determined using TNCs that are a certain threshold distance away from the position of the mutation to be monitored. In some embodiments, the threshold distance is used to exclude TNCs with high errors (e.g., it is known that the confidence of nucleotides located at the end of a sequence read can be lower than the confidence of nucleotides located at the beginning of a sequence read). Reference is made herein and in the following sections. Figure 3 Aspects of TNC error rate and background region are described. In some embodiments, the background region does not include the position of the mutation to be monitored.

[0061] In some embodiments, for a first position among the positions of mutations to be monitored, the consensus sequence reads include a first group of positive-strand consensus sequence reads associated with a positive-strand primer binding sequence and a second group of negative-strand consensus sequence reads associated with a negative-strand primer binding sequence, the positive-strand primer binding sequence being located at the 3' end of each positive-strand consensus sequence read in the first group, and the negative-strand primer binding sequence being located at the 3' end of the negative-strand consensus sequence read in the second group, wherein using the consensus sequence reads to determine the multiple TNC error rates of the corresponding multiple trinucleotide field (TNC) error types includes: determining the multiple TNC error rates using: nucleotides within a second threshold distance range of the positive-strand primer binding sequence in any sequence read in the first group of positive-strand consensus sequence reads, and nucleotides within a third threshold distance range of the negative-strand primer binding sequence in any sequence read in the second group of negative-strand consensus sequence reads. Thus, the TNC error rate can be determined using a TNC at a threshold distance from the start of the sequence read, as determined by the location of the positive strand primer binding sequence or the negative strand primer binding sequence. Various aspects of the second and third thresholds are described herein, including in the following sections with reference to Figure 3 Give a description.

[0062] In some embodiments, using the consensus sequence reads to determine the plurality of TNC error rates for the corresponding plurality of trinucleotide field (TNC) error types comprises: determining the plurality of TNC error rates based on a background region of the consensus sequence reads, wherein the consensus sequence reads comprise a first consensus sequence read, and the background region comprises a first background region of the first consensus sequence reads, wherein the TNC error rate is determined based on how frequently each of the TNC error types occurs in the first background region of the first consensus sequence. Thus, in some embodiments, a TNC error rate may be calculated using only the background region of the sequence reads. In other embodiments, a TNC error rate may be calculated using both the background region and the positions of mutations to be monitored; such embodiments rely on the understanding that the vast majority of TNCs in the sequence reads are not monitored for mutations, and therefore the number of positions of mutations to be monitored may be so small that, in some embodiments, including them in determining sequencing errors may not substantially change sequencing errors.

[0063] In some embodiments, the method includes identifying and / or removing positions in background regions where the error rate is >= 0.5%, >= 1%, >= 1.5%, >= 2%, or >= 3%. This can remove positions that may bias estimates of background errors (e.g., positions where the genomic sequence of a biological sample differs from a reference genomic sequence). In some embodiments, a reference genomic sequence is a genomic sequence for a specific patient (e.g., the patient's wild-type sequence). In some embodiments, a reference genomic sequence is a reference genome (e.g., hg19).

[0064] In some embodiments, the method further includes: after determining the multiple trinucleotide field (TNC) error rates, and before grouping at least some of the multiple TNC error rates into multiple TNC error rate groups, determining a confidence interval for the TNC error rate (e.g., a 99% binomial distribution confidence interval for each TNC error rate); and selecting at least some of the multiple TNC error rates for grouping based on the confidence interval of the TNC error rate when the confidence interval exceeds a threshold (e.g., a 99% binomial distribution confidence interval for all TNC error rates). This step can allow the highest TNC error rate (which may be abnormally high and reduce the accuracy of the algorithm) to be removed before determining the first value. Various aspects of selecting TNC error rates for grouping are discussed herein, including reference to Figure 4 Therefore, in some embodiments, high TNC error rates may be eliminated before determining the first value.

[0065] This article describes various aspects of grouping TNC error rates, including reference Figure 5 Described. In some embodiments, each TNC error rate group includes at least 1 TNC error rate. In some embodiments, grouping at least some of the multiple TNC error rates into multiple TNC error rate groups includes grouping using a suitable clustering method, and non-limiting examples of the suitable clustering method include k-means clustering, hierarchical aggregation clustering, partitioning around midpoints (PAM) clustering, etc. In some embodiments, grouping at least some of the multiple TNC error rates into multiple TNC error rate groups includes grouping using partitioning around midpoints (PAM) clustering. In some embodiments, grouping at least some of the multiple TNC error rates includes grouping into 4 TNC error rate groups. In some embodiments, in order to ensure that there is sufficient sequence read depth to accurately determine the TNC group error rate, grouping the TNC error rates into TNC error rate groups can be performed.

[0066] In some embodiments, determining the first value indicative of the expected number of mutations present in the sequencing data can be performed using at least some of the TNC group error rates and using the number of times at least some of the positions of the mutations to be monitored are covered by sequence reads in the first subset of sequence reads. In some embodiments, the TNC group error rate for a given TNC error rate group can be determined by calculating a population-weighted average of the TNC error rate group. In some embodiments, the number of times at least some of the positions of the mutations to be monitored are covered by sequence reads in the first subset of sequence reads can be determined by counting the number of times each of the positions to be monitored is covered by sequence reads in the subset of sequence reads of the sequencing data. Various aspects of TNC group error rates are described herein, including reference to Figure 5 Described.

[0067] In some embodiments, determining the first value indicative of an expected number of mutations present in sequencing data comprises determining the first value as a weighted linear combination of the TNC error group rates, wherein each specific TNC error group rate in the TNC error group rates is weighted by the number of times the position of the mutation to be monitored is covered by sequence reads corresponding to a TNC error type belonging to the specific TNC error group in the first subset of the sequence reads. Thus, in some embodiments, the first value can be calculated based on information from all of the TNC error rate groups, which increases the amount of data used to estimate the TNC error rate and can improve the estimation of the TNC error rate.

[0068] The methods described herein for determining and using TNC error group rates can be applied to any suitable NC (eg, single NC, dinucleotide blocks, TNCs, etc.).

[0069] As described above, in some embodiments, the method includes (C) using at least a second subset of the sequence reads to determine a second value indicating the actual number of mutations present at the position of the mutation to be monitored. In some embodiments, the second subset of sequence reads includes sequence reads containing the position of the mutation to be monitored. In some embodiments, the first subset of sequence reads and the second subset of sequence reads can be the same subset of sequence reads. In some embodiments, the second value can be determined based on counting the number of times the position of each mutation to be monitored mutates in the second subset of reads. In some embodiments, performing (C) further includes: using at least the second subset of the sequence reads to generate a second consensus sequence read, wherein each consensus sequence read in the second consensus sequence read can be generated by a sequence read associated with a corresponding common unique molecular identifier (UMI) in at least the second subset of sequence reads, wherein determining the second value indicating the actual number of mutations present at the position of the mutation to be monitored can be performed using the second consensus sequence read. In some embodiments, the second subset of consensus sequence reads can include a consensus read constructed using 2-20 reads. Therefore, similar to the calculation of the first value, in some embodiments, the second value can also be calculated using common reads for the same reasons (e.g., controlling PCR amplification bias).

[0070] As described above, in some embodiments, the method includes (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the position of the mutation to be monitored to determine whether the sequencing data provides an indication that the subject has minimal residual disease. In some embodiments, (D) can be performed using a suitable statistical test with a null hypothesis (e.g., a one-sided Poisson test or a t-test), wherein the distribution associated with the null hypothesis has one or more parameters that depend on the first value. In some embodiments, the distribution can be a Poisson distribution with a mean value (λ), which is set to the first value. In some embodiments, using a statistical test includes determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the null hypothesis. In some embodiments, when the second value indicates that the null hypothesis can be rejected, the subject may have minimal residual disease. In some embodiments, performing the one-sided Poisson hypothesis test includes setting the mean value (λ) of the Poisson distribution to the first value, and determining a measure of the likelihood of observing the actual number of mutations indicated by the second value under the Poisson distribution. In some embodiments, the sequencing data can use the measure of the likelihood from the Poisson test to provide an indication that the subject has minimal residual disease (MRD). In some embodiments, the indication of MRD can be based on a p-value obtained from a statistical test that is lower than a predetermined α (e.g., p-value ≤ 0.01). In some embodiments, the rejection of the null hypothesis of the Poisson test can be an indication of MRD. In some embodiments, failure to reject the null hypothesis of the Poisson test may not be an indication of MRD. In some embodiments, (D) further includes: providing an indication that the subject has minimal residual disease. Various aspects of the indication of MRD are described herein, including a description in the section referred to below as "Indications of Minimal Residual MRD."

[0071] In some embodiments, the presence of MRD is determined using a chi-square test across each monitored position comparing observed to expected (e.g., as in (D)). In some embodiments, positions without deep alternate observations (DAO) are filtered out.

[0072] In some embodiments, the presence of MRD is determined by subtracting the first value indicating the expected number of mutations present in the sequencing data from the second value indicating the actual number of mutations present at the location of the mutation to be monitored and determining whether the difference is greater than a threshold value. In some embodiments, the threshold value can be zero.

[0073] In some embodiments, any embodiment of this method can be repeated as described above using additional biological samples collected from the subject. In some embodiments, the methods described herein can encompass a method of monitoring MRD, the method comprising obtaining one or more additional sequencing data previously generated by sequencing one or more additional biological samples from the subject over time, and analyzing these samples according to the above embodiments. Therefore, in some embodiments, the MRD of a subject can be monitored gradually over a period of weeks, months, years, or for the remainder of the subject's life. In some embodiments, the method further comprises, when MRD is identified, referring the subject for treatment, and when MRD is not identified, not referring the subject for treatment.

[0074] In some embodiments, the method includes: obtaining second sequencing data previously generated by sequencing a second biological sample (e.g., an additional biological sample) of the subject, the second sequencing data including second sequence reads covering the location of the mutation to be monitored; using at least a first subset of the second sequence reads to determine a third value indicating the expected number of mutations due to sequencing errors present in the second sequencing data (e.g., in the same manner as determining the first value, but using the second sequencing data), the determination including: determining a second plurality of trinucleotide field (TNC) error rates corresponding to a plurality of TNC error types; grouping at least some of the second plurality of TNC error rates into a second plurality of TNC error rate groups; using at least some of the second plurality of TNC error rates to determine a third value indicating the expected number of mutations due to sequencing errors present in the second sequencing data (e.g., in the same manner as determining the first value, but using the second sequencing data); C error rates to determine a second TNC group error rate of the second plurality of TNC error rate groups; and using the second TNC group error rate to determine the third value indicating the expected number of mutations present in the second sequencing data (for example, in the same manner as determining the first value, but using the second sequencing data); using at least a second subset of the second sequence reads to determine a fourth value indicating the actual number of mutations present at the position; and using the third value indicating the expected number of mutations present in the second sequencing data due to sequencing errors and the fourth value indicating the actual number of mutations present in the second sequencing data at the position of the mutation to be monitored to determine whether the second sequencing data provides the indication that the subject suffers from minimal residual disease.

[0075] In some embodiments, the method further includes using the at least one computer hardware processor to perform the following operations: obtaining one or more additional sequencing data (e.g., second sequence data, third sequence data, etc.) previously generated by sequencing one or more additional biological samples of the subject (e.g., a second biological sample, a third biological sample, etc. collected over time), each of the one or more additional sequencing data includes additional sequence reads covering the location of the mutation to be monitored, and for each additional sequence read of the additional sequence reads in the one or more additional sequencing data: using at least a first subset of the additional sequence reads to determine an additional first value indicating an expected number of mutations due to sequencing errors present in the corresponding sequencing data, the determination comprising: determining additional multiple TNC error rates for corresponding multiple trinucleotide field (TNC) error types; and converting the additional multiple TNC error rates into additional first values. At least some of the TNC error rates are grouped into an additional plurality of TNC error rate groups; additional TNC error rates among at least some of the additional plurality of TNC error rates are used to determine additional TNC group error rates for the additional plurality of TNC error rate groups; and the additional TNC group error rates are used to determine the additional first value indicating the expected number of mutations present in the corresponding sequencing data; an additional second value indicating the actual number of mutations present at the position is used to determine the additional second value; and the additional first value indicating the expected number of mutations present in the corresponding sequencing data due to sequencing errors and the additional second value indicating the actual number of mutations present in the corresponding sequencing data at the position of the mutation to be monitored are used to determine whether the corresponding sequencing data provides the indication that the subject suffers from minimal residual disease.

[0076] Some embodiments of the present disclosure provide a method for determining whether sequencing data of a biological sample of a subject has a mutation (e.g., a cancer-associated mutation). In these embodiments, the method may include using at least one computer hardware processor to perform the following operations: (A) obtaining the sequencing data, the sequencing data being previously generated by sequencing the biological sample of the subject, the sequencing data including sequence reads covering the positions of the mutations to be monitored; (B) using at least a first subset of the sequence reads to determine a first value indicating the expected number of mutations due to sequencing errors present in the sequencing data, the determination comprising: using the first subset of the sequence reads to determine a plurality of TNC error rates corresponding to a plurality of trinucleotide field (TNC) error types; grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups; using the plurality of TNC error (c) using at least a second subset of the sequence reads to determine a second value indicating the actual number of mutations present at the locations of the mutations to be monitored; and (d) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the locations of the mutations to be monitored to determine whether the sequencing data provides an indication that the subject has a mutation at the location of the mutation to be monitored.

[0077] The following is a more detailed disclosure of various concepts and embodiments, as well as methods and compositions, related to MRD-indicating described herein.

[0078] Figure 1 is a diagram depicting an illustrative technique 100 for determining and outputting an indication 107 of MRD in a subject 101. Technique 100 involves collecting a biological sample 102 (e.g., plasma) from the subject 101; sequencing polynucleotides (e.g., ctDNA) from the biological sample using a sequencing device 103 to generate sequencing data 104; inputting the sequencing data 104 and the location 106 of the mutation to be monitored into a computing device 105 (e.g., a laptop, a desktop computer, one or more servers, a cloud computing device, a smartphone, a tablet computer, and / or any other suitable computing device), the computing device being configured to execute software 108, which, when executed, causes the computing device 105 to perform the techniques described herein for detecting an indication of MRD.

[0079] Subjects

[0080] In some embodiments, the subject can be a mammal (e.g., a human, non-human primate, dog, cat, horse, goat, sheep, mouse, or rat), bird, reptile, amphibian, fish, or laboratory model organism (e.g., mouse and rat). In some embodiments, the subject can be a human. In some embodiments, the subject can be an adult human (e.g., over 18 years of age). In some embodiments, the subject can be a human child. In some embodiments, the subject can be a human infant.

[0081] In some embodiments, the subject may be in remission of a disease. In some embodiments, the subject may have been treated for a disease (e.g., cancer). In some embodiments, the subject may have been previously treated by one or more of surgery, chemotherapy, radiotherapy, immunotherapy, and / or hormone therapy. In some embodiments, the subject may be in remission of cancer. In some embodiments, the subject may be in remission of lung cancer, brain cancer, liver cancer, kidney cancer, immune cancer, breast cancer, skin cancer, bone cancer, uterine cancer, prostate cancer, testicular cancer, or colon cancer. In some embodiments, the subject may be in remission of non-small cell lung cancer (NSCLC). In some embodiments, the subject may be in remission of small cell lung cancer (SCLC). In some embodiments, the subject may be in remission of lung adenocarcinoma. In some embodiments, the subject may be in remission of squamous cell carcinoma. In some embodiments, the subject may be in remission of melanoma. In some embodiments, the biological sample includes ctDNA released into the subject's body fluid by cancer cells. In some embodiments, the subject may be in remission of a cancer selected from: NSCLC, colorectal cancer (CRC), bladder cancer, pancreatic cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, and hematological cancers (e.g., leukemia, lymphoma, and multiple myeloma). These cancers may be prone to releasing DNA in body fluids.

[0082] In some embodiments, the subject has or is suspected of having cancer or a tumor.

[0083] Biological samples

[0084] like Figure 1As shown, a biological sample 102 is collected from a subject. In some embodiments, the biological sample includes any cell, tissue, biological fluid or bone from a subject or any other part of a subject. In some embodiments, the biological sample includes ctDNA of a subject. In some embodiments, the biological sample can be a tissue biopsy (e.g., a tumor biopsy). In some embodiments, the tissue can be brain tissue, lung tissue, liver tissue, kidney tissue, skin tissue, pancreatic tissue, connective tissue, muscle tissue or neural tissue. In some embodiments, the biological sample can be a biological fluid from a subject. In some embodiments, the biological fluid can be saliva, semen, vaginal secretions, urine, feces, nasal mucus, sweat, earwax, cerebrospinal fluid, blood, serum or plasma from a subject. In some embodiments, the biological sample can be plasma from a subject. In some embodiments, the biological sample from a subject can be blood or a blood product (e.g., serum or plasma), wherein the blood or blood product includes tumor cells and / or ctDNA. In some embodiments, the biological sample (e.g., blood product, blood, plasma or serum) includes cell-free DNA (cfDNA) or ctDNA. Cell-free DNA from a subject with cancer or a tumor generally includes ctDNA.

