Identification and exclusion of variants with outlier allele fractions in tumor fraction estimations

Abstract

The present disclosure relates to a method for improving tumor fraction estimation by identifying and excluding variants having outlier allele fractions. For example, the methods described herein include assessing, for each identified somatic variant, an allele fraction, wherein if the allele fraction for the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant is excluded from a TF estimation.

Description

IDENTIFICATION AND EXCLUSION OF VARIANTS WITH OUTLIER ALLELE FRACTIONS IN TUMOR FRACTION ESTIMATIONSBACKGROUND1. FIELD OF ART

[0001] This disclosure relates to improving tumor fraction estimation.2. DESCRIPTION OF THE RELATED ART

[0002] Computational techniques can be used on DNA sequencing data to identify mutations or variants in DNA that can correspond to various types of cancer or other diseases. To conduct a survey of tumor fractions across cfDNA samples, targeted sequencing panels (small nucleotide variants, not methylation variants) are designed to identify somatic variants called from whole-genome sequencing of each participant’s tumor biopsy. Rather than ordering one panel per participant, the sequencing data from multiple participants can be combined into a single panel, both to reduce cost and simplify operations. Another important consideration that arises in the process of combining sequencing data from multiple participants into a single panel is how tumor fraction estimation is impacted by variants with outlier allele fractions. One approach would be to figure out a way to identify and exclude the variants having outlier allele fractions.SUMMARY

[0003] The methods described herein provide a method for improving tumor fraction (TF) estimation in a sample by identifying and excluding one or more somatic variants comprising an outlier allele fraction from the TF estimation. Here, the analysis begins with the aim of identifying a somatic variant comprising an outlier allele fraction from the TF estimation. By excluding the identified somatic variants having an outlier allele fraction from the TF estimation, the overall TF estimation can be improved (e.g., lower the limit of detection). The method includes identifying somatic variants in the sequencing data; determining the current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction, wherein if the allele fraction for the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more excluded somatic variants. In some instances, determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = #of total observations across all variants for this individual. The method can also include iterating through the steps until all or a portion of the identified somatic variants having a shedding rate that significantly deviate from the current shedding rate are excluded from the TF estimation. Here, a somatic value can be swapped for a feature value that can be used in tumor fraction estimation. In one example, a feature value is a somatic variant and the somatic variant is identified because the somatic variant is a variant (SNP) that has been previously shown to indicate a disease presence and / or a disease type.

[0004] In another embodiment, this disclosure features a method for detecting clonal hematopoiesis of indeterminate potential (CHIP) in a sample comprising cell-free DNA (cfDNA), the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining the current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; and determining, based on the allele fraction, if an identified somatic variant derives from CHIP, wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant the somatic variant is identified as deriving from CHIP.

[0005] Overall, this disclosure provides a framework for using feature values (e.g., somatic variants) for improving tumor fraction (TF) estimation. In some case, this is applied to detection of clonal hematopoiesis of indeterminate potential (CHIP) where detecting of CHIP (or CHIP variants) and removing these variants enables improved TF estimation.BRIEF DESCRIPTION OF DRAWINGS

[0006] FIG. 1 is flowchart of a method for preparing a nucleic acid sample for sequencing according to one embodiment.

[0007] FIG. 2 is block diagram of a processing system for processing sequence reads according to one embodiment.

[0008] FIG. 3 is flowchart of a method for determining variants of sequence reads according to one embodiment.

[0009] FIG. 4 is a flow chart of a workflow for improving sequencing panel assignment according to one embodiment.

[0010] FIG. 5 is a plot showing Tumor Fraction (TF) estimation for patient-specific panels (y-axis) over whole genome sequencing (WGS)-based TF estimation.

[0011] FIG. 6 is a plot showing the altcount (total count minus the reference count) over the total count for data from the same patient specific panels shown in FIG. 5.

[0012] FIGs. 7A-7B shows a plot of the number of removed “sites” per panel participant where a binomial p-value filter removes “sites” if they have a p-value < IE-6. FIG. 7A shows number of removed sites per participant with the x-axis value presented up to 200 sites. FIG. 7B shows number of removed sites per participant with the x-axis values present up to 10 sites.