[0085] In some embodiments, the biological sample can be collected from any part of the subject's body, including, but not limited to, hair, skin (including portions of the epidermis, dermis, and / or hypodermis), oropharynx, hypopharynx, esophagus, stomach, bronchus, salivary glands, tongue, oral cavity, nasal cavity, vaginal cavity, anal cavity, bone, bone marrow, brain, thymus, spleen, small intestine, appendix, colon, rectum, anus, liver, bile duct, pancreas, kidney, ureter, bladder, urethra, uterus, vagina, vulva, ovary, cervix, scrotum, penis, prostate, testicle, seminal vesicle, any body fluid [such as blood (e.g., whole blood, serum, or plasma), saliva, tears, synovial fluid, cerebrospinal fluid, pleural effusion, pericardial fluid, ascites, and / or urine] and / or any type of tissue (e.g., muscle tissue, epithelial tissue, connective tissue, or neural tissue).

[0086] Any biological sample described herein can be collected using any suitable technique, including the methods described in Vaught et al., Cancer Epidemiol Biomarkers Prev. 2012 Feb;21(2):253-5) in Biospecimens and biorepositories: from afterthought to science and Vaught and Henderson (IARC Sci Publ. 2011;(163):23-42) in Biological sample collection, processing, storage and information management.

[0087] In some embodiments, multiple biological samples can be collected from a subject and sequenced to obtain sequencing data. In some embodiments, the multiple biological samples can be collected from a subject in sequence within a specified time period, and then the multiple biological samples can be sequenced to obtain sequencing data. In some embodiments, the specified time period can start after the end of cancer treatment and can continue for the rest of the subject's life. In some embodiments, the frequency of collecting biological samples from a subject can be any frequency suitable for monitoring MRD. In some embodiments, biological samples can be collected from a subject weekly. In some embodiments, biological samples can be collected from a subject about twice a month. In some embodiments, biological samples can be collected from a subject about once a month. In some embodiments, biological samples can be collected from a subject about once every three months. In some embodiments, biological samples can be collected from a subject about once every six months. In some embodiments, biological samples can be collected from a subject at least twice a month. In some embodiments, biological samples can be collected from a subject at least once a month. In some embodiments, biological samples can be collected from a subject at least once every three months. In some embodiments, biological samples can be collected from a subject at least once every six months. In some embodiments, the frequency with which biological samples may be collected from the subject may be determined based on the type of disease (e.g., cancer type) to be monitored in the subject, the expected likelihood of recurrence, and the rate of disease progression following recurrence.

[0088] Biological specimens can be stored using suitable methods. In some embodiments, the biological specimen can be stored using cryopreservation. Non-limiting examples of cryopreservation include, but are not limited to, progressive freezing, blast freezing, direct immersion freezing, rapid freezing, slow freezing using a programmable freezer, and vitrification. In some embodiments, the biological specimen can be stored using freeze drying. In some embodiments, the biological specimen can be placed in a container already containing a preservative (e.g., RNALater for preserving RNA), and then frozen (e.g., by rapid freezing) after the biological specimen is collected from a subject. In some embodiments, this storage in a frozen state can be performed immediately after the biological specimen is collected. In some embodiments, before freezing, the biological specimen can be kept at room temperature or 4° C. for a certain time (e.g., up to one hour, up to 8 hours, or up to 1 day or several days) in a preservative or a buffer without a preservative.

[0089] Non-limiting examples of preservatives include formalin solution, formaldehyde solution, RNALater or other equivalent solutions, TriZol or other equivalent solutions, DNA / RNA Shield or equivalent solutions, EDTA (e.g., Buffer AE (10 mM Tris·Cl; 0.5 mM EDTA, pH 9.0) and other coagulants, and citrate dextrose (e.g., for blood samples).

[0090] In some embodiments, special containers can be used to collect and / or store biological samples. For example, vacuum containers can be used to store blood. In some embodiments, vacuum containers can include preservatives (e.g., coagulants or anticoagulants). In some embodiments, for the purpose of better preservation or to avoid contamination, the container in which the biological sample is stored can be contained in a secondary container.

[0091] Any of the biological samples from the subject described herein can be stored under any conditions that maintain the stability of the biological sample. In some embodiments, the biological sample can be stored at a temperature that maintains the stability of the biological sample. In some embodiments, the sample can be stored between 18°C ​​and 28°C (e.g., 25°C). In some embodiments, the sample can be stored under refrigeration (e.g., 4°C). In some embodiments, the sample is stored under freezing conditions (e.g., -20°C). In some embodiments, the sample can be stored under ultra-low temperature conditions (-50°C to -800°C). In some embodiments, the sample can be stored under liquid nitrogen (e.g., -1700°C). In some embodiments, the biological sample can be stored at -60°C to -80°C (e.g., -70°C) for up to 5 years. In some embodiments, the biological sample can be stored at -60°C to -80°C (e.g., -70°C) for up to 1 month, up to 2 months, up to 3 months, up to 4 months, up to 5 months, up to 6 months, up to 7 months, up to 8 months, up to 9 months, up to 10 months, up to 11 months, up to 1 year, up to 2 years, up to 3 years, up to 4 years, or up to 5 years. In some embodiments, the biological sample can be stored as described in any of the methods described herein for up to 5 years, up to 10 years, up to 15 years, or up to 20 years.

[0092] Sequencing equipment

[0093] like Figure 1 As shown, the biological sample 102 can be sequenced using a sequencing device 103. In some embodiments, the sequencing device 103 can be any suitable next generation sequencing device or any high throughput or massively parallel sequencing device. In some embodiments, the sequencing device 103 can include any suitable sequencing device and / or any sequencing system including one or more devices. In some embodiments, the sequencing device for sequencing the biological sample can be selected from any suitable platform known in the art, including but not limited to Nanopore-based sequencing, Sanger sequencing, or 454TM. In some embodiments, the sequencing device used to sequence the biological sample is an Illumina sequencing device (e.g., or ).

[0094] Sequencing data

[0095] like Figure 1 As shown, the sequencing platform 103 processes the sample 102 to generate sequencing data 104 .

[0096] In some embodiments, sequencing data 104 may include sequence reads (e.g., the positive strand and the negative strand) of a polynucleotide sequence of the biological sample from the subject. In some embodiments, a sequence read may include a nucleotide representation of the nucleotides of the polynucleotide sequence. In some embodiments, the nucleotide representation may have any reasonable form, including but not limited to an alphabetical representation, a numerical representation, an alphanumeric representation, or a symbolic representation. In some embodiments, the sequence read comprises A (representing adenosine), C (representing cytosine), G (representing guanosine), and T (representing thymine).

[0097] In some embodiments, sequence data includes sequence reads annotated with additional information (e.g., mapped position, length, sample source, collection date, etc.). In some embodiments, sequence reads are mapped to a reference sequence (e.g., a reference genome), thereby providing mapped sequence reads. Thus, in some embodiments, sequence data includes a plurality of annotated sequence reads. Annotated sequence reads and / or sequence data can be provided in a suitable digital format (e.g., a BAM file).

[0098] In some embodiments, the sequencing data may include sequence reads of any suitable polynucleotides of the biological sample. In some embodiments, the sequencing data includes sequence reads of RNA of the biological sample. In some embodiments, the sequencing data includes sequence reads of DNA of the biological sample. In some embodiments, the sequencing data includes sequence reads of tumor DNA of the biological sample. In some embodiments, the sequencing data includes sequence reads of cell-free DNA (e.g., from healthy cells and cancer cells (e.g., ctDNA)). In some embodiments, the sequencing data includes sequence reads of circulating tumor DNA (ctDNA) of the biological sample. In some embodiments, the sequencing data includes sequence reads of whole exome sequencing of the biological sample. In some embodiments, the sequencing data includes sequence reads of whole genome sequencing of the biological sample. In some embodiments, the sequencing data includes sequence reads covering the position of the mutation to be monitored (e.g., a position associated with MRD). In some embodiments, the sequencing data includes sequence reads obtained using a targeted gene sequencing kit.

[0099] It should be understood that a sequence read is data generated by sequencing a polynucleotide (e.g., by using a sequencing device). Thus, a sequence read does not include a physical molecule, but rather includes data representing the physical molecule. Thus, a reference to a nucleotide in a sequence read is a reference to information about the nucleotide (e.g., information representing the type of the nucleotide - such as "A" or "G" or "C" or "T").

[0100] In some embodiments, the sequence read includes a positive strand primer binding sequence and a negative strand primer binding sequence. In some embodiments, the positive strand primer binding sequence and the negative strand primer binding sequence can be complementary to primers used to amplify polynucleotides. In some embodiments, the positive strand primer binding sequence and the negative strand primer binding sequence are determined (e.g., when generating sequencing data) to design positive strand sequence primers and negative strand sequence primers for amplifying specific polynucleotides (e.g., polynucleotides including positions of mutations to be monitored). For example, for each polynucleotide including a position of a mutation to be monitored, a positive strand primer binding sequence and a negative strand primer binding sequence can be determined, in which case the positive strand primer binding sequence and the negative strand primer binding sequence flank the position of the mutation to be monitored (e.g., the 3' and 5' ends of the position).

[0101] In some embodiments, the positive strand primer binding sequence is within 50 nucleotides, within 40 nucleotides, within 30 nucleotides, within 20 nucleotides, within 10 nucleotides, or within 5 nucleotides of the 3' end of the sequence read. In some embodiments, the positive strand primer binding sequence is within 50 nucleotides, within 40 nucleotides, within 30 nucleotides, within 20 nucleotides, within 10 nucleotides, or within 5 nucleotides of the 5' end of the sequence read. In some embodiments, the negative strand primer binding sequence is within 50 nucleotides, within 40 nucleotides, within 30 nucleotides, within 20 nucleotides, within 10 nucleotides, or within 5 nucleotides of the 3' end of the sequence read. In some embodiments, the negative strand primer binding sequence is within 50 nucleotides, within 40 nucleotides, within 30 nucleotides, within 20 nucleotides, within 10 nucleotides, or within 5 nucleotides of the 5' end of the sequence read. In some embodiments, the positive strand primer binding sequence and the negative strand primer binding sequence are 15-30 nucleotides in length.

[0102] In some embodiments, a target gene sequencing kit can be used to specifically sequence only certain polynucleotides from the biological sample (e.g., a polynucleotide comprising a position of a mutation to be monitored). In some embodiments, a target gene sequencing kit can include a representation of a polynucleotide sequence comprising a position of a mutation to be monitored. In some embodiments, the representation of a polynucleotide sequence can be used to determine a polynucleotide for amplification (e.g., using a polymerase chain reaction (PCR)). In some embodiments, amplification can be accomplished using PCR primers (e.g., a positive strand sequence primer and a negative strand sequence primer) that are complementary to a polynucleotide comprising a position of a mutation to be monitored. In some embodiments, the positive strand sequence primer is complementary to a positive strand primer binding sequence of the polynucleotide corresponding to a position of a mutation to be monitored. In some embodiments, the negative strand sequence primer is complementary to a negative strand primer binding sequence of the polynucleotide corresponding to a position of a mutation to be monitored. Methods for designing primers for amplifying specific polynucleotides are well known in the art, for example, as described in Untergasser, Andreas et al., Nucleic Acids Research 40.15 (2012): e115-e115. In some embodiments, primers are designed using the ArcherDx suite design algorithm. The primers may be designed using other suitable suite design algorithms and may be used without departing from the scope of the present invention.

[0103] In some embodiments, the amplification method used with the targeted gene sequencing kit is anchor-multiplex PCR (AMP). AMP is a multiplex PCR enrichment chemistry that introduces strand-specific primers and incorporates unique molecular identifiers (UMIs) into sequenced reads, which is well known in the art, for example, as described in Zheng Z et al., Nature Medicine 20.12 (2014): 1479-1484.

[0104] Location of mutation to be monitored

[0105] like Figure 1As shown, the indication 104 of the position of the mutation to be monitored is input into the computer 105 and the software 108. In some embodiments, the indication of the position of the mutation to be monitored may include a representation of a polynucleotide containing a disease-associated mutation (e.g., a cancer-associated mutation). In some embodiments, the disease-associated mutation may be the position of the mutation to be monitored. In an embodiment, the position of the mutation to be monitored is a set of predetermined positions in a reference genome (e.g., hg19) or a portion thereof. In some embodiments, the position of the mutation to be monitored is a position in a reference sequence, in which case the reference sequence may be completely arbitrary, it may be a reference genome assembly or a custom reference. In some embodiments, the position of the mutation to be monitored corresponds to the position where a mutation occurs in the genome of the tumor of the subject. In some embodiments, the position of the mutation to be monitored may be a position associated with MRD. In some aspects, the position of the mutation to be monitored may be a position associated with MRD. In some embodiments, the position of the mutation to be monitored may be monitored to determine the indication of MRD.

[0106] In some embodiments, the location of the mutation to be monitored can be determined by methods well known in the art. In some embodiments, the location of the mutation to be monitored can be determined before determining the indication of MRD. In some embodiments, determining the location of the mutation to be monitored includes sequencing polynucleotides from diseased cells of the subject. In some embodiments, determining the location of the mutation to be monitored includes sequencing polynucleotides from cancer cells of the subject. In some embodiments, determining the location of the mutation to be monitored includes sequencing tumors from the subject. In some embodiments, determining the location of the mutation to be monitored includes sequencing primary tumors from the subject. In some embodiments, determining the location of the mutation to be monitored includes sequencing secondary tumors from the subject. In some embodiments, determining the location of the mutation to be monitored includes sequencing ctDNA from the subject before treatment (e.g., before surgery). Thus, in some embodiments, the location of the mutation to be monitored can be specific to a given subject. In some embodiments, the location of the mutation to be monitored can be determined using the ArcherDx set design algorithm. Without departing from the scope of the present invention, other suitable set design algorithms can be used.

[0107] In some embodiments, the number of positions to be monitored for mutations includes at least 10 positions, at least 25 positions, at least 50 positions, at least 75 positions, at least 100 positions, at least 125 positions, at least 150 positions, at least 175 positions, at least 200 positions, at least 250 positions, or at least 300 positions. In some embodiments, the number of positions to be monitored for mutations can be 10-200 positions. In some embodiments, the number of positions to be monitored for mutations can be 25-200 positions. In some embodiments, the number of positions to be monitored for mutations can be 50-200 positions. In some embodiments, the number of positions to be monitored for mutations can be 75-200 positions. In some embodiments, the number of positions to be monitored for mutations can be 100-200 positions. In some embodiments, the number of positions to be monitored for mutations can be 10-150. In some embodiments, the number of positions to be monitored for mutations can be 25-150 positions. In some embodiments, the number of positions to be monitored for mutations may be 50-150 positions. In some embodiments, the number of positions to be monitored for mutations may be 75-150 positions. In some embodiments, the number of positions to be monitored for mutations may be 100-150 positions.

[0108] like Figure 1 As shown, the computing device 105 and software 108 obtain the sequencing data, obtain the indication of the position of the mutation to be monitored, and determine the indication of MRD. In some embodiments, the computing device 105 can be one or more computing devices of any suitable type. In some embodiments, the computing device 105 can be but not limited to a laptop computer, a desktop computer, a cloud computer, a server, a phone or a tablet computer. In some embodiments, the computing device 105 can be located in a single physical location or can be located across multiple physical locations. In some embodiments, the computing device 105 can be located in a facility operated by an entity (e.g., a hospital or a research institution). In some embodiments, the software 105 can determine the indication of MRD based on the trinucleotide field (TNC) in the sequence read of the sequencing data.

[0109] In some embodiments, the computing device 105 can be operated by a user such as a researcher, patient, doctor, clinician, or other individual. For example, the user can provide the sequencing data 104 as input to the computing device 105 (e.g., by uploading a file).

[0110] Trinucleotide Nucleotide Field (TNC)

[0111] In some embodiments, TNC can be a series of three consecutive nucleotides in a sequence read. Typically, the nucleotides in a sequence read can include, but are not limited to, A (representing adenine), C (representing cytosine), G (representing guanine), and T (representing thymine). In some embodiments, a trinucleotide field can be any arrangement of any three of A, C, T, and G. For example, TNC can be, but is not limited to, any one of GAT, TTC, TAT, CAT, TTT, and TAA. In some embodiments, a sequence read includes multiple TNCs. In some embodiments, a given nucleotide representation associated with a given position included in a first TNC may not be included in a second TNC. In some embodiments, TNC corresponds to an amino acid codon and / or anticodon. In some embodiments, TNC can be used to determine the first value indicating the expected number of mutations due to sequencing errors present in the sequencing data, which can then be used when determining an indication of MRD. One advantage of using TNC to calculate error rates over using single positions is that the error rate can be estimated more accurately when more data are available. For example, trinucleotide errors in TNC can provide richer information than single nucleotide errors (Deng, Shibing et al., BMC Bioinformatics 19.1 (2018): 1-7).

[0112] Minimal residual disease (MRD)

[0113] Minimal residual disease refers to any residual disease (e.g., diseased cells or ctDNA) that may be present in a subject after the subject has received or completed treatment for the disease. For example, minimal residual disease associated with cancer may be present when cancer cells, cancer RNA, and / or circulating tumor DNA (e.g., ctDNA) are present in the subject after treatment. In some embodiments, MRD can be detected based on ctDNA detection before detecting cancer recurrence using standard monitoring imaging (e.g., computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET)). Some cancer types can cause DNA (e.g., ctDNA) to fall off, and the DNA eventually reaches the bloodstream of the subject. Thus, in some embodiments, minimal residual disease can be monitored based on sequencing of ctDNA from a biological sample (e.g., plasma) from blood. In some embodiments, the likelihood or probability of an indication of minimal residual disease can increase over time. For example, because cancer cells that survive treatment can continue to replicate and / or metastasize, this may cause additional ctDNA to fall off.

[0114] Indications for minimal residual disease (MRD)

[0115] Any technology described herein can be used to determine the indication of MRD. In some embodiments, the indication of MRD may include any information that provides an estimate of the possibility and / or probability that MRD may exist. For example, the indication of MRD can be a letter representation, a digital representation, a symbolic representation, or an alphanumeric representation of the possibility and / or probability of MRD existence. In a further example, the indication of MRD can be based on a scale of 0-1, where zero indicates that the subject may have the lowest possibility or probability of MRD, and 1 indicates that the subject may have the highest possibility or probability of MRD. In some embodiments, the indication of MRD can be binary (e.g., "yes" or "no"; "true" or "false"; etc.). In some embodiments, the indication of MRD can be an instruction to perform additional tests to confirm the indication of MRD. In some embodiments, the indication of MRD can be derived based on sequencing data from multiple biological samples of the same subject. For example, the indication of MRD can be determined when sequencing data from at least two biological samples of the subject are analyzed. In another example, an indication of MRD can be determined when analysis of sequencing data of at least one of two or more biological samples from the subject indicates MRD.

[0116] In some embodiments, an indication of MRD can be an estimate of the likelihood and / or probability of the presence of MRD in a biological sample of a subject based on a statistical test that compares the sequencing error in the analysis of the biological sample (e.g., a first value) to the actual number of mutations at the locations of the mutations to be monitored present in the biological sample (e.g., a second value). In some embodiments, an indication of minimal residual disease can be derived based on the number of locations of mutations to be monitored.

[0117] In some embodiments, a statistical test can be used to determine the indication of MRD. In some embodiments, the statistical test can be selected from the group consisting of: Poisson test, binomial test, T test or any other suitable statistical test. In some embodiments, the statistical test can be a non-parametric test (e.g., Wilcoxon rank-sum test). In some embodiments, the statistical test can be a one-sided test. In some embodiments, the statistical test can be a one-sided Poisson test. In some embodiments, the statistical test can compare the null distribution based on the first value with the alternative hypothesis based on the second value. In some embodiments, the indication of MRD can be derived based on the p value of the statistical test being less than a predetermined value α. In some embodiments, the α can be at most 0.2, at most 0.15, at most 0.1, at most 0.05, at most 0.01, at most 0.005 or at most 0.001. In some embodiments, α is 0.2, 0.15, 0.1, 0.05, 0.01, 0.005, or 0.001. In some embodiments, α can be 0.01. Various aspects of determining an indication for MRD are described herein, including reference Figure 6 Described.

[0118] Figure 2A A flow chart of an illustrative process 200 for determining an indication of minimal residual disease (MRD) based on sequencing data of a biological sample from the subject is depicted. Process 200 has the following steps: step 201, obtaining the sequencing data including sequence reads covering the positions of mutations to be monitored; step 202, determining a first value indicating the expected number of mutations present in the sequencing data due to background errors; step 203, determining a second value indicating the actual number of mutations present at the positions of the mutations to be monitored; and step 204, using the first value and the second value to determine an indication of minimal residual disease. In some embodiments, process 200 is embodied in a program code (e.g., processor executable instructions) portion of software 108 and is performed using the computing device 105.

[0119] like Figure 2AAs shown, process 200 begins at step 201 of obtaining sequencing data. In some embodiments, the sequencing data can be generated from a biological sample collected from a subject (e.g., a subject who has received treatment for cancer) as described above. In some embodiments, the sequencing data includes sequence reads covering the locations of mutations to be monitored. In some embodiments, sequencing data can be obtained from a data store, a cloud storage device, a compressed disk, an external storage device, a sequencing instrument, a public repository, a desktop computer, a laptop computer, a phone, a tablet computer, a flash drive, or any other suitable source. In some embodiments, the sequencing data can be obtained by generating sequencing data using a sequencing system (e.g., sequencing the biological sample of the subject). Various aspects of sequencing data are described herein, including references to Figure 1 Described.

[0120] In some embodiments, low-quality sequence reads can be identified in the sequencing data before determining the first value and the second value. In some embodiments, low-quality sequence reads can be identified and removed from the sequencing data before determining the first value and the second value. Methods for removing low-quality sequence reads from sequence data are well known in the art, for example, Chen, Shifu et al., BMC Bioinformatics 18.3 (2017): described in 91-100.

[0121] In some implementations, after step 201 and before steps 202 and 203, a consensus sequence read may be generated. In some implementations, the consensus sequence read may be generated based on a unique molecular identifier (UMI) associated with the consensus sequence read. The UMI may be a plurality of unique nucleotide sequences that may be linked to polynucleotide fragments (e.g., DNA fragments from a biological sample) to detect sample preparation bias. Prior to sequencing, polynucleotides extracted from a biological sample (e.g., ctDNA) are typically amplified using a polymerase chain reaction to produce sufficient material for sequencing. However, the polymerase chain reaction may produce amplification bias because some polynucleotides may be amplified more efficiently than other polynucleotides. To control for this, the UMI may be linked to the polynucleotide prior to amplification. During the amplification phase, each polynucleotide-UMI molecule may be copied many times, and regardless of the number of copies made, the UMI may indicate which copies are from the same original polynucleotide. The polynucleotide-UMI may be sequenced to generate sequence reads. In some implementations, a consensus sequence read may be generated by aligning all of the sequence reads having the same UMI and generating a single consensus sequence read. Methods for generating consensus sequence reads are well known in the art, such as the methods described in Chen, Shifu et al., BMC Bioinformatics 20.23 (2019): 1-8. In some implementations, each of the consensus sequence reads may be generated by at least a threshold number of sequence reads associated with a corresponding common UMI. In some implementations, the threshold number of sequence reads associated with a corresponding common UMI may be 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or any other suitable threshold.

[0122] In some embodiments, the subset of consensus sequence reads may be selected prior to step 202 and step 203. In some embodiments, the subset of consensus sequence reads may be selected to facilitate the use of high quality data when calculating the first value and the second value. In some embodiments, selecting the subset of sequence reads may be performed based on a measure of similarity (e.g., complementarity) between the positive strand consensus sequence reads and the negative strand consensus sequence reads. In some embodiments, selecting the subset may be performed based on the relative number of positive strand consensus sequence reads and corresponding negative strand consensus sequence reads. For example, if the proportion of reads associated with a given sequence that comes from the positive strand consensus sequence reads or the negative strand consensus sequence reads does not exceed 95%, a subset of the sequence reads is selected. In some embodiments, selecting the subset may be performed based on the relative number of read 1 consensus sequence reads and corresponding read 2 consensus sequence reads. The skilled person will appreciate that some types of sequencing technologies (e.g., ) can sequence a given polynucleotide (e.g., the positive strand or the negative strand) from the 3' end and the 5' end of the polynucleotide, and thereby generate two sequence reads, namely read 1 and read 2.

[0123] Next, process 200 proceeds to steps 202 and 203. Figure 2A As shown, these steps can be performed in parallel. However, this is not necessarily the case, because in some embodiments, these steps can also be performed in sequence (e.g., 202 is performed before 203 or 203 is performed before 202), because the various aspects of the technology described herein are not limited in this regard. Nevertheless, for clarity, step 202 is described first.

[0124] In step 202, a computing device performing process 200 determines a first value indicating an expected number of mutations present in the sequencing data due to sequencing errors. An example of a first value indicating an expected number of mutations is a first value representing an expected number of mutations. In another example, the first value indicating an expected number of mutations is a first value corresponding to the expected number of mutations present in the sequencing data. In some embodiments, determining the first value may include, but is not limited to, determining an indication of an error rate at which mutations are observed at a location of a mutation to be monitored. In some embodiments, the first value is determined using a subset of sequence reads of the sequencing data.

[0125] The first subset of sequence reads may include one, some or all of the sequence reads. In some embodiments, the first subset of sequence reads includes sequence reads covering the positions of mutations to be monitored. It should be understood that sequence reads covering the positions of mutations to be monitored may also cover background areas. In some embodiments, the first subset of sequence reads consists of reads selected based on the data quality filters described herein. In some embodiments, the first subset of sequence reads includes reads associated with positions indicating MRD when a mutation occurs. In some embodiments, the first subset of sequence reads may be the same as the second subset of sequence reads.

[0126] Various aspects of determining the first value are further described herein, including reference to Figure 2BDescribed. In step 203, the computing device performing process 200 determines a second value indicating the actual number of mutations present at the position of the mutation to be monitored. In some embodiments, the second value can be determined based on the number of times the mutation is observed at the position of the mutation to be monitored. In some embodiments, the second value can be determined based on the number of times the mutation is observed at the position of multiple mutations to be monitored. For example, the second value can be determined based on the number of times the mutation is observed at the position of at least 10 mutations to be monitored. In other examples, the second value can be determined based on the number of times the mutation is observed at at least 16 positions, at least 25 positions, at least 50 positions, at least 75 positions, at least 100 positions, at least 125 positions, at least 150 positions, at least 175 positions, at least 200 positions, at least 250 positions, or at least 300 positions. In some embodiments, the second value can be determined based on the number of times mutations are observed at the positions of 10-200, 16-200, 25-200, 50-200, 75-200, 100-200, 125-200, 150-200, 75-200, 25-150, 50-150, 75-150, 100-150, 25-100, 50-100, 75-100 mutations to be monitored. In some embodiments, the second value can be determined based on the number of times mutations are observed at the positions of 16-100 mutations to be monitored. In some embodiments, the second value can be determined based on the number of times mutations are observed at the positions of 50-100 mutations to be monitored. In some embodiments, the second value can be determined based on the number of times mutations are observed at the positions of 50-200 mutations to be monitored. In some embodiments, determining the second value comprises counting the number of mutations present at the position of each mutation to be monitored. In some embodiments, determining the second value can be performed based on the number of mutations present at the position of each mutation to be monitored. In some embodiments, determining the second value can be performed based on the frequency of mutations present at the position of each mutation to be monitored. In some embodiments, determining the second value can be performed using a consensus sequence read. Aspects of the consensus read sequence are further described herein.

[0127] In some embodiments, the second value can be determined based on a second subset of sequence reads. The second subset of sequence reads may include one, some or all of the sequence reads. In some embodiments, the second subset of sequence reads includes only sequence reads covering the position of the mutation to be monitored. In some embodiments, the reads covering the position of the mutation to be monitored may also cover the background area. In some embodiments, the second subset of sequence reads consists of reads selected based on the data quality filter described herein. In some embodiments, the second subset of sequence reads includes reads associated with positions indicating MRD when a mutation occurs. In some embodiments, the second subset of sequence reads may be the same as the first subset of sequence reads. In some embodiments, the second value is based on the number of times the position of the mutation to be monitored mutates. In some embodiments, the second value is the sum of the number of times at least some of the positions of the mutation to be monitored mutate. In some embodiments, the second value is the sum of the number of times each position of the mutation to be monitored mutates.

[0128] After steps 202 and 203 are completed, process 200 proceeds to step 204, in which the information determined during these steps (e.g., the first value and the second value) can be used to determine an indication that the subject has MRD. Examples of determining an indication that a subject has MRD are provided herein, including the portion referred to above as an indication of minimal residual disease (MRD) and referenced Figure 6 .

[0129] Figure 6 Various aspects of the likelihood that an MRD may be present are shown using a statistical test (e.g., a one-sided Poisson test) to use the first value (calculated at step 202 and indicating the expected number of mutations present due to sequencing errors) and the second value (calculated at step 204 and indicating the actual number of mutations present) to determine. In some embodiments, using the statistical test can include determining the likelihood of observing the second value (indicating the actual number of mutations present) assuming that the distribution of mutations due to sequencing errors follows a distribution associated with the null hypothesis of the statistical test. The distribution can be parameterized by one or more parameters. For example, the distribution can be parameterized by one or more parameters that indicate the expected number of mutations present in the sequencing data due to sequencing errors. For example, in some embodiments, the distribution can be a Poisson distribution, and its mean (λ) can be set to the first value indicating the expected number of mutations present in the sequencing data.

[0130] For example, Figure 6As shown, a one-sided Poisson test can be used to determine the likelihood of an observation 602 under the null hypothesis (e.g., the number of actual observations present) where the expected number of mutations present is distributed according to a Poisson distribution (601) whose mean (λ) 603 is set to the first value (calculated at step 202 of process 200) indicating the expected number of mutations due to sequencing error. In some embodiments, the presence of the MRD can be determined when the p-value of the statistical test is less than a threshold value (e.g., 0.1, 0.01, 0.001, 0.0001, etc.).

[0131] Figure 2B A flow chart of an illustrative process 210 for determining an expected number of mutations due to sequencing errors is depicted. The process 210 includes the following steps: step 211, determining a plurality of trinucleotide field (TNC) error rates corresponding to a plurality of TNC error types; step 212, grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups; step 213, determining TNC group error rates for the plurality of TNC error rate groups using the TNC error rates of at least some of the plurality of TNC error rates; and step 214, determining the first value indicative of the expected number of mutations present in the sequencing data using the TNC group error rates.

[0132] At step 211 of process 210, a computing device determines a plurality of TNC error rates corresponding to a plurality of trinucleotide field (TNC) error types. Examples of trinucleotide fields are provided herein, including the portion referred to above as "trinucleotide field (TNC)". In some embodiments, a TNC error rate can be determined based on the number of times a TNC error type is observed in sequencing data. In some embodiments, a TNC error rate can be determined based on the frequency with which a TNC error type is observed in the sequencing data. In some embodiments, determining a plurality of TNC error rates corresponding to a plurality of TNC error types using the first subset of sequence reads comprises determining a frequency of occurrence of each of the TNC error types in the first subset of sequence reads.

[0133] In some embodiments, a TNC error type may be a mutant variant of a wild-type TNC. In some embodiments, a wild-type TNC is a TNC that may occur naturally at a given position in a genome or a portion thereof (e.g., a reference genome such as GRCh38 or hg19). In some embodiments, a mutant variant TNC of a wild-type TNC may be a TNC having a mutation in at least one position of the TNC relative to the wild-type TNC. For example, if the reference genome comprises the following sequence 367-ATGGTACTGCGTACG-381, the wild-type TNC at positions 367-369 will be ATG. In this example, the TNC error type of ATG may be, but is not limited to, TTG, AAG, and ATC. In some embodiments, a TNC error type may depend on the wild-type TNC. For example, a TNC error type of ATG to AAG is different from a TNC error type of ACG to AAG, although the sequence of each mutant variant of the wild-type TNC is the same. In some embodiments, the TNC error type corresponds to a mutation in any position of a given TNC. In some embodiments, a TNC error type includes a mutation located in a middle position of the TNC and does not include a mutation located in a first position or a third position of the TNC (a middle position TNC error type). Various aspects of a middle position TNC error type are described herein, including reference to Figure 3 In some implementations, TNC error rates corresponding to a plurality of TNC error types may be determined.

[0134] Figure 3 An example of determining the TNC error rate is shown. Figure 3 In some embodiments, consensus sequence reads 304 (described below) that all cover the same position 313 of the mutation to be monitored can be compared to a reference sequence 303 (e.g., a known sequence that includes the position of the mutation to be monitored). In some embodiments, the reference sequence is the human reference genome assembly hg19. In some embodiments, the reference sequence is the human reference genome assembly hg38. In some embodiments, the TNC of the compared consensus sequence reads can be identified as being within predetermined background regions 308 and 309. In some embodiments, background region 308 is specific to the negative strand (-) consensus sequence reads, and background region 309 is specific to the positive strand (+) consensus sequence reads. Background region 308 is specified by threshold 1 (305) and threshold 3 (307). Background region 309 is specified by threshold 1 (305) and threshold 2 (306). Additional aspects of background regions are discussed below.

[0135] In some embodiments, determining a TNC error rate comprises counting the number of times the wild-type TNC 300 and the TNC error type 309 appear in a consensus sequence read. In some embodiments, the counting can be performed based on the wild-type sequence 301. For each TNC in each consensus sequence read, a determination can be made as to whether the TNC is a wild-type TNC or a TNC error type by comparing the TNC of the consensus sequence read to the TNC at the same position in the wild-type sequence 301. In some embodiments, a count 314 of each occurrence 310 of the wild-type TNC and each occurrence 310 of each TNC error type can be determined. In some embodiments, a total number 311 of wild-type TNC occurrences and TNC error type occurrences is determined.