[0013] FIG. 8 shows a plot of mutations per patient in cfDNA (y-axis) over age (x-axis) in cancer versus non-cancer for the number of WBC-matched nonsynonmous mutations pers patient in cfDNA.

[0014] FIG. 9 is a plot showing the percent of nonsynonymous variants (y-axis) with the number of patients in whom the mutation is found on the x-axis.

[0015] FIG. 10 is a bar graph for the number of CHIP variants per gene with the top 15 genes presented. In total 1072 CHIP variants are represented in FIG. 10, which represent about 49% of the total number of CHIP variants

[0016] FIG. 11 is a plot showing count number of filtered variants in lOOkb bins across the genome.

[0017] FIG. 12 is a plot showing the fraction of variants filtered for each of the indicated ages on the x-axis.

[0018] FIGs. 13A-13B are tables that show the observed allele fraction for the filtered variants in white blood cells. FIG. 13 A shows non-normalized data, and FIG. 13B shows normalized data.

[0019] FIG. 14A is a plot of allele fraction (AF) in patient-specific panels versus AF in WBC. Data shows that 40 of the filtered variants have a WBC alt! = 0.

[0020] FIG. 14B is a plot of count versus P(WBC alt == 0) (from FIG. 14A), which shows that 82 filtered variants have WBC alt == 0.

[0021] FIG. 15A is a plot of counts for each fraction of MAPQ60 fragments as indicated on the x-axis.

[0022] FIG. 15B is a plot of counts for each of the WBC pvalues (using MAPQ60) indicated on the x-axis.

[0023] FIG. 16 is a plot of estimated tumor fraction for 23 healthy WBC samples following deconvolution. Black bars are references ranges take from https: / / www.stemcell.com / media / files / wallchart / WA10006- Frequencies_Cell_Types_Human_ Peripheral_Blood.pdf.

[0024] FIG. 17A is a plot showing filtered variants (y-axis)) for various tumor fractions.

[0025] FIG. 17B is a plot showing panel-based TF estimates (filtered variants (y-axis)) for various WGS-based TF estimates (tumor fraction sv (x-axis)). Box indicates the samples shown in FIG. 17 A.DETAILED DESCRIPTIONI. DEFINITIONS

[0026] The term “individual” refers to a human individual. The term “healthy individual” refers to an individual presumed to not have a cancer or disease. The term “subject” refers to an individual who is known to have, or potentially has, a cancer or disease.

[0027] The term “sequence reads” refers to nucleobase sequences read from a sample obtained from an individual. Sequence reads can be obtained through various methods known in the art.

[0028] The term “read segment” or “read” refers to any nucleobase sequences including sequence reads obtained from an individual and / or nucleobase sequences derived from the initial sequence read from a sample obtained from an individual. For example, a read segment can refer to an aligned sequence read, a collapsed sequence read, or a stitched read. Furthermore, a read segment can refer to an individual nucleobase base, such as a single nucleobase variant.

[0029] The term “single nucleobase variant” or “SNV” refers to a substitution of one nucleobase to a different nucleobase at a position (e.g., site) of a nucleobase sequence, e.g., a sequence read from an individual. A substitution from a first nucleobase X to a second nucleobase Y can be denoted as “X>Y.” For example, a cytosine to thymine SNV can be denoted as “OT.”

[0030] The term “indel” refers to any insertion or deletion of one or more base pairs having a length and a position (which can also be referred to as an anchor position) in a sequence read. An insertion corresponds to a positive length, while a deletion corresponds to a negative length.

[0031] The term “mutation” refers to one or more SNVs or indels.

[0032] The term “true positive” refers to a mutation that indicates real biology, for example, presence of a potential cancer, disease, or germline mutation in an individual. True positives are not caused by mutations naturally occurring in healthy individuals (e.g., recurrent mutations) or other sources of artifacts such as process errors during assay preparation of nucleic acid samples.

[0033] The term “false positive” refers to a mutation incorrectly determined to be a true positive. Generally, false positives can be more likely to occur when processing sequence reads associated with greater mean noise rates or greater uncertainty in noise rates.