[0136] In some embodiments, the TNC error types that are counted are TNC error types 309 that have a mutation in the middle position of the TNC (relative to wild-type TNC) but not in the first (5') position or the last (3') position of the TNC (e.g., middle position TNC error types). For example, middle position TNC error types for ATG can include, but are not limited to, AAG, ACG, and AGG. In some embodiments, there are 192 different middle position TNC error types. The values ​​of the 192 different middle TNC error types can be determined as follows: Based on the four common nucleotides A, T, G, and C, there are 64 possible TNCs because each position of the TNC (first, middle, and last) can have any of A, T, G, or C (e.g., 4 common nucleotides ^ 3 positions). Each middle position of a TNC can be mutated in three different ways (e.g., ATG to AAG, ACG, or AGG). Thus, in some embodiments, there are 64 TNC times 3 possible middle position mutations (e.g., 192 different possible middle position TNC error types).

[0137] In some embodiments, the plurality of TNC error types includes any combination of any number of intermediate position TNC error types 309. In some embodiments, the plurality of TNC error types includes 10, 25, 50, 75, 96, 100, 125, 150, 176, 192, or any other reasonable number of TNC error types. In some embodiments, the plurality of TNC error types includes at least 10, at least 25 TNC error types, at least 50 TNC error types, at least 75 TNC error types, at least 96 TNC error types, at least 100 TNC error types, at least 125 TNC error types, at least 150 TNC error types, at least 176 TNC error types, or at least 192 TNC error types. In some embodiments, the plurality of TNC error types includes 192 intermediate position TNC error types. In some embodiments, the plurality of TNC error types involves 96 intermediate position TNC error types. In some embodiments, a plurality of TNC error types relates to 10-25, 10-75, 10-125, 10-176, 10-192, 25-50, 25-75, 25-125, 25-176, 25-192, 50-75, 50-125, 50-176, 50-192 TNC error types. In some embodiments, a plurality of TNC error types can be determined based on the TNC error types present in the sequence reads. In some embodiments, a TNC error rate for each TNC error type in the plurality of TNC error types is determined.

[0138] In some embodiments, a TNC error rate 312 can be a function of the number of occurrences 310 and the total number 311. In some embodiments, a TNC error rate 312 for a given TNC error type 309 can be the number of occurrences 310 of the TNC error type divided by the total number 311. In some embodiments, the TNC error rate 312 for a given TNC error type 309 can be determined based on the number of occurrences 310 of the TNC error type among all identical wild-type TNCs and the total number 311 among all identical wild-type TNCs. For example, a wild-type TNC, "ATG", can occur in two different positions in a wild-type sequence 300, namely position one and position two. The number of occurrences of the wild-type TNC at position one in the consensus sequence read can be 15, and the number of occurrences of the wild-type TNC at position two can be 19. Additionally, at position one, the number of occurrences of a TNC error type "AGG" in the consensus sequence read can be 2, and at position two, the number of occurrences can be 3. In this example, the TNC error rate may be calculated as (2+3) / (15+19). In some aspects, the TNC error rate 312 for a given TNC error type 309 may be determined using the following formula:

[0139]

[0140] Where n is a TNC with the same sequence but located at a different position in the background region corresponding to the TNC error type (type) (e.g., 3- ATG TTTCATTTG ATG-17 )

[0141] In some embodiments, the TNC error rates include error rates for each TNC error type. In some embodiments, the TNC error rates include error rates for each intermediate position TNC error type. In some embodiments, the TNC error rates include 10, 25, 50, 75, 100, 125, 150, 176, 192, or any other reasonable number of TNC error rates. In some embodiments, the TNC error rates include 10-25, 10-75, 10-125, 10-176, 10-192, 25-50, 25-75, 25-125, 25-176, 25-192, 50-75, 50-125, 50-176, 50-192, or any other reasonable number of TNC error rates. In some embodiments, a plurality of trinucleotide field (TNC) error rates comprises 192 TNC error rates. In some embodiments, a plurality of trinucleotide field (TNC) error rates comprises 96 TNC error rates. In some embodiments, determining a plurality of TNC error rates corresponding to a plurality of TNC error types using the first subset of sequence reads comprises determining a frequency of occurrence of each of the TNC error types in the background region in the first subset of sequence reads.

[0142] In some embodiments, the sequence reads described herein include background regions. In some aspects, background regions can identify regions in the sequence that can be used to determine a first value indicating the expected number of mutations present in the sequencing data due to background errors. In some embodiments, the background region does not include the location of any mutations to be monitored.

[0143] In some embodiments, the background area in the plurality of background areas comprises at least 25 nucleotides, at least 50 nucleotides, at least 100 nucleotides, at least 125 nucleotides, at least 150 nucleotides, at least 175 nucleotides, at least 200 nucleotides, at least 250 nucleotides, at least 300 nucleotides, at least 400 nucleotides, at least 500 nucleotides, at least 750 nucleotides, at least 1000 nucleotides, at least 2000 nucleotides, at least 5000 nucleotides or at least 10000 nucleotides. In some embodiments, the background area comprises at least 400 nucleotides. In some embodiments, the background region comprises at least 10 trinucleotides, at least 25 trinucleotides, at least 50 trinucleotides, at least 100 trinucleotides, at least 125 trinucleotides, at least 150 trinucleotides, at least 175 trinucleotides, at least 200 trinucleotides, at least 250 trinucleotides, at least 300 trinucleotides, at least 400 trinucleotides, at least 500 trinucleotides, at least 750 trinucleotides, at least 1000 trinucleotides, at least 2000 trinucleotides, at least 3000 trinucleotides, at least 5000 trinucleotides or at least 10,000 trinucleotides. In some embodiments, the background region comprises at least 130 trinucleotides.

[0144] In some embodiments, the background region can be identified based on a first threshold, a second threshold, and / or a third threshold.

[0145] In some embodiments, the background area can be identified based on the first threshold value. In some embodiments, the first threshold value sets the nucleotide distance between the position of the sudden change to be monitored and the beginning of the background area. In some embodiments, the first threshold value is 0 nucleotide, which means that the position of the sudden change to be monitored is included in the background area. In some embodiments, the first threshold value is 1 nucleotide, which means that the position of the sudden change to be monitored is not included in the background area. In some embodiments, the first threshold value is 2 nucleotides, which means that the position of the sudden change to be monitored and the nucleotide on either side of the position are not included in the background area. In some embodiments, the first threshold value is 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 50, 100, 150, 200 or more nucleotides. In some embodiments, the first threshold value can be 2-5, 2-10, 2-20, 2-50, 2-100 or 2-200 nucleotides. In some embodiments, the first threshold value can be 5-10, 5-20, 5-50, 5-100 or 5-200 nucleotides. In some embodiments, the first threshold value can be 20-50, 20-100 or 20-200 nucleotides.

[0146] In some embodiments, the background area can be identified based on a second threshold value. In some embodiments, the second threshold value can be specific to the positive strand sequence read. In some embodiments, the second threshold value can be the distance from the start of the positive strand sequence read (e.g., the positive strand primer binding sequence). For example, in the case where the second threshold value is 100 nucleotides, any nucleotide falling within the first 100 nucleotides of the positive strand sequence read can be included in the background area. In some embodiments, the second threshold value can be 50, 100, 150, 200, 250, 300 or 400 nucleotides. In some embodiments, the second threshold value can be at least 50 nucleotides, at least 100 nucleotides, at least 150 nucleotides, at least 200 nucleotides, at least 250 nucleotides, at least 300 nucleotides or at least 400 nucleotides. In some embodiments, the second threshold value can be between 50-100 nucleotides, 100-150 nucleotides, 150-200 nucleotides, 200-250 nucleotides, 250-300 nucleotides, 300-400 nucleotides or 100-400 nucleotides. In some embodiments, the second threshold value can be determined based on the sequencing quality associated with the nucleotides of the sequence read. It is known that sequencing performed near the end of the sequence read can produce data of lower quality than sequencing performed at the beginning of the sequence read. Therefore, the technician can set a second threshold value to remove low-quality nucleotide identifications near the end of the sequence read based on the data quality of the sequence read.

[0147] In some embodiments, the background area can be identified based on a third threshold. In some embodiments, the third threshold may be specific to the negative strand sequence read. In some embodiments, the third threshold may be a distance from the start of the negative strand sequence read (e.g., at the negative strand primer binding sequence). For example, in the case where the third threshold is 100 nucleotides, any nucleotide falling within the first 100 nucleotides of the third sequence read may be included in the background area. In some embodiments, the third threshold may be 50, 100, 150, 200, 250, 300 or 400 nucleotides. In some embodiments, the third threshold may be at least 50 nucleotides, at least 100 nucleotides, at least 150 nucleotides, at least 200 nucleotides, at least 250 nucleotides, at least 300 nucleotides or at least 400 nucleotides. In some embodiments, the third threshold value can be between 50-100 nucleotides, 100-150 nucleotides, 150-200 nucleotides, 200-250 nucleotides, 250-300 nucleotides, 300-400 nucleotides or 100-400 nucleotides. In some embodiments, the third threshold value can be determined based on the sequencing quality associated with the nucleotides of the sequence read. It is known that sequencing near the end of the sequence read will produce data with lower sequencing quality than at the beginning of the sequence read. Therefore, the technician can set a third threshold value based on the data quality of the sequence read to remove low-quality nucleotide identifications near the end of the sequence read. In some embodiments, the first threshold value and the second threshold value are the same number of nucleotides from their corresponding primer binding sequences.

[0148] In some embodiments, after step 211 and before step 212, a TNC error rate may be selected for grouping. In some embodiments, selecting a TNC error rate for grouping may be based on an error rate threshold (e.g., Figure 4401). For example, in some embodiments, if the TNC error rate does not exceed the error rate threshold, the TNC error rate can be selected for grouping. In some embodiments, the error rate threshold is 0.0001%-1%, 0.001%-0.1%, or 0.005%-0.05%. In some embodiments, the error rate threshold is 0.001%, 0.002%, 0.003%, 0.004%, 0.005%, 0.006%, 0.007%, 0.008%, 0.009%, 0.01%, 0.02%, 0.03%, 0.04%, 0.05%, 0.06%, 0.07%, 0.08%, or 0.09%, 0.1%. In some embodiments, the error rate threshold is 0.01%. In some implementations, a TNC error rate may be selected if an upper limit of a confidence interval for the TNC error rate is less than the error rate threshold.

[0149] In some aspects, Figure 4 TNC error rates (e.g., 403 and 405) are shown selected for grouping based on their corresponding 99% binomial confidence interval (CI) upper limits (e.g., 402 and 404). In some embodiments, the error rate threshold 401 is a static threshold (e.g., 0.001%-0.10%). In some embodiments, for each individual TNC error rate (e.g., 402 and 404), a binomial confidence interval 406 (e.g., 99% confidence interval) is also calculated. In some embodiments, for each individual TNC error rate binomial CI, if the upper limit of the TNC error rate binomial CI (e.g., 404) is less than the threshold 401, then the TNC error rate is selected for grouping. Therefore, selecting TNC error rates can allow abnormally high error rates (e.g., exceeding the confidence interval 406) to be eliminated, thereby reducing the accuracy error estimate (e.g., the error in determining the first value).

[0150] Next, process 210 proceeds to step 212, "Group at least some of the plurality of TNC error rates into a plurality of TNC error rate groups". In some embodiments, grouping may be performed using any known grouping method. In some embodiments, grouping may be performed using a clustering algorithm. In some embodiments, the clustering algorithm may be selected from the group consisting of: affinity propagation, agglomerative clustering, balanced iterative reduction and hierarchical clustering (BIRCH), density-based clustering with noise (DBSCAN), K-means, Mini-Batch K-Means, mean shift, point ordering identification cluster structure (OPTICS), spectral clustering, mixture of Gaussians, and partitioning around medoids (PAM) clustering. In some embodiments, grouping is performed by a process comprising nearest neighbor clustering or hierarchical agglomerative clustering.

[0151] In some implementations, grouping can be performed using Partitioning Around Medoids (PAM) clustering.

[0152] In some embodiments, the TNC error rates can be grouped into 2, 3, 4, 5, 6, 7, 8, 9, 10, or any other suitable number of TNC error rate groups. In some embodiments, a suitable number of TNC error rate groups can be determined based on the goal of obtaining sufficient read depth in each group to enhance statistical power. In some embodiments, different TNC error rate groups are associated with different numbers of TNC error rates. For example, group 1 can be associated with 40 TNC error rates and group 2 can be associated with 25 TNC error rates.

[0153] Next, process 210 proceeds to step 213, "Determine a TNC group error rate for said plurality of TNC error rate groups using TNC error rates of at least some of said plurality of TNC error rates." In some embodiments, said TNC error rates include TNC error rates selected as described above. In some embodiments, said TNC error rates include TNC error rates associated with 10, 25, 50, 75, 96, 100, 125, 150, 176, 192, or any other possible number of TNC error types. In some embodiments, the TNC error rates include TNC error rates associated with at least 10 TNC error types, at least 25 TNC error types, at least 50 TNC error types, at least 75 TNC error types, at least 96 TNC error types, at least 100 TNC error types, at least 125 TNC error types, at least 150 TNC error types, at least 176 TNC error types, or at least 192 TNC error types. In some embodiments, the TNC error rates include TNC error rates associated with 192 TNC error types. In some embodiments, the TNC error rates include TNC error rates associated with 96 TNC error types. In some embodiments the TNC error rate comprises TNC error rates associated with 10-25, 10-75, 10-125, 10-176, 10-192, 25-50, 25-75, 25-125, 25-176, 25-192, 50-75, 50-125, 50-176, 50-192 TNC error types. In some embodiments the TNC error rate comprises a TNC error rate based on the TNC error types present in the sequence reads.

[0154] In some embodiments, determining a TNC group error rate is performed based on TNC error rates associated with the group. In some embodiments, the TNC group error rate may be determined as an average of the TNC error rates associated with the TNC error rate group. In some embodiments, the TNC group error rate may be determined using a weighted average of the TNC error rates associated with the TNC error rate group. In some embodiments, the TNC group error rate may be determined using a population weighted average of the TNC error rates associated with the TNC error rate group. In some embodiments, the TNC group error rate may be determined using a mean or median of the TNC error rates associated with the TNC error rate group.

[0155] In some aspects, Figure 5 TNC group error rates 501 - 504 for a corresponding plurality of TNC error rate groups 509 are shown. Figure 5 Confidence intervals for the TNC error rates grouped into any of the groups 1-4 are shown, e.g., 505 and 506 (e.g., 99% binomial confidence intervals). In some embodiments, the TNC group error rates 501-504 can be determined by calculating the population weighted average for each group (501 (μ1), 502 (μ2), 503 (μ3), 504 (μ4)). In some embodiments, the TNC group error rates 501-504 can be determined by calculating the median for each group. In some embodiments, the TNC group error rates can be determined by calculating the average for each group 501-504.

[0156] Next, process 210 proceeds to step 214, "Determine a first value indicating the expected number of mutations present in the sequencing data using the TNC group error rates." In some embodiments, the first value can be based on at least one of the TNC group error rates. In some embodiments, the first value can be based on each of the TNC group error rates. In some embodiments, the first value can be based on the number of times the position of each mutation to be monitored is covered by sequence reads. In some embodiments, the first value can be based on the number of times the position of each mutation to be monitored in each TNC error rate group is covered by sequence reads. In some embodiments, the first value can be based on a function of the TNC group error rates and the number of times the position of the mutation to be monitored in each TNC error rate group is covered by sequence reads. In some embodiments, the first value can be based on a linear combination of the TNC group error rates and the number of times the position of each mutation to be monitored in each TNC error rate group is covered by sequence reads. In some embodiments, the first value can be determined as follows:

[0157] First value = (μ1×r1)+(μ2×r2).(μn×rn)

[0158] Where n can be the total number of TNC error rate groups, μ1 to μn can be the TNC group error rates associated with TNC error rate groups 1 to n (e.g., a population-weighted average of the TNC error rates associated with the TNC error rate groups), and r1 to rn can be the sum of the number of times each position of a mutation to be monitored in a given TNC error rate group 1 to n is covered by sequence reads (both mutated and non-mutated positions). Or in other words, r1 to rn can be the sum of the number of sequences (e.g., consensus sequences) covering the position of a mutation to be monitored in a given TNC error rate group 1 to n.

[0159] Computer Implementation

[0160] Fig.12 , which can be used with the techniques described herein (e.g., Figure 2A-2B 1303). The computer system 1300 includes one or more processors 1302 and one or more articles of manufacture, the one or more articles of manufacture including non-transitory computer-readable storage media (e.g., memory 1305 and one or more non-volatile storage media 1303). The processor 1302 can control writing data to the memory 1305 and the non-volatile storage device 1303 and reading data from the memory and non-volatile storage device in any suitable manner, as the various aspects of the technology described herein are not limited to any specific technology for writing or reading data. In order to perform any of the functions described herein, the processor 1302 can execute one or more processor-executable instructions stored in one or more non-transitory computer-readable storage media (e.g., the memory 1305), which can serve as non-transitory computer-readable storage media that store processor-executable instructions for execution by the processor 1302.

[0161] The computing device 1300 may also include a network input / output (I / O) interface 1301, through which the computing device can communicate with other computing devices (e.g., over a network), and may also include one or more user I / O interfaces 1306, through which the computing device can provide output to a user and receive input from a user. The user I / O interface may include devices such as a keyboard, a mouse, a microphone, a display device (e.g., a display or a touch screen), a speaker, a camera, and / or various other types of I / O devices.

[0162] The above-mentioned embodiments can be implemented in any of a variety of ways. For example, these embodiments can be implemented using hardware, software or a combination thereof. When implemented with software, the software code can be executed on any suitable processor (e.g., microprocessor) or processor set, whether the processor or processor set is set in a single computing device or distributed in multiple computing devices. It should be understood that any component or set of components that perform the above-mentioned functions can generally be regarded as one or more controllers that control the above-mentioned functions. The one or more controllers can be implemented in a variety of ways such as using dedicated hardware, or using general-purpose hardware (e.g., one or more processors), and the hardware is programmed by microcode or software to perform the above-mentioned functions.

[0163] At this point, it should be understood that one implementation of the embodiments described herein includes at least one computer-readable storage medium (e.g., RAM, ROM, EEPROM, flash memory or other storage technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device or other magnetic storage device, or other tangible, non-transitory computer-readable storage medium encoded with a computer program (i.e., a plurality of executable instructions)) that, when executed on one or more processors, performs the above-described functions of one or more embodiments. The computer-readable medium may be transportable so that the program stored thereon may be loaded onto any computing device to implement various aspects of the technology described herein. Furthermore, it should be understood that the computer program, when executed, may perform any of the functions described above and is not limited to an application program running on a host computer. Instead, the terms "computer program" and "software" are used herein as broad concepts to refer to any type of computer code (e.g., application software, firmware, microcode, or any other form of computer instructions) that can be used to program one or more processors to implement various aspects of the technology discussed herein.

[0164] The foregoing description of the embodiments provides illustration and description, but is not intended to be exhaustive or restrictive of the embodiments to necessarily adopt the disclosed exact form. In view of the above teachings, modifications and variations are possible, or can be obtained from the practice of this teaching. In other embodiments, the methods described in these figures may include fewer operations, different operations, operations in different orders, and / or additional operations. Further, unrelated steps may be performed in parallel.

[0165] It will be apparent that, as described above, the exemplary aspects may be implemented in many different forms of software, firmware, and hardware in the embodiments shown in the drawings. Further, certain portions of the embodiments may be implemented as a "module" that performs one or more functions. This module may include hardware, such as a processor, an application specific integrated circuit (ASIC), or a field programmable gate array (FPGA), or a combination of hardware and software.

[0166] Having thus described several aspects and embodiments of the technology set forth in this disclosure, it should be understood that those skilled in the art will readily conceive of various changes, modifications and improvements. Such changes, modifications and improvements are intended to be within the spirit and scope of the technology described herein. For example, a person of ordinary skill in the art will readily conceive of various other devices and / or structures for performing the functions described herein and / or obtaining one or more of these results and / or these advantages, and each of such changes and / or modifications is deemed to be within the scope of the embodiments described herein. Using only routine experiments, those skilled in the art will recognize or be able to determine many equivalent forms of the specific embodiments described herein. Therefore, it should be understood that the aforementioned embodiments are presented only by way of example, and within the scope of the appended claims and their equivalents, embodiments of the present invention may be practiced in a manner different from that specifically described. In addition, if two or more features, systems, manufactured products, materials, kits and / or methods described herein are not mutually contradictory, any combination of such features, systems, manufactured products, materials, kits and / or methods is included within the scope of this disclosure.

[0167] The above-mentioned embodiments can be implemented in any of a variety of ways. One or more aspects and embodiments of the execution of the process or method disclosed herein can be executed or controlled by program instructions that can be executed by a device (e.g., a computer, a processor or other device). In this regard, each inventive concept can be embodied as a computer-readable storage medium (or multiple computer-readable media) (e.g., a computer memory, one or more floppy disks, compressed disks, optical disks, tapes, flash memories, circuit configurations in field programmable gate arrays or other semiconductor devices, or other tangible computer storage media), the storage medium is encoded with one or more programs, and the one or more programs, when executed on one or more computers or other processors, will perform one or more methods in the above-mentioned various embodiments. The computer-readable medium can be transportable so that one or more programs stored thereon can be loaded onto one or more different computers or other processors to implement various aspects of the aspects described above. In some embodiments, the computer-readable medium can be a non-transitory medium.

[0168] The term "program" or "software" is used in a broad sense herein to refer to any type of computer code or computer executable instruction set that can be used to program a computer or other processor to implement the various aspects described above. In addition, it should be understood that, according to one aspect, one or more computer programs that are capable of performing the methods of the present disclosure when executed do not necessarily reside on a single computer or processor, but can be modularly distributed among multiple different computers or processors to implement various aspects of the present disclosure.

[0169] Computer executable instructions may be in a variety of forms, such as program modules that are executed by one or more computers or other devices. Typically, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. Typically, the functions of the program modules may be combined or distributed as desired in various implementations.

[0170] Moreover, the data structure may be stored in a computer readable medium in any suitable form. To simplify the description, the data structure may be shown as having fields that are associated with each other through positions in the data structure. Similarly, this association may be achieved by allocating storage locations that reflect the relationship between the fields to each field in a computer readable medium. However, any suitable mechanism may be used to establish associations between the information in the fields of the data structure, including by using pointers, tags, or other mechanisms that establish associations between data elements.

[0171] When implemented in software, the software code may be executed on any suitable processor or collection of processors, whether the processor or collection of processors is provided in a single computer or distributed among multiple computers. Further, it should be understood that the computer may be embodied in any of a variety of forms, such as a rack-mounted computer, a desktop computer, a laptop computer, or a tablet computer, as non-limiting examples. In addition, the computer may be embedded in a device that is not generally considered a computer but has suitable processing capabilities, including a personal digital assistant (PDA), a smart phone, a tablet computer, or any other suitable portable or fixed electronic device.

[0172] Moreover, the computer may have one or more input and output devices. These devices may be used to provide an interface for a user. Examples of output devices that may be used to provide a user interface include a printer or display screen for visually presenting the output, and a speaker or other sound generating device for auditorily presenting the output. Examples of input devices that may be used for a user interface include a keyboard and a pointing device, such as a mouse, a touch pad, and a digitized tablet computer. As another example, a computer may receive input information by voice recognition or in other audio formats.

[0173] Such computers may be interconnected in any suitable form via one or more networks, including local area networks or wide area networks, such as enterprise networks and intelligent networks (IN) or the Internet. Such networks may be based on any suitable technology and may operate according to any suitable protocol, and may include wireless networks, wired networks, or fiber optic networks.

[0174] In addition, as described, some aspects may be embodied as one or more methods. The steps performed as part of the method may be ordered in any suitable manner. Therefore, in a constructed embodiment, the order of execution of the steps may be different from the order shown, and the embodiment may include performing some steps simultaneously, even if these steps are shown as sequential steps in an illustrative embodiment.

[0175] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0176] As used herein in the specification and claims, the indefinite articles "a" and "an" should be understood to mean "at least one", unless explicitly stated to the contrary.

[0177] As used herein in the specification and in the claims, the phrase "and / or" should be understood to mean "either or both" of the elements so combined, i.e., elements that exist together in some cases and separately in other cases. Multiple elements listed with "and / or" should be interpreted in the same manner, i.e., "one or more elements" of the elements so combined. In addition to the elements specifically identified by the "and / or" clause, other elements may optionally be present, whether related or unrelated to those specifically identified. Thus, as a non-limiting example, when used in conjunction with open language such as "including", a reference to "A and / or B" may refer to only A (optionally including elements other than B) in one embodiment; to only B (optionally including elements other than A) in another embodiment; to both A and B (optionally including other elements) in yet another embodiment; etc.

[0178] As used in this specification and claims, the phrase "at least one" with respect to a list of one or more elements should be understood to refer to at least one element selected from any one or more elements in the list of elements, but not necessarily including at least one element of each element specifically listed in the list of elements, and not excluding any combination of elements in the list of elements. This definition also allows that there may be optionally elements other than the elements specifically identified in the list of elements to which the phrase "at least one" refers, whether related or unrelated to those specifically identified elements. Thus, as a non-limiting example, "at least one of A and B" (or equivalently, "at least one of A or B", or equivalently "at least one of A and / or B"), in one embodiment, may refer to at least one A (optionally including more than one A), the absence of B (and optionally including elements other than B); in another embodiment, may refer to at least one B (optionally including more than one B), the absence of A (and optionally including elements other than A); in yet another embodiment, may refer to at least one A (optionally including more than one A), and at least one B (optionally including more than one B) (and optionally including other elements); etc.

[0179] In the claims and the above description, all transitional phrases such as "comprising", "including", "carrying", "having", "containing", "involving", "containing", "consisting of", etc. should be understood as open-ended, i.e., meaning including but not limited to. Only the transitional phrases "consisting of" and "consisting essentially of" should be regarded as closed or semi-closed transitional phrases, respectively.

[0180] The methods illustratively disclosed herein may suitably be practiced in the absence of any element not specifically disclosed herein.Thus, for example, in each instance herein, the term "comprising" may be replaced by "consisting essentially of" or "consisting of."

[0181] The terms "approximately," "substantially," and "about" may be used to mean within ±20% of a target value in some embodiments, within ±10% of a target value in some embodiments, within ±5% of a target value in some embodiments, within ±2% of a target value in some embodiments. The terms "approximately," "substantially," and "approximately" may include the target value itself.

[0182] Example

[0183] Embodiment A1: A method for determining whether sequencing data of a biological sample of a subject provides an indication that the subject has minimal residual disease, the method comprising:

[0184] Use at least one computer hardware processor to:

[0185] (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored;

[0186] (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising:

[0187] determining a plurality of nucleotide field (NC) error rates corresponding to a plurality of NC error types using the first subset of sequence reads;

[0188] grouping at least some of the plurality of NC error rates into a plurality of NC error rate groups;

[0189] determining an NC group error rate for the plurality of NC error rate groups using NC error rates from the at least some of the plurality of NC error rates; and

[0190] using the NC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data;

[0191] (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and

[0192] (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

[0193] Embodiment A2: The method according to embodiment A1, wherein NC is selected from a single nucleotide field (SNC), a dinucleotide field (DNC), a trinucleotide field (TNC), a tetranucleotide field (4NC), a pentanucleotide field (5NC), a hexanucleotide field (6NC), a heptanucleotide field (7NC) and an octanucleotide field (8NC).

[0194] Embodiment A3: The method according to embodiment A1 or A2, wherein the sequence reads cover at least 10 positions of mutations to be monitored.

[0195] Embodiment A4: The method according to any one of embodiments A1 to A3, wherein the sequence reads cover 10-200 positions of mutations to be monitored.

[0196] Embodiment A5: The method according to any one of embodiments A1 to A4, wherein the sequence reads cover 50-200 positions of mutations to be monitored.

[0197] Embodiment A6: According to the method according to any one of embodiments A1 to A5, it further comprises: obtaining the sequencing data by sequencing the biological sample.

[0198] Embodiment A7: The method according to any one of Embodiments A1 to A6, wherein the biological sample is a body fluid or a sample obtained from a body fluid.

[0199] Embodiment A8: The method of any one of Embodiments A1 to A7, wherein the sequencing data comprises sequence reads from circulating tumor DNA (ctDNA).

[0200] Embodiment A9: The method according to any one of Embodiments A1 to A8, wherein each of the sequence reads covers at least one of the positions of the mutation to be monitored.

[0201] Embodiment A10: The method of any one of Embodiments A1 to A9, wherein the sequence reads are obtained using whole exome sequencing.

[0202] Embodiment A11: The method according to any one of embodiments A1 to A11, wherein the sequence reads are obtained using a targeted gene sequencing kit.

[0203] Embodiment A12: The method according to embodiment A11, wherein the targeted gene sequencing kit targets sequences covering the positions of mutations to be monitored.

[0204] Embodiment A13: The method according to embodiment A11, wherein primers are used to amplify the sequence including the position of the mutation to be monitored.

[0205] Embodiment A14: The method according to embodiment A11, wherein the sequence targeted by the targeted gene sequencing kit is determined using sequence data from the primary tumor of the subject.

[0206] Embodiment A15: The method of any one of Embodiments A1 to 14, wherein the first subset of sequence reads and the second subset of sequence reads are the same.

[0207] Embodiment A16: A method according to any one of Embodiments A1 to A15, wherein (B) is performed using at least the first subset of the sequence reads and one or more sequence reads in the sequencing data that do not cover the position of the mutation to be monitored.

[0208] Embodiment A17: The method of any one of Embodiments A1 to A16, wherein performing (B) further comprises:

[0209] generating a consensus sequence read using at least said first subset of said sequence reads,

[0210] wherein each of the consensus sequence reads is generated from sequence reads in at least the first subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI),

[0211] Wherein determining the plurality of NC error rates for the corresponding plurality of NC error types is performed using the generated consensus sequence reads.

[0212] Embodiment A18: The method of Embodiment A17, wherein each of the consensus sequence reads is generated from at least a threshold number of sequence reads associated with a corresponding common UMI.

[0213] Embodiment A19: The method of Embodiment A18, wherein the threshold number of sequence reads is between 2 and 20.

[0214] Embodiment A20: The method according to any one of embodiments A17 to A19, further comprising:

[0215] selecting a subset of the consensus sequence reads,

[0216] Wherein determining the plurality of NC error rates for the corresponding plurality of NC error types is performed using only a subset of the selected consensus sequence reads.

[0217] Embodiment A21: A method according to embodiment A20, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to a measure of similarity between corresponding positive strand consensus sequence reads and negative strand consensus sequence reads.

[0218] Embodiment A22: A method according to embodiment A20, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and a subset of the consensus sequence reads is selected using one or more criteria applicable to the positive strand consensus sequence reads and the negative strand consensus sequence reads.

[0219] Embodiment A23: A method according to any one of Embodiments A20 to A22, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to the relative number of positive strand consensus sequence reads and corresponding negative strand consensus sequence reads.

[0220] Embodiment A24: The method of any one of Embodiments A17 to A23, wherein using the common sequence reads to determine the plurality of NC error rates for the corresponding plurality of NC error types comprises:

[0221] determining the plurality of NC error rates using a background region of the consensus sequence reads,

[0222] The position of the mutation to be monitored includes the first position,

[0223] wherein the consensus sequence reads include a first consensus sequence read covering the first position, and the background region includes a first background region of the first consensus sequence reads,

[0224] The first background region comprises nucleotides in the first consensus sequence read that are at least a first threshold distance away from the first position.

[0225] Embodiment A25: The method according to embodiment A24, wherein the background region does not include the position of the mutation to be monitored.

[0226] Embodiment A26: The method according to any one of embodiments A17 to A23,

[0227] wherein for a first position among the positions of the mutation to be monitored, the consensus sequence reads include a first group of positive-strand consensus sequence reads associated with a positive-strand primer binding sequence and a second group of negative-strand consensus sequence reads associated with a negative-strand primer binding sequence, the positive-strand primer binding sequence being located at the 3' end of each positive-strand consensus sequence read in the positive-strand consensus sequence reads in the first group, and the negative-strand primer binding sequence being located at the 3' end of the negative-strand consensus sequence reads in the second group,

[0228] Wherein using the common sequence reads to determine the plurality of NC error rates corresponding to the plurality of NC error types comprises:

[0229] Determine the multiple NC error rates using the following:

[0230] nucleotides in any sequence read in the first group of forward strand consensus sequence reads that are within a second threshold distance of the forward strand primer binding sequence, and

[0231] A nucleotide in any sequence read in the second group of negative strand consensus sequence reads that is within a third threshold distance of the negative strand primer binding sequence.

[0232] Embodiment A27: The method of any one of Embodiments A17 to A26, wherein using the consensus sequence reads to determine the plurality of NC error rates comprises determining a frequency of occurrence of each of the TNC error types in the consensus sequence reads.

[0233] Embodiment A28: The method of Embodiment A21, wherein using the common sequence reads to determine the plurality of NC error rates for the corresponding plurality of NC error types comprises:

[0234] determining the plurality of NC error rates based on a background region of the consensus sequence reads,

[0235] wherein the consensus sequence reads comprise a first consensus sequence read, and the background region comprises a first background region of the first consensus sequence read,

[0236] The NC error rate is determined based on how frequently each of the NC error types occurs in the first background region of the first common sequence read.

[0237] Embodiment A29: The method of any one of Embodiments A1 to A28, wherein the NC error type corresponds to a mutation in any position of a given NC.

[0238] Embodiment A30: The method of any one of embodiments A1 to A29, wherein each of the NC error types corresponds to a specific mutation of an intervening nucleotide in a given NC.