[0034] The term “cell-free nucleic acid,” “cell-free DNA,” or “cfDNA” refers to nucleic acid fragments that circulate in an individual’s body (e.g., bloodstream) and originate from one or more healthy cells and / or from one or more cancer cells. cfDNA can be obtained from a blood sample.

[0035] The term “circulating tumor DNA” or “ctDNA” refers to nucleic acid fragments that originate from tumor cells or other types of cancer cells, which can be released into an individual’s bloodstream as result of biological processes such as apoptosis or necrosis of dying cells or actively released by viable tumor cells. In some cases, ctDNA is DNA found in cfDNA.

[0036] The term “genomic nucleic acid,” “genomic DNA,” or “gDNA” refers to nucleic acid including chromosomal DNA that originates from one or more healthy cells. In some cases, white blood cells are assumed to be healthy cells.

[0037] The term “white blood cell DNA,” or “wbcDNA” refers to nucleic acid including chromosomal DNA that originates from white blood cells. Generally, wbcDNA is gDNA and is assumed to be healthy DNA.

[0038] The term “tissue nucleic acid,” “cancerous tissue DNA,” or “tDNA” refers to nucleic acid including chromosomal DNA from tumor cells or other types of cancer cells that are obtained from cancerous tissue or a tumor. In some cases, tDNA is obtained from a biopsy of a tumor.

[0039] The term “alternative allele” or “ALT” refers to an allele having one or more mutations relative to a reference allele, e.g., corresponding to a known gene.

[0040] The term “sequencing depth” or “depth” refers to a total number of read segments from a sample obtained from an individual.

[0041] The term “alternate depth” or “AD” refers to a number of read segments in a sample that support an ALT, e.g., include mutations of the ALT.

[0042] The term “alternate frequency” or “AF” refers to the frequency of a given ALT. The AF can be determined by dividing the corresponding AD of a sample by the depth of the sample for the given ALT.

[0043] As used herein, the term “feature value” refers to a characteristic or set of characteristics used to define a sample. A “feature value” can include a single characteristicat a single genomic locus or a combination of characteristics across a plurality of genomic loci.

[0044] As used herein, the term “sequencing panel” refers to a combination of sequencing data from two or more sample (e.g., individuals).

[0045] As used herein, the terms “mean target coverage” or “MTC” refer to the total number of targeted bases divided by the targeted region size.II . EXAMPLE AS SAY PROTOCOL

[0046] FIG. 1 is flowchart of a method for preparing a nucleic acid sample for sequencing according to one embodiment. The workflow 100 includes, but is not limited to, the following steps. For example, any step of the workflow 100 can comprise a quantitation sub-step for quality control or other laboratory assay procedures known to one skilled in the art.

[0047] In step 110, a nucleic acid sample (DNA or RNA) is extracted from a subject. In the present disclosure, DNA and RNA can be used interchangeably unless otherwise indicated. That is, the following embodiments for using error source information in variant calling and quality control can be applicable to both DNA and RNA types of nucleic acid sequences. However, the examples described herein can focus on DNA for purposes of clarity and explanation. The sample can be any subset of the human genome, including the whole genome. The sample can be extracted from a subject known to have or suspected of having cancer. The sample can include blood, plasma, serum, urine, fecal, saliva, other types of bodily fluids, or any combination thereof. In some cases, the sample can include tissue or bodily fluids extracted from tissue. In some embodiments, methods for drawing a blood sample (e.g., syringe or finger prick) can be less invasive than procedures for obtaining a tissue biopsy, which can require surgery. The extracted sample can include cfDNA and / or ctDNA. For healthy individuals, the human body can naturally clear out cfDNA and other cellular debris. If a subject has a cancer or disease, ctDNA in an extracted sample can be present at a detectable level for diagnosis.

[0048] Additionally, the extracted sample can include wbcDNA. Extracting the nucleic acid sample can further include separating the cfDNA and / or ctDNA from the wbcDNA. Extracting the wbcDNA from the cfDNA and / or ctDNA can occur when the DNA is separated from the sample. In the case of a blood sample, the wbcDNA is obtained from a buff coat fraction of the blood sample. The wbcDNA can be sheared to obtain wbcDNA fragments less than 300 base pairs in length. Separating the wbcDNA from the cfDNAand / or ctDNA allows the wbcDNA to be sequenced independently from the cfDNA and / or ctDNA. Generally the sequencing process for wbcDNA is similar to the sequencing process for cfDNA and / or ctDNA.