[0239] Embodiment A31: The method according to any one of embodiments A1 to A30, further comprising:

[0240] After determining the multiple NC error rates and before grouping at least some of the multiple NC error rates into multiple NC error rate groups, determine confidence intervals for the NC error rates; and select at least some of the multiple NC error rates for grouping using criteria applicable to the confidence intervals for the NC error rates.

[0241] Embodiment A32: The method of any one of Embodiments A1 to A31, wherein grouping at least some of the plurality of NC error rates into a plurality of NC error rate groups comprises clustering the plurality of NC error rates.

[0242] Embodiment A33: A method according to any one of Embodiments A1 to A32, wherein grouping at least some of the plurality of NC error rates into a plurality of NC error rate groups comprises grouping using Partitioning Around Medoids (PAM) clustering.

[0243] Embodiment A34: The method of any one of Embodiments A1 to A33, wherein grouping at least some of the plurality of NC error rates comprises grouping into 4 NC error rate groups.

[0244] Embodiment A35: A method according to any one of Embodiments A1 to A34, wherein determining the first value indicating the expected number of mutations present in the sequencing data is performed using at least some of the NC group error rates and the number of times at least some of the positions of the mutations to be monitored are covered by sequence reads in the first subset of the sequence reads.

[0245] Embodiment A36: The method of any one of Embodiments A1 to A35, wherein determining the first value indicative of the expected number of mutations present in the sequencing data comprises:

[0246] The first value is determined as a weighted linear combination of the NC error group rates, wherein each specific NC error group rate among the NC error group rates is weighted according to the number of times the position to be monitored is covered by sequence reads corresponding to the NC error type belonging to the specific NC error group in the first subset of sequence reads.

[0247] Embodiment A37: The method of any one of Embodiments A1 to A36, wherein performing (C) further comprises:

[0248] generating second consensus sequence reads using at least the second subset of the sequence reads, wherein each consensus sequence read in the second consensus sequence reads is generated from sequence reads in at least the second subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI),

[0249] Wherein determining the second value indicative of the actual number of mutations present at the position of the mutation to be monitored is performed using the second consensus sequence reads.

[0250] Embodiment A38: A method according to any one of Embodiments A1 to A37, wherein (D) is performed using a statistical hypothesis test with a null hypothesis by comparing the second value with a distribution associated with the null hypothesis, wherein the distribution has one or more parameters that depend on the first value.

[0251] Embodiment A39: A method according to Embodiment A38, wherein the distribution is a Poisson distribution with a mean (λ), and the mean is set to the first value.

[0252] Embodiment A40: The method according to Embodiment A38 or Embodiment A39, wherein using the statistical hypothesis test comprises determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the null hypothesis.

[0253] Embodiment A41: The method of any one of Embodiments A1 to A40, wherein (D) is performed using a one-sided Poisson hypothesis test.

[0254] Embodiment A42: A method according to embodiment A41, wherein using the one-sided Poisson hypothesis test includes: setting the mean (λ) of the Poisson distribution to the first value, and determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the Poisson distribution.

[0255] Embodiment A43: According to the method of Embodiment A42, using the measure of likelihood to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

[0256] Embodiment A44: The method according to any one of Embodiments A39 to A43, wherein when the second value indicates that the null hypothesis can be rejected, the subject is likely to have minimal residual disease.

[0257] Embodiment A45: The method of any one of Embodiments A1 to A44, wherein (D) further comprises: providing the indication that the subject suffers from minimal residual disease.

[0258] Embodiment A46: The method of any one of Embodiments A1 to A45, further comprising using the at least one computer hardware processor to perform the following operations:

[0259] obtaining one or more additional sequencing data previously generated by sequencing one or more additional biological samples of the subject, each of the one or more additional sequencing data comprising additional sequence reads covering the position of the mutation to be monitored, and for each additional sequence read of the additional sequence reads in the one or more additional sequencing data:

[0260] Determining an additional first value indicative of an expected number of mutations due to sequencing errors present in the corresponding sequencing data using at least a first subset of the additional sequence reads, the determining comprising:

[0261] determining an additional plurality of NC error rates corresponding to the plurality of NC error types;

[0262] grouping at least some of the additional plurality of NC error rates into an additional plurality of NC error rate groups;

[0263] determining additional NC group error rates for the additional plurality of NC error rate groups using additional NC error rates of the at least some of the additional plurality of NC error rates; and

[0264] using the additional NC group error rate to determine the additional first value indicative of the expected number of mutations present in the corresponding sequencing data;

[0265] using at least a second subset of the additional sequence reads to determine an additional second value indicating the actual number of mutations present at the position; and using the additional first value indicating the expected number of mutations due to sequencing errors present in the corresponding sequencing data and the additional second value indicating the actual number of mutations present in the corresponding sequencing data at the position of the mutation to be monitored to determine whether the corresponding sequencing data provides the indication that the subject suffers from minimal residual disease.

[0266] Example 1: A method for determining whether sequencing data of a biological sample of a subject provides an indication that the subject has minimal residual disease, the method comprising:

[0267] Use at least one computer hardware processor to:

[0268] (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored;

[0269] (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising:

[0270] determining a plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types using the first subset of sequence reads;

[0271] grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups;

[0272] determining a TNC group error rate for the plurality of TNC error rate groups using TNC error rates of at least some of the plurality of TNC error rates; and

[0273] using the TNC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data;

[0274] (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and

[0275] (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

[0276] Embodiment 2: The method according to embodiment 1, wherein the sequence reads cover at least 10 positions of mutations to be monitored.

[0277] Embodiment 3: The method according to any one of embodiments 1 to 2, wherein the sequence reads cover 10-200 positions of mutations to be monitored.

[0278] Embodiment 4: The method according to any one of embodiments 1 to 3, wherein the sequence reads cover 50-200 positions of mutations to be monitored.

[0279] Embodiment 5: The method according to any one of embodiments 1 to 4 further comprises: obtaining the sequencing data by sequencing the biological sample.

[0280] Embodiment 6: The method according to any one of embodiments 1 to 5, wherein the biological sample is a body fluid or a sample obtained from a body fluid.

[0281] Embodiment 7: The method according to any one of embodiments 1 to 6, wherein the sequencing data comprises sequence reads from circulating tumor DNA (ctDNA).

[0282] Embodiment 8: The method according to any one of embodiments 1 to 7, wherein each of the sequence reads covers at least one of the positions of the mutation to be monitored.

[0283] Embodiment 9: The method according to any one of embodiments 1 to 8, wherein the sequence reads are obtained using whole exome sequencing.

[0284] Embodiment 10: The method according to any one of embodiments 1 to 9, wherein the sequence reads are obtained using a targeted gene sequencing kit.

[0285] Embodiment 11: The method according to embodiment 10, wherein the targeted gene sequencing kit targets the sequence covering the position of the mutation to be monitored.

[0286] Embodiment 12: The method according to embodiment 10, wherein primers are used to amplify the sequence including the position of the mutation to be monitored.

[0287] Embodiment 13: The method according to embodiment 10, wherein the sequence targeted by the targeted gene sequencing kit is determined using sequence data from the primary tumor of the subject.

[0288] Claims 10 to 27 relate to (B): determining the expected number of mutations due to sequencing errors.

[0289] Embodiment 14: The method of any one of Embodiments 1 to 13, wherein the first subset of sequence reads and the second subset of sequence reads are the same.

[0290] Embodiment 15: A method according to any one of Embodiments 1 to 14, wherein (B) is performed using at least the first subset of the sequence reads and one or more sequence reads in the sequencing data that do not cover the position of the mutation to be monitored.

[0291] Embodiment 16: The method according to any one of embodiments 1 to 15, wherein performing (B) further comprises:

[0292] generating a consensus sequence read using at least said first subset of said sequence reads,

[0293] wherein each of the consensus sequence reads is generated from sequence reads in at least the first subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI),

[0294] Wherein determining the plurality of trinucleotide nucleoside (TNC) error rates for the corresponding plurality of TNC error types is performed using the generated consensus sequence reads.

[0295] Embodiment 17: The method of embodiment 16, wherein each of the consensus sequence reads is generated from at least a threshold number of sequence reads associated with a corresponding common UMI.

[0296] Embodiment 18: The method of embodiment 17, wherein the threshold number of sequence reads is between 2 and 20.

[0297] Embodiment 19: The method according to embodiment 16 further comprises:

[0298] selecting a subset of the consensus sequence reads,

[0299] Wherein determining the plurality of trinucleotide field (TNC) error rates for the corresponding plurality of TNC error types is performed using only a subset of the selected consensus sequence reads.

[0300] Embodiment 20: A method according to embodiment 19, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to a measure of similarity between corresponding positive strand consensus sequence reads and negative strand consensus sequence reads.

[0301] Embodiment 21: A method according to embodiment 19, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and a subset of the consensus sequence reads is selected using one or more criteria applicable to the positive strand consensus sequence reads and the negative strand consensus sequence reads.

[0302] Embodiment 22: A method according to any one of embodiments 19 to 21, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to the relative number of positive strand consensus sequence reads and corresponding negative strand consensus sequence reads.

[0303] Embodiment 23: The method of any one of embodiments 16 to 22, wherein using the consensus sequence reads to determine the plurality of trinucleotide field (TNC) error rates for the corresponding plurality of TNC error types comprises:

[0304] determining the plurality of TNC error rates using a background region of the consensus sequence reads,

[0305] The position of the mutation to be monitored includes the first position,

[0306] wherein the consensus sequence reads include a first consensus sequence read covering the first position, and the background region includes a first background region of the first consensus sequence reads,

[0307] The first background region comprises nucleotides in the first consensus sequence read that are at least a first threshold distance away from the first position.

[0308] Embodiment 24: The method according to embodiment 23, wherein the background region does not include the position of the mutation to be monitored.

[0309] Embodiment 25: The method according to any one of embodiments 16 to 22,

[0310] wherein for a first position among the positions of the mutation to be monitored, the consensus sequence reads include a first group of positive-strand consensus sequence reads associated with a positive-strand primer binding sequence and a second group of negative-strand consensus sequence reads associated with a negative-strand primer binding sequence, the positive-strand primer binding sequence being located at the 3' end of each positive-strand consensus sequence read in the positive-strand consensus sequence reads in the first group, and the negative-strand primer binding sequence being located at the 3' end of the negative-strand consensus sequence reads in the second group,

[0311] Wherein using the consensus sequence reads to determine the plurality of trinucleotide nucleoside (TNC) error rates for the corresponding plurality of TNC error types comprises:

[0312] The multiple TNC error rates were determined using:

[0313] nucleotides in any sequence read in the first group of forward strand consensus sequence reads that are within a second threshold distance of the forward strand primer binding sequence, and

[0314] A nucleotide in any sequence read in the second group of negative strand consensus sequence reads that is within a third threshold distance of the negative strand primer binding sequence.

[0315] Embodiment 26: A method according to any one of embodiments 16 to 25, wherein using the common sequence reads to determine the plurality of trinucleotide field (TNC) error rates comprises determining the frequency of occurrence of each of the TNC error types in the common sequence reads.

[0316] Embodiment 27: The method of embodiment 20, wherein using the common sequence reads to determine the plurality of trinucleotide field (TNC) error rates for the corresponding plurality of TNC error types comprises:

[0317] determining the plurality of TNC error rates based on a background region of the consensus sequence reads,

[0318] wherein the consensus sequence reads comprise a first consensus sequence read, and the background region comprises a first background region of the first consensus sequence read,

[0319] The TNC error rate is determined based on how frequently each of the TNC error types occurs in the first background region of the first consensus sequence read.

[0320] Embodiment 28: The method according to any one of embodiments 1 to 27, wherein the TNC error type corresponds to a mutation in any position of a given TNC.

[0321] Embodiment 29: The method according to any one of embodiments 1 to 28, wherein each of the TNC error types corresponds to a specific mutation of an intervening nucleotide in a given TNC.

[0322] Embodiment 30: The method according to any one of embodiments 1 to 29, further comprising:

[0323] After determining the multiple trinucleotide field (TNC) error rates, and before grouping at least some of the multiple TNC error rates into multiple TNC error rate groups, determine confidence intervals for the TNC error rates; and select at least some of the multiple TNC error rates for grouping using criteria applicable to the confidence intervals for the TNC error rates.

[0324] Embodiment 31: The method of any one of Embodiments 1 to 30, wherein grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups comprises clustering the plurality of TNC error rates.

[0325] Embodiment 32: The method of any one of Embodiments 1 to 31, wherein grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups comprises grouping using Partitioning Around Medoid (PAM) clustering.

[0326] Embodiment 33: The method of any one of Embodiments 1 to 32, wherein grouping at least some of the plurality of TNC error rates comprises grouping into four TNC error rate groups.

[0327] Embodiment 34: A method according to any one of Embodiments 1 to 33, wherein determining the first value indicating the expected number of mutations present in the sequencing data is performed using at least some of the TNC group error rates and the number of times at least some of the positions of the mutations to be monitored are covered by sequence reads in the first subset of the sequence reads.

[0328] Embodiment 35: The method according to any one of embodiments 1 to 34, wherein determining the first value indicating the expected number of mutations present in the sequencing data comprises:

[0329] The first value is determined as a weighted linear combination of the TNC error group rates, wherein each specific TNC error group rate among the TNC error group rates is weighted according to the number of times the position to be monitored is covered by sequence reads corresponding to a TNC error type belonging to the specific TNC error group in the first subset of the sequence reads.

[0330] Embodiment 36: The method according to any one of embodiments 1 to 35, wherein performing (C) further comprises:

[0331] generating second consensus sequence reads using at least the second subset of the sequence reads, wherein each consensus sequence read in the second consensus sequence reads is generated from sequence reads in at least the second subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI),

[0332] Wherein determining a second value indicative of an actual number of mutations present at the position of the mutation to be monitored is performed using the second consensus sequence reads.

[0333] Embodiment 37: A method according to any one of Embodiments 1 to 36, wherein (D) is performed using a statistical hypothesis test with a null hypothesis by comparing the second value to a distribution associated with the null hypothesis, wherein the distribution has one or more parameters that depend on the first value.

[0334] Embodiment 38: The method of Embodiment 37, wherein the distribution is a Poisson distribution with a mean (λ), the mean being set to the first value.

[0335] Embodiment 39: The method according to embodiment 37 or embodiment 38, wherein using the statistical hypothesis test comprises determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the null hypothesis.

[0336] Embodiment 40: The method according to any one of embodiments 1 to 39, wherein (D) is performed using a one-sided Poisson hypothesis test.

[0337] Example 41: A method according to Example 40, wherein using the one-sided Poisson hypothesis test includes: setting the mean (λ) of the Poisson distribution to the first value, and determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the Poisson distribution.

[0338] Example 42: According to the method of Example 41, using the measure of likelihood to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

[0339] Embodiment 43: The method according to any one of Embodiments 38 to 42, wherein when the second value indicates that the null hypothesis can be rejected, the subject is likely to have minimal residual disease.

[0340] Embodiment 44: The method according to any one of Embodiments 1 to 43, wherein (D) further comprises: providing the indication that the subject has minimal residual disease.

[0341] Embodiment 45: The method of any one of Embodiments 1 to 44, further comprising using the at least one computer hardware processor to perform the following operations:

[0342] obtaining one or more additional sequencing data previously generated by sequencing one or more additional biological samples of the subject, each of the one or more additional sequencing data comprising additional sequence reads covering the position of the mutation to be monitored, and for each additional sequence read of the additional sequence reads in the one or more additional sequencing data:

[0343] Determining an additional first value indicative of an expected number of mutations due to sequencing errors present in the corresponding sequencing data using at least a first subset of the additional sequence reads, the determining comprising:

[0344] determining an additional plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types; grouping at least some of the additional plurality of TNC error rates into an additional plurality of TNC error rate groups;

[0345] determining an additional TNC group error rate for the additional plurality of TNC error rate groups using additional TNC error rates of at least some of the additional plurality of TNC error rates; and

[0346] using the additional TNC group error rate to determine the additional first value indicative of the expected number of mutations present in the corresponding sequencing data;

[0347] using at least a second subset of the additional sequence reads to determine an additional second value indicating the actual number of mutations present at the position; and using the additional first value indicating the expected number of mutations due to sequencing errors present in the corresponding sequencing data and the additional second value indicating the actual number of mutations present in the corresponding sequencing data at the position of the mutation to be monitored to determine whether the corresponding sequencing data provides the indication that the subject suffers from minimal residual disease.

[0348] Embodiment 46: A system for determining whether sequencing data of a biological sample of a subject provides an indication that the subject has minimal residual disease, the system comprising:

[0349] at least one computer hardware processor; and

[0350] at least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by the at least one computer hardware processor, cause the at least one computer hardware processor to:

[0351] (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored;

[0352] (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising:

[0353] determining a plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types using a first subset of the sequence reads;

[0354] grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups;

[0355] determining a TNC group error rate for the plurality of TNC error rate groups using TNC error rates of at least some of the plurality of TNC error rates; and

[0356] using the TNC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data;

[0357] (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and

[0358] (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

[0359] Embodiment 47: The system according to Embodiment 46, wherein the at least one computer hardware processor stores processor-executable instructions, and the processor-executable instructions cause the at least one computer hardware processor to perform the method according to any one of claims 2 to 45.