[0049] In step 120, a sequencing library is prepared. During library preparation, unique molecular identifiers (UMI) are added to the nucleic acid molecules (e.g., DNA molecules) through adapter ligation. The UMIs are short nucleic acid sequences (e.g., 4-10 base pairs) that are added to ends of DNA fragments during adapter ligation. In some embodiments, UMIs are degenerate base pairs that serve as a unique tag that can be used to identify sequence reads originating from a specific DNA fragment. During PCR amplification following adapter ligation, the UMIs are replicated along with the attached DNA fragment, which provides a way to identify sequence reads that came from the same original fragment in downstream analysis.

[0050] In step 130, targeted DNA sequences are enriched from the library. During enrichment, hybridization probes (also referred to herein as “probes”) are used to target, and pull down, nucleic acid fragments informative for the presence or absence of cancer (or disease), cancer status, or a cancer classification (e.g., cancer type or tissue of origin). For a given workflow, the probes can be designed to anneal (or hybridize) to a target (complementary) strand of DNA or RNA. The target strand can be the “positive” strand (e.g., the strand transcribed into mRNA, and subsequently translated into a protein) or the complementary “negative” strand. The probes can range in length from 10s, 100s, or 1000s of base pairs. In one embodiment, the probes are designed based on a gene panel to analyze particular mutations or target regions of the genome (e.g., of the human or another organism) that are suspected to correspond to certain cancers or other types of diseases. Moreover, the probes can cover overlapping portions of a target region. By using a targeted gene panel rather than sequencing all expressed genes of a genome, also known as “whole exome sequencing,” the workflow 100 can be used to increase sequencing depth of the target regions, where depth refers to the count of the number of times a given target sequence within the sample has been sequenced. Increasing sequencing depth reduces required input amounts of the nucleic acid sample. After a hybridization step, the hybridized nucleic acid fragments are captured and can also be amplified using PCR.

[0051] In step 140, sequence reads are generated from the enriched DNA sequences. Sequencing data can be acquired from the enriched DNA sequences by known means in the art. For example, the workflow 100 can include next generation sequencing (NGS)techniques including synthesis technology (Illumina), pyrosequencing (454 Life Sciences), ion semiconductor technology (Ion Torrent sequencing), single-molecule real-time sequencing (Pacific Biosciences), sequencing by ligation (SOLiD sequencing), nanopore sequencing (Oxford Nanopore Technologies), or paired-end sequencing. In some embodiments, massively parallel sequencing is performed using sequencing-by-synthesis with reversible dye terminators.

[0052] In some embodiments, the sequence reads can be aligned to a reference genome using known methods in the art to determine alignment position information. The alignment position information can indicate a beginning position and an end position of a region in the reference genome that corresponds to a beginning nucleobase base and end nucleobase base of a given sequence read. Alignment position information can also include sequence read length, which can be determined from the beginning position and end position. A region in the reference genome can be associated with a gene or a segment of a gene. As cfDNA and / or ctDNA and wbcDNA are sequenced independently, sequence reads for both cfDNA and or ctDNA and wbcDNA are independently generated.

[0053] In various embodiments, a sequence read is comprised of a read pair denoted as and R2. For example, the first read R can be sequenced from a first end of a nucleic acid fragment whereas the second read R2can be sequenced from the second end of the nucleic acid fragment. Therefore, nucleobase base pairs of the first read R and second read R2can be aligned consistently (e.g., in opposite orientations) with nucleobase bases of the reference genome. Alignment position information derived from the read pair R and R2can include a beginning position in the reference genome that corresponds to an end of a first read (e.g., R- ) and an end position in the reference genome that corresponds to an end of a second read (e.g., R2). In other words, the beginning position and end position in the reference genome represent the likely location within the reference genome to which the nucleic acid fragment corresponds. An output file having SAM (sequence alignment map) format or BAM (binary) format can be generated and output for further analysis such as variant calling, as described below with respect to FIG. 2.III. EXAMPLE PROCESSING SYSTEM

[0054] FIG. 2 is block diagram of a processing system 200 for processing sequence reads and generating sequence panel assignments according to one embodiment. The processing system 200 includes a sequence processor 205, sequence database 210, model database 215, machine learning engine 220, models 225 (for example, including a classifier model and oneor more Bayesian hierarchical models or joint models), parameter database 230, score engine 235, variant caller 240, and a sequencing panel generator 250.III. A FEATURE VALUES / SOMATIC VA IANTS

[0055] In FIG. 2, the sequencing panel generator 250 selects a feature value(s) (e.g., somatic variant) from the sequencing data using various features, scores, sequences, or a combination thereof. In some embodiments, the feature value(s) are pre-determined (e.g., a pre-determined set of variants). In some embodiments, the feature value(s) are determined by feature value identifier 260. In one embodiment, a feature value indicates a presence or absence of a variant, a mean allele frequency, a total number of small variants, and an allele frequency of true variants.

[0056] The sequencing panel generator 250 can include a feature value identifier 260 that stores features values (e.g., store somatic variants). The feature can be any one or more of, without limitation, a presence or absence of a variant, a mean allele frequency, a total number of small variants, and an allele frequency of true variants.

[0057] In one embodiment, a feature value comprises a variant. In such cases, the variant comprises one or more of the following of a single nucleotide variant, an insertion, and a deletion. In one embodiment, a variant is associated with a genomic region described herein and / or as described U.S. Pat. Pub. No. 2019 / 0073445 Al.III.B DETERMINING VARIANTS (FEATURE VALUES) FROM SEQUENCING DATA

[0058] FIG. 3 is an example where the feature value is a variant and the sum of the information gathered for each variant can be considered a feature value and used according to the methods described herein. In particular, FIG. 3 is a flowchart of a workflow for determining variants of sequence reads according to one embodiment.

[0059] In some embodiments, the processing system 200 uses the workflow 300 to perform variant calling (e.g., for SNVs and / or indels) based on input sequencing data. Further, the processing system 200 can obtain the input sequencing data from an output file associated with nucleic acid sample prepared using the workflow 100 described above. The workflow 300 includes, but is not limited to, the following steps, which are described with respect to the components of the processing system 200. In other embodiments, one or more steps of the workflow 300 can be replaced by a step of a different process for generating variant calls, e.g., using Variant Call Format (VCF), such as HaplotypeCaller, VarScan, Strelka, or SomaticSniper.

[0060] At step 310, the sequence processor 205 collapses aligned sequence reads of the input sequencing data. In one embodiment, collapsing sequence reads includes using UMIs, and optionally alignment position information from sequencing data of an output file (e.g., from the workflow 100 shown in FIG. 1) to collapse multiple sequence reads into a consensus sequence for determining the most likely sequence of a nucleic acid fragment or a portion thereof. Since the UMIs are replicated with the ligated nucleic acid fragments through enrichment and PCR, the sequence processor 205 can determine that certain sequence reads originated from the same molecule in a nucleic acid sample. In some embodiments, sequence reads that have the same or similar alignment position information (e.g., beginning and end positions within a threshold offset) and include a common UMI are collapsed, and the sequence processor 205 generates a collapsed read (also referred to herein as a consensus read) to represent the nucleic acid fragment. The sequence processor 205 designates a consensus read as “duplex” if the corresponding pair of collapsed reads have a common UMI, which indicates that both positive and negative strands of the originating nucleic acid molecule is captured; otherwise, the collapsed read is designated “non-duplex.” In some embodiments, the sequence processor 205 can perform other types of error correction on sequence reads as an alternate to, or in addition to, collapsing sequence reads.

[0061] At step 315, the sequence processor 205 stitches the collapsed reads based on the corresponding alignment position information. In some embodiments, the sequence processor 205 compares alignment position information between a first read and a second read to determine whether nucleobase base pairs of the first and second reads overlap in the reference genome. In one use case, responsive to determining that an overlap (e.g., of a given number of nucleobase bases) between the first and second reads is greater than a threshold length (e.g., threshold number of nucleobase bases), the sequence processor 205 designates the first and second reads as “stitched”; otherwise, the collapsed reads are designated “unstitched.” In some embodiments, a first and second read are stitched if the overlap is greater than the threshold length and if the overlap is not a sliding overlap. For example, a sliding overlap can include a homopolymer run (e.g., a single repeating nucleobase base), a dinucleobase run (e.g., two-nucleobase base sequence), or a trinucleobase run (e.g., three- nucleobase base sequence), where the homopolymer run, dinucleobase run, or trinucleobase run has at least a threshold length of base pairs.