[0360] Embodiment 48: At least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to:

[0361] (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored;

[0362] (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising:

[0363] determining a plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types using the first subset of sequence reads;

[0364] grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups; determining TNC group error rates for the plurality of TNC error rate groups using TNC error rates of at least some of the plurality of TNC error rates; and

[0365] using the TNC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data;

[0366] (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and

[0367] (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

[0368] Embodiment 49: According to at least one non-temporary computer-readable storage medium storing processor-executable instructions as described in Embodiment 48, wherein the at least one computer hardware processor stores processor-executable instructions, and the processor-executable instructions cause the at least one computer hardware processor to perform a method according to any one of Embodiments 2 to 45.

[0369] Example 50: Determining the Indication of MRD in Non-Small Cell Lung Cancer (NSCLC)

[0370] summary

[0371] Data are presented using a method of tumor-informed ctDNA detection from specific patients (e.g., indications of MRD) that detects a median of 200 somatic variants per patient from surgically resected tumors. To evaluate test performance, 712 construction samples were analyzed and studied together with 1093 clinically annotated plasma samples from 197 prospectively recruited patients with early-stage operable NSCLC, 76 of whom experienced disease recurrence. Analysis of 383 extracranial surveillance scans revealed that MRD status helps explain uncertain results on surveillance imaging and has the potential to guide early, definitive interventions at these sites.

[0372] introduction

[0373] Circulating tumor DNA (ctDNA) is a multifaceted biomarker with the potential to accelerate innovation, especially in the field of early interventional trials. ctDNA detected after surgery, if indicative of molecular residual disease (MRD), is a specific indicator of impending recurrence of NSCLC. Here, this application reports a personalized, tumor-informed anchored multiplex PCR (AMP) approach to determine whether there is an indication of MRD in a study of NSCLC for the determination of progression (NCT01888601).

[0374] MRD detection

[0375] Compared to previous work, this application increases the number of mutations tracked by patient-specific enrichment panels (e.g., PSP or targeted gene sequencing panels) and adopts methods that enable higher-resolution characterization of NSCLC recurrence ( Figure 7 PSP targets a median of 200 mutations (range 72 to 201). The algorithm designed by PSP to select variants from tumor exome data incorporates parameters including predicted low background sequencing errors and elevated tumor copy number status to maximize sensitivity ( Figure 8 and Fig. 9 ). In cfDNA analysis, only deep reads were considered to minimize sequencing errors (see Methods). Deep reads are reads supported by a minimum number (e.g., at least 5) of sequence reads associated with a single unique identifier [UMI].

[0376] The MRD detection algorithm evaluates background (non-variant) sequencing positions to estimate the error rate of trinucleotide fields within the library, thereby enabling ctDNA detection based on a single library (Methods, see Figure 7 ). Below in the "Analysis and Verification Experiment" (including reference Fig.11E) will detail the determination of the MRD judgment threshold in the pilot trial data of 10 patients, the analytical validation of the sensitivity of variant DNA detection in 712 simulated samples analyzed by PSP of 50 variants, and the orthogonal validation of the positive judgment results of preoperative ctDNA in NSCLC using digital droplet PCR.

[0377] Postoperative ctDNA testing is specific for recurrent NSCLC

[0378] Postoperative cfDNA samples from 45 recurrence-free patients were analyzed ( Fig. 10A ). All samples have p-values ​​exceeding 0.1 ( Fig.11A Postoperative cfDNA samples from 20 patients who developed new second primary tumors during follow-up were also analyzed (PSP is specific for resected primary NSCLC and is not expected to detect second primary tumors, Fig. 10B ). 462 of 472 postoperative samples (97.9%) from these patients tested negative for ctDNA, including 62 of 65 patients (95%) who had no detectable ctDNA after surgery.

[0379] about Figures 10A-10C , the circle on the left of day 0 is the preoperative time point when the patient's tumor was still in situ. The circle on the right of day 0 is the sample time point collected after surgical resection of the primary NSCLC. If the color of the circle is dark, it reflects a positive ctDNA test. The light gray rectangles (rectangles 1 and 3) indicate whether the patient received chemotherapy, the dark gray rectangles (rectangles 2 and 4) indicate whether the patient received radiotherapy, and the medium gray shaded rectangle (rectangle 5) indicates whether the patient received surgery after recurrence. The triangles represent CT, PET, or MRI imaging results under standard postoperative care, and the images are classified as no disease (medium gray, triangle 8), uncertain images (very light gray, triangle 9), or clear imaging evidence of extracranial recurrence (dark gray, triangle 10). The medium gray triangle (triangle 6) indicates no evidence of intracranial recurrence, and the very dark gray triangle (triangle 7) indicates intracranial recurrence. The black vertical line represents the patient's event date (if an event such as death, second primary tumor, NSCLC recurrence, etc. occurred), otherwise the vertical line represents the patient's follow-up cutoff date for the NSCLC study (NCT01888601).

[0380] ctDNA detection in patients with relapsed NSCLC

[0381] Four hundred post-operative plasma samples from 76 patients with recurrence of their NSCLC were analyzed ( Fig. 10CAmong their postoperative hospitalized patients with recurrent NSCLC, 1 of 13 reads had an MRD p value between 0.1 and 0.01 ( Fig. 11B ) were made within this range. The remaining calls were made with a p value less than 0.01. Among patients with detectable ctDNA before surgery, postoperative ctDNA was detected in 54 of 61 patients (89%). In contrast, among patients who lacked detectable ctDNA before surgery, ctDNA was detected postoperatively in 10 of 14 cases (71%, Fig. 10C ) ctDNA was detected. Lead time (number of days from first postoperative ctDNA detection to confirmed radiographic recurrence) was assessed. Patients in whom ctDNA was not detected postoperatively or ctDNA was detected after clinical recurrence were assigned a lead time of 0 days. Lead time was evaluable in 67 of 75 patients (patients without preoperative plasma sampling for clinical recurrence [CRUK0048, 0557, 0516, 0674, 0640] or patients whose postoperative imaging showed incomplete resection of the disease [CRUK0230, 0234, 0291, and 0387] were excluded). Among patients with detectable ctDNA before surgery, the median lead time for all patients was 119 days, and the mean lead time was 236 days (range 0 to 1137 days, n=54, Fig. 10C In patients without detectable ctDNA before surgery, the median lead time was 0 days and the mean lead time was 114 days (range 0 to 589 days, n=14, Fig. 10C ). In the initial postoperative ctDNA sampling, 26 of 52 patients were ctDNA positive, and 26 patients became ctDNA positive during the monitoring period after a median of 2 (range 1-9) ctDNA-negative time points ( Fig. 10C ).

[0382] For reference, regarding preoperative ctDNA determination from the pilot cohort, 7 patients had positive ctDNA in plasma before surgery, with p values ​​less than 0.01 (<0.01) for all determinations ( Fig. 11C To evaluate the specificity of the ctDNA determination tool, an in silico analysis was performed to generate 3157 simulated MRD panels in the pilot patient library and to evaluate the p-value of the ctDNA determination tool ( Fig.11D At a threshold of p value < 0.1, 121 of the 157 simulated mimetic panels were ctDNA positive (96.2% in silico specificity), and at a threshold of p value less than 0.01 (<0.01), 22 of the 3157 simulated mimetic panels were ctDNA positive (99.3% in silico specificity).

[0383] Imaging and MRD testing

[0384] A total of 386 reports of extracranial surveillance imaging from 131 patients who underwent postoperative plasma sampling were reviewed (including 343 CT scans of the neck, chest, abdomen, pelvis, colon, kidney, bladder, and spine, 7 magnetic resonance imaging [MRI], spine, liver, or femur scans, and 36 whole-body positron emission tomography scans). In 15 of 11 patients, MRD was detected before imaging scans showed no new abnormalities; 9 of these patients subsequently had a recurrence of NSCLC ( Fig. 10D ). It is common for ctDNA to be detected before imaging scans show new, undefined abnormalities. These data suggest that MRD status could guide early, definitive therapeutic intervention (eg, surgery, radiation, or ablative therapy) when the anatomical site is unclear.

[0385] discuss

[0386] Detection of ctDNA in the post-curative setting indicates impending disease recurrence. A total of 1096 preoperative and postoperative plasma samples from 197 patients were analyzed. Superior performance of MRD monitoring was observed in patients with preoperative ctDNA detection (median lead time of 119 days vs. 0 days in preoperative ctDNA-positive patients vs. preoperative ctDNA-negative patients), suggesting that MRD monitoring can be prioritized over conventional radiation monitoring in patients who present with preoperative ctDNA detection. Of the 12 patients in whom MRD was detected during the preoperative neoadjuvant therapy phase, only one patient remained recurrence-free after receiving adjuvant radiation therapy (patient CRUK0086 received postoperative radiation therapy to the mediastinal bed and died of intestinal perforation 880 days after surgery without cancer recurrence). The remaining patients all experienced NSCLC recurrence within 2.8 years after ctDNA detection. This suggests that in NSCLC, the 8% of patients who benefit from adjuvant chemotherapy at a 5-year survival benefit may not be identified by the detection limit of current tumor-informed assays for MRD, perhaps because the metastatic tumor burden required to detect a positive MRD result in the preoperative adjuvant setting is already beyond the range that can be cured with systemic therapy.

[0387] method

[0388] Library preparation using anchored multiplex PCR (AMP) Anchored multiplex PCR (AMP) is a nested multiplex PCR enrichment chemistry that combines strand-specific primers and the introduction of unique molecular identifiers (UMIs) to incorporate UMIs into sequencing reads9. Cell-free DNA, fragmented peripheral blood mononuclear cell (PBMC) DNA, or fragmented normal tissue DNA is end-repaired, phosphorylated, and A-tailed. Adapters containing universal primer sites, multiple indexes, and UMIs are then ligated to the DNA. One round of target-specific PCR amplification is performed with gene-specific primer 1 (GSP1), which amplifies against the P5 primer in the adapter, and an additional round of PCR is performed with a second nested gene-specific primer (GSP2) and a primer that incorporates a second primer containing the P7 index. Strand-specific primer amplification is performed in both rounds of amplification, which helps to identify positive-strand input DNA molecules and negative-strand input DNA molecules during informatics analysis.

[0389] The method aims to sequence each library to 10 million reads. The libraries are then evaluated for the target deduplication ratio, which describes the ratio of raw target reads to unique molecular identifier [UMI]-supported target reads (UMI-supported reads contain 5 or more supporting raw reads with matching molecular indices). In samples where the initial sequencing depth resulted in a target deduplication ratio of less than 10:1, additional sequencing was performed to maximize the recovery of unique molecular identifier (UMI) families. 6000 System or system PBMC and normal tissue libraries were sequenced on the same platform.

[0390] MRD determination algorithm | An MRD determination tool was developed that studies the MRD of background sequencing noise based on the internal conditions of the library ( Figure 7 ). MRD determination tool use The information pipeline is used to clean the input reads and generate deduplicated UMI-supported reads. The cleaned, deduplicated, and error-corrected UMI-supported reads are aligned to hg19 and used to evaluate alternative base observations at predetermined positions (based on tumor information) of specific variants present in patient tumors. Only "deep" common reads supported by 5 or more PCR replicates are used to infer the expected sequencing noise and calculate the signal of the MRD call algorithm.

[0391] For alternative bases at the position of tumor information, a set of strict high-quality filters will be used to remove artifact signals, which consists of off-target filters, read strand bias filters, sequencing strand bias filters, background error rate filters, and variant allele frequency outlier filters. The variant allele frequency outlier filter performs clustering analysis on the variant allele frequencies (VAFs) of the positions of tumor information that pass the above filters by performing PAM (partitioning around the center point) clustering, thereby generating high VAF groups and low VAF groups. In the clustering algorithm, K is set to 2. If the VAF of one of the two clusters is significantly higher, as indicated by the non-overlapping confidence intervals between the highest VAF of the low VAF cluster and the lowest VAF of the higher VAF cluster, and the cluster contains 2 or fewer tumor-specific variants, these variants will be removed from the subsequent algorithm analysis.

[0392] Then, the background error rates (ERs) within the library are calculated. ER is used to establish the level of noise present in each library, which must be clearly exceeded to make an MRD call. To calculate the background library ER, the number of UMI-supported alternative base observations (deep alternate observations, DAOs) at each possible alternative base position of the positive strand based on the reference sequence across the detection region of interest (ROI) is counted. The ER corresponding to each TNC alternative base is calculated as DAO / DDP (deep UMI-corrected depth across a TNC alternate, DDP). In order to measure only PCR and sequencing errors, if the VAF of a particular alternative base at the position in the ROI is > 1%, the position is not included in the calculation of the TNC ER (based on the fact that this may represent a clonal hematopoiesis-related mutation or single nucleotide polymorphism).

[0393] A plot of observed tumor variants and their associated TNC ER was generated. Observed tumor variants where any corresponding TNC ER upper confidence interval was above 0.01% were filtered from the MRD call algorithm. TNC error rates for four "D groups" were generated using PAM clustering, which were selected from the qualified TNC error rates. The population-weighted average TNC error rate for each of the four D groups was calculated based on the product of the TNC error rates included in each D group cluster and the total DDP for each TNC. The generation of four D groups ensured that there was sufficient DDP coverage within the library of each D group to accurately estimate the ER of the variants within each group.

[0394] To determine whether ctDNA is present in a sample, the total number of DAOs observed across tumor-specific locations remaining after filter processing was compared to the number of DAOs expected based on the background ER as determined by the D cohort. A one-tailed exact Poisson test was applied, in which the total number of DAOs observed remaining was taken as the value to be detected, and the number of DAOs expected due to error served as the parameter λ of the Poisson distribution. If the resulting P value of the test was below the alpha threshold pre-set to 0.01, the sample was classified as MRD positive. Figure 11 provides data on how the pre-specified alpha threshold of 0.01 used in these analyses was generated.

[0395] To investigate the presence of a single mutation targeted by the panel, a specific trinucleotide error rate corresponding to the mutation of interest was utilized, and a one-tailed Poisson test was performed to assess whether the number of DAOs across the mutation of interest was above the expected background ER. A variant was considered reliably detected if the number of DAOs was above the expected background error using an alpha threshold of 0.01.

[0396] Designing AMP-MRD Enrichment Panels | Tumor-informed personalized AMP-MRD enrichment panels were designed for 197 patients enrolled in a NSCLC study (NCT01888601). Using the ArcherDx panel design algorithm, a median of 50 variants were selected in each panel (range 0 to 50), and a median of 150 variants were selected (range 34 to 153) using variants selected from multi-region exome sequencing data from a NSCLC study (NCT01888601). When selecting Archer variants, WES sequencing data from the highest purity tumor regions and paired germline DNA data were used. The algorithm then identifies high-confidence variants that are present and are not artifacts but tumor-specific variants. The algorithm then uses the ArcherDX AMP set to determine which variants can be targeted for testing, and from this set of variants, the 50 most informative mutations are selected as targeted targets based on the following criteria: the quality of the primers targeting the variant (to ensure high sequencing coverage of the target variant), the predicted error rate of the variant in the error-corrected bin, and the localizability. The predicted error rate for each variant is determined based on analysis of AMP cfDNA libraries sequenced on the NovaSeq instrument. This error rate analysis is performed by performing targeted variant calls on each possible SNV in the set of Archer LiquidPlex cfDNA libraries. The variants used for the NSCLC study were screened out after sequencing the NSCLC primary tumor WES pipeline 7.

[0397] Each personalized enrichment set also contains 90 primers that target 45 common single nucleotide polymorphisms (SNPs). During analysis, the zygosity of these SNPs in the cfDNA library was compared with their zygosity in the whole exome sequencing data of the patient to confirm that sample mismatch did not occur. In addition, the coverage provided by these primers helps to determine the background PCR and sequencing error rate of the library. These 45 SNPs are selected based on their presence at a frequency of 25%-75% in each Gnomad subgroup to maximize the utility of detecting sample swapping.

[0398] Analytical Validation Experiments | For Experiment LODI, a sample of 634 fragmented DNA (bottle genomic DNA, NA24385) with a known SNP profile was added to a background of four other fragmented bottle genomic inputs (NA24149, NA24631, NA24694, and NA24695). Six AMP enrichment panels were generated that target 50 heterozygous SNPs located in NA24385 and not present in the other four cell lines. To generate mock samples, NA24385 DNA was spiked into the background of the other four samples at a ratio of 0.006% to 0.2% by mass to target variant allele frequencies ranging from 0.003% to 0.1% (since the frequency of heterozygous variants in pure NA24538 is 50%). Mixtures with target allele frequencies of 1%, 5%, and 10% were also prepared as part of the same dilution series. These mixtures were used as input samples for AMP library preparation to confirm that quality-based mixing achieved the desired target allele fraction. The spike variant fractions in these higher AF libraries were measured by adding the number of deep alternative base reads across the targeted SNPs and dividing by the total coverage of the deep reads across the targeted SNPs. This analysis confirmed that the spike achieved the target AF. 2ng to 80ng of fragmented DNA were used as input in the experiment to reflect the range of DNA inputs encountered in the clinical setting. In general, 564 of the 634 samples were considered to be useful for LOD1 analysis (62 samples failed due to incorrect DNA input used, with the judgment criteria being <30 or >400 target reads for each 3 primer per ng input DNA, and 8 samples failed due to less than 10 million reads). Clinical samples were used in the validation of AMP-MRD (LOD2), and the samples were prepared using a method similar to that for the preparation of the genomic mixture in the bottle. Whole exome sequencing data from four patients were used to design patient-specific panels containing 50 SNVs with the ArcherDx panel design algorithm. The panels were used to prepare libraries using cfDNA from each patient, and the overall tumor variant AF for each sample was calculated by summing the total number of deep unique reads containing the targeted tumor-specific variants and dividing by the sum of deep unique coverage across all targeted tumor variants. The total AF of the cfDNA libraries of the four patients was greater than 1% (>1%). A single mixture was prepared using cfDNA from healthy donors and used to dilute the patient cfDNA. These dilutions were performed in serial dilutions. The first dilution was made with a target of 1% total AF, and this mixture was used to prepare the library. The total AF of this sample was measured, and a dilution correction factor was calculated to account for differences in conversion efficiency between background cfDNA.For example, if the target AF was 1% and the actual observed AF was 1.3%, this indicates that the patient's cfDNA was converted into libraries more efficiently than the background, so more background DNA was used. Mixing was then performed to prepare mixtures with AFs of 0.1%, 0.05%, 0.01%, 0.008%, and 0.005%. A total of 100 libraries were prepared under the conditions of 5 AFs and 3 input qualities. 48 blank samples (DNA donated by 22 healthy donors) were analyzed to assess the specificity of the assay. The method for calculating the observed allele frequency of the panel is to count the number of deep alternative base reads recorded across the AMP panel, subtract the estimated background error, and divide by the coverage of the depth across the panel. For assay sensitivity for specific spike categories, Clopper-Pearson binomial two-sided 95% confidence intervals were calculated in Appendix 2e-f using DescTools software in R and the BinomCI function.

[0399] Simulated analysis to assess specificity | Assess trinucleotide fields of tumor-specific SNVs within each set of the AMP-MRD pilot cohort for the NSCLC study (NCT01888601). Based on these data, simulated tumor signatures (i.e., genomic positions covered by enriched primers with expected error rates similar to the target SNVs) were generated. A simulated variant was added to a simulated signature if the following criteria were met: the variant was bidirectionally covered by primers used for MRD detection; the variant contained a TNC cohort error rate that was identical to the TNC cohort error rate of the true MRD variant it replaced; the variant was not a known population SNP variant as defined by Ensemble's Variant Effect Predictor version 94.5; the variant had an error-corrected coverage difference (δ) of no more than 2,000 compared to the true MRD variant; and the variant was not used in any other simulated tumor signature, including itself. Therefore, the generated simulated signatures targeted bases that were not mutated in the primary tumor, and any positive MRD calls detected from these simulated signatures were by default false positives. MRD-positive calls were assessed for 3157 simulated signatures across 91 pilot cfDNA libraries. The simulated ctDNA fraction for each sample was estimated by measuring the number of deep alternative base reads recorded in the simulated signature, minus the estimated background error, and dividing by the coverage across the depth of the simulated signature.

[0400] Orthogonal Validation of Droplet Digital PCR | Orthogonal analysis of droplet digital polymerase chain reaction (ddPCR) was performed on 30 preoperative plasma samples from patients in the NSCLC study (NCT01888601) whose preoperative plasma was also analyzed using the tumor-informed AMP personalized approach. In addition, 8 negative control samples (preoperative plasma from patients diagnosed with non-malignant diseases after surgery) were analyzed. Patients in the NSCLC study (NCT01888601) were selected because they had clonal driver mutations that could be targeted by a single ddPCR. The ddPCR assay used was Determination (SAGA diagnostics), and has been designed and developed on the QX200 droplet digital PCR system of BioRad. ddPCR analysis is performed at SAGA, which receives plasma (median value is 4.8ml, ranging from 2.5ml to 5.2ml). cfDNA is extracted using QiaAMPMinElute ccfDNA Midi kit (Qiagen). cfDNA is eluted in 40ul buffer EB. In each case, all cfDNA materials are used as input, and ddPCR analysis is performed in 4 replicate reaction wells for each sample.

Claims

1. A method for determining whether sequencing data of a biological sample of a subject provides an indication that the subject has minimal residual disease, the method comprising: Use at least one computer hardware processor to: (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored; (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising: determining a plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types using the first subset of sequence reads; grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups; determining a TNC group error rate for the plurality of TNC error rate groups using TNC error rates of at least some of the plurality of TNC error rates; and using the TNC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data; (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

2. The method according to claim 1, wherein the sequence reads cover at least 10 positions of mutations to be monitored.

3. The method according to claim 1 or claim 2, wherein the sequence reads cover 10-200 positions of mutations to be monitored.

4. The method according to any one of claims 1 to 3, wherein the sequence reads cover 50-200 positions of mutations to be monitored.

5. The method according to any one of claims 1 to 4, further comprising: The sequencing data is obtained by sequencing the biological sample.

6. The method according to any one of claims 1 to 5, wherein the biological sample is a body fluid or a sample obtained from a body fluid.

7. The method according to any one of claims 1 to 6, wherein the sequencing data comprises sequence reads from circulating tumor DNA (ctDNA).

8. The method according to any one of claims 1 to 7, wherein each of the sequence reads covers at least one of the positions of the mutation to be monitored.

9. The method according to any one of claims 1 to 8, wherein the sequence reads are obtained using whole exome sequencing.

10. The method according to any one of claims 1 to 9, wherein the sequence reads are obtained using a targeted gene sequencing kit.

11. The method according to claim 10, wherein the targeted gene sequencing kit targets sequences covering the positions of mutations to be monitored.

12. The method of claim 10, wherein primers are used to amplify a sequence including the position of the mutation to be monitored.

13. The method of claim 10, wherein the sequence targeted by the targeted gene sequencing kit is determined using sequence data from a primary tumor of the subject.

14. The method of any one of claims 1 to 13, wherein the first subset of sequence reads and the second subset of sequence reads are identical.

15. The method according to any one of claims 1 to 14, wherein (B) is performed using at least the first subset of the sequence reads and one or more sequence reads in the sequencing data that do not cover the position of the mutation to be monitored.

16. The method according to any one of claims 1 to 15, wherein performing (B) further comprises: generating a consensus sequence read using at least said first subset of said sequence reads, wherein each of the consensus sequence reads is generated from sequence reads in at least the first subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI), Wherein determining the plurality of trinucleotide nucleoside (TNC) error rates for the corresponding plurality of TNC error types is performed using the generated consensus sequence reads.

17. The method of claim 16, wherein each of the consensus sequence reads is generated from at least a threshold number of sequence reads associated with a corresponding common UMI. The method of claim 17 , wherein the threshold number of sequence reads is between 2 and 20.

19. The method of claim 16, further comprising: selecting a subset of the consensus sequence reads, Wherein determining the plurality of trinucleotide field (TNC) error rates for the corresponding plurality of TNC error types is performed using only a subset of the selected consensus sequence reads.

20. The method of claim 19, wherein the consensus sequence reads comprise positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to a measure of similarity between corresponding positive strand consensus sequence reads and negative strand consensus sequence reads.

21. The method of claim 19, wherein the consensus sequence reads comprise positive strand consensus sequence reads and negative strand consensus sequence reads, and a subset of the consensus reads is selected using one or more criteria applicable to the positive strand consensus sequence reads and the negative strand consensus sequence reads.

22. A method according to any one of claims 19 to 21, wherein the consensus sequence reads include positive strand consensus sequence reads and negative strand consensus sequence reads, and wherein selecting the subset is performed using a criterion applicable to the relative number of positive strand consensus sequence reads and corresponding negative strand consensus sequence reads.

23. The method of any one of claims 16 to 22, wherein using the consensus sequence reads to determine the plurality of trinucleotide field (TNC) error rates for the corresponding plurality of TNC error types comprises: determining the plurality of TNC error rates using a background region of the consensus sequence reads, The position of the mutation to be monitored includes the first position, wherein the consensus sequence reads include a first consensus sequence read covering the first position, and the background region includes a first background region of the first consensus sequence reads, The first background region comprises nucleotides in the first consensus sequence read that are at least a first threshold distance away from the first position. The method according to claim 23 , wherein the background region does not include the position of the mutation to be monitored.

25. The method according to any one of claims 16 to 22, wherein for a first position among the positions of the mutation to be monitored, the consensus sequence reads include a first group of positive-strand consensus sequence reads associated with a positive-strand primer binding sequence and a second group of negative-strand consensus sequence reads associated with a negative-strand primer binding sequence, the positive-strand primer binding sequence being located at the 3' end of each positive-strand consensus sequence read in the positive-strand consensus sequence reads in the first group, and the negative-strand primer binding sequence being located at the 3' end of the negative-strand consensus sequence reads in the second group, Wherein using the consensus sequence reads to determine the plurality of trinucleotide nucleoside (TNC) error rates for the corresponding plurality of TNC error types comprises: The multiple TNC error rates were determined using: nucleotides in any sequence read in the first group of forward strand consensus sequence reads that are within a second threshold distance of the forward strand primer binding sequence, and A nucleotide in any sequence read in the second group of negative strand consensus sequence reads that is within a third threshold distance of the negative strand primer binding sequence.

26. The method of any one of claims 16 to 25, wherein using the consensus sequence reads to determine the plurality of trinucleotide field (TNC) error rates comprises determining a frequency of occurrence of each of the TNC error types in the consensus sequence reads.

27. The method of claim 20, wherein using the consensus sequence reads to determine the plurality of trinucleotide field (TNC) error rates for the corresponding plurality of TNC error types comprises: determining the plurality of TNC error rates based on a background region of the consensus sequence reads, wherein the consensus sequence reads comprise a first consensus sequence read, and the background region comprises a first background region of the first consensus sequence read, The TNC error rate is determined based on how frequently each of the TNC error types occurs in the first background region of the first consensus sequence read.

28. The method according to any one of claims 1 to 27, wherein the TNC error type corresponds to a mutation in any position of a given TNC.

29. The method of any one of claims 1 to 28, wherein each of the TNC error types corresponds to a specific mutation of an intervening nucleotide in a given TNC.

30. The method according to any one of claims 1 to 29, further comprising: After determining the plurality of trinucleotide field (TNC) error rates and before grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups, determining confidence intervals for the TNC error rates; And selecting said at least some of said plurality of TNC error rates for grouping using criteria applicable to said confidence intervals of said TNC error rates.

31. The method of any one of claims 1 to 30, wherein grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups comprises clustering the plurality of TNC error rates.

32. The method of any one of claims 1 to 31, wherein grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups comprises grouping using Partitioning Around Medoids (PAM) clustering.

33. The method of any one of claims 1 to 32, wherein grouping at least some of the plurality of TNC error rates comprises grouping into four TNC error rate groups.

34. A method according to any one of claims 1 to 33, wherein determining the first value indicative of the expected number of mutations present in the sequencing data is performed using at least some of the TNC group error rates and the number of times at least some of the positions of the mutations to be monitored are covered by sequence reads in the first subset of the sequence reads.

35. The method of any one of claims 1 to 34, wherein determining the first value indicative of the expected number of mutations present in the sequencing data comprises: The first value is determined as a weighted linear combination of the TNC error group rates, wherein each specific TNC error group rate among the TNC error group rates is weighted according to the number of times the position to be monitored is covered by sequence reads corresponding to a TNC error type belonging to the specific TNC error group in the first subset of the sequence reads.

36. The method of any one of claims 1 to 35, wherein performing (C) further comprises: generating second consensus sequence reads using at least the second subset of the sequence reads, wherein each consensus sequence read in the second consensus sequence reads is generated from sequence reads in at least the second subset of the sequence reads that are associated with a corresponding common unique molecular identifier (UMI), Wherein determining the second value indicative of the actual number of mutations present at the position of the mutation to be monitored is performed using the second consensus sequence reads.

37. A method according to any one of claims 1 to 36, wherein (D) is performed using a statistical hypothesis test with a null hypothesis by comparing the second value to a distribution associated with the null hypothesis, wherein the distribution has one or more parameters that depend on the first value.

38. The method of claim 37, wherein the distribution is a Poisson distribution with a mean (λ), the mean being set to the first value.

39. The method of claim 37 or claim 38, wherein using the statistical hypothesis test comprises determining a measure of the likelihood of observing the actual number of mutations represented by the second value under the null hypothesis.

40. The method of any one of claims 1 to 39, wherein (D) is performed using a one-sided Poisson hypothesis test.

41. The method of claim 40, wherein using the one-sided Poisson hypothesis test comprises: The mean (λ) of the Poisson distribution is set to the first value, and a measure of the likelihood of observing the actual number of mutations represented by the second value under the Poisson distribution is determined.

42. The method of claim 41, using the measure of likelihood to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

43. The method of any one of claims 38 to 42, wherein if the second value indicates that the null hypothesis can be rejected, the subject is likely to have minimal residual disease.

44. The method of any one of claims 1 to 43, wherein (D) further comprises: The indication that the subject has minimal residual disease is provided.

45. The method of any one of claims 1 to 44, further comprising using the at least one computer hardware processor to: obtaining one or more additional sequencing data previously generated by sequencing one or more additional biological samples of the subject, each of the one or more additional sequencing data comprising additional sequence reads covering the position of the mutation to be monitored, and for each additional sequence read of the additional sequence reads in the one or more additional sequencing data: Determining an additional first value indicative of an expected number of mutations due to sequencing errors present in the corresponding sequencing data using at least a first subset of the additional sequence reads, the determining comprising: determining additional multiple trinucleotide nucleoside (TNC) error rates corresponding to multiple TNC error types; grouping at least some of the additional plurality of TNC error rates into an additional plurality of TNC error rate groups; determining an additional TNC group error rate for the additional plurality of TNC error rate groups using additional TNC error rates from at least some of the additional plurality of TNC error rates; as well as using the additional TNC group error rate to determine the additional first value indicative of the expected number of mutations present in the corresponding sequencing data; using at least a second subset of the additional sequence reads to determine an additional second value indicating the actual number of mutations present at the position; and using the additional first value indicating the expected number of mutations due to sequencing errors present in the corresponding sequencing data and the additional second value indicating the actual number of mutations present in the corresponding sequencing data at the position of the mutation to be monitored to determine whether the corresponding sequencing data provides the indication that the subject suffers from minimal residual disease.

46. ​​A system for determining whether sequencing data of a biological sample of a subject provides an indication that the subject has minimal residual disease, the system comprising: at least one computer hardware processor; as well as at least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by the at least one computer hardware processor, cause the at least one computer hardware processor to: (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored; (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising: determining a plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types using the first subset of sequence reads; grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups; determining a TNC group error rate for the plurality of TNC error rate groups using TNC error rates of at least some of the plurality of TNC error rates; and using the TNC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data; (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

47. The system of claim 46, wherein the at least one computer hardware processor stores processor-executable instructions that cause the at least one computer hardware processor to perform the method of any one of claims 2 to 45.

48. At least one non-transitory computer-readable storage medium storing processor-executable instructions that, when executed by at least one computer hardware processor, cause the at least one computer hardware processor to: (A) obtaining the sequencing data, which was previously generated by sequencing the biological sample of the subject, and which includes sequence reads covering the positions of the mutations to be monitored; (B) determining a first value indicating an expected number of mutations due to sequencing errors present in the sequencing data using at least a first subset of the sequence reads, the determining comprising: determining a plurality of trinucleotide nucleoside (TNC) error rates corresponding to a plurality of TNC error types using the first subset of sequence reads; grouping at least some of the plurality of TNC error rates into a plurality of TNC error rate groups; determining a TNC group error rate for the plurality of TNC error rate groups using TNC error rates of at least some of the plurality of TNC error rates; and using the TNC group error rate to determine the first value indicative of the expected number of mutations present in the sequencing data; (C) using at least a second subset of the sequence reads to determine a second value indicative of an actual number of mutations present at the location of the mutation to be monitored; and (D) using the first value indicating the expected number of mutations present in the sequencing data due to sequencing errors and the second value indicating the actual number of mutations present in the sequencing data at the location of the mutation to be monitored to determine whether the sequencing data provides the indication that the subject has minimal residual disease.

49. The at least one non-transitory computer-readable storage medium storing processor-executable instructions of claim 48, wherein the at least one computer hardware processor stores processor-executable instructions that cause the at least one computer hardware processor to perform the method of any one of claims 2 to 45.

Images