[0062] At step 320, the sequence processor 205 assembles reads into paths. In some embodiments, the sequence processor 205 assembles reads to generate a directed graph, forexample, a de Bruijn graph, for a target region (e.g., a gene). Unidirectional edges of the directed graph represent sequences of k nucleobase bases (also referred to herein as “k- mers”) in the target region, and the edges are connected by vertices (or nodes). The sequence processor 205 aligns collapsed reads to a directed graph such that any of the collapsed reads can be represented in order by a subset of the edges and corresponding vertices.

[0063] In some embodiments, the sequence processor 205 determines sets of parameters describing directed graphs and processes directed graphs. Additionally, the set of parameters can include a count of successfully aligned k-mers from collapsed reads to a k-mer represented by a node or edge in the directed graph. The sequence processor 205 stores, e.g., in the sequence database 210, directed graphs and corresponding sets of parameters, which can be retrieved to update graphs or generate new graphs. For instance, the sequence processor 205 can generate a compressed version of a directed graph (e.g., or modify an existing graph) based on the set of parameters. In one use case, in order to filter out data of a directed graph having lower levels of importance, the sequence processor 205 removes (e.g., “trims” or “prunes”) nodes or edges having a count less than a threshold value and maintains nodes or edges having counts greater than or equal to the threshold value.

[0064] In one embodiment, the processing system 200 can store sequencing data in a database 210 (e.g., variants and normals), which can be used to detect presence, absence, or level of a feature values (e.g., a presence or absence of a variant, a mean allele frequency, a total number of small variants, and an allele frequency of true variants) in a sample from a subject, and / or otherwise predict cost associated with the variant (e.g., per-site cost values and relative costs). The sequence database 210 can also store sequencing data processed by the system 200, but can also store sequencing data not processed by the system 200, such as sequencing data uploaded from an external source and / or otherwise retrieved from external or publicly available databases.

[0065] At step 325, the variant caller 240 generates candidate variants from the paths assembled by the sequence processor 205. In one embodiment, the variant caller 240 generates the candidate variants by comparing a directed graph (which can have been compressed by pruning edges or nodes in step 310) to a reference sequence of a target region of a genome. The variant caller 240 can align edges of the directed graph to the reference sequence and records the genomic positions of mismatched edges and mismatched nucleobase bases adjacent to the edges as the locations of candidate variants. Additionally, the variant caller 240 can generate candidate variants based on the sequencing depth of atarget region. In particular, the variant caller 240 can be more confident in identifying variants in target regions that have greater sequencing depth, for example, because a greater number of sequence reads help to resolve (e.g., using redundancies) mismatches or other base pair variations between sequences.

[0066] In one embodiment, the variant caller 240 generate candidate variants using a variant model 225 to determine expected noise rates for sequence reads from a subject. The variant model 225 can be a Bayesian hierarchical model, though in some embodiments, the processing system 200 uses one or more different types of models. Moreover, a Bayesian hierarchical model can be one of many possible model architectures that can be used to generate candidate variants and which are related to each other in that they all model position-specific noise information in order to improve the sensitivity / specificity of variant calling. More specifically, the machine learning engine 220 trains the variant model 225 using samples from healthy individuals to model the expected noise rates per position of sequence reads.

[0067] Further, multiple different models can be stored in the model database 215 or retrieved for application post-training. For example, a first model is trained to model SNV noise rates and a second model is trained to model indel noise rates. Further, the score engine 235 can use parameters of the variant model 225 to determine a likelihood of one or more true positives in a sequence read. The score engine 235 can determine a quality score (e.g., on a logarithmic scale) based on the likelihood. For example, the quality score is a Phred quality score Q = — 10 ■ log10P , where P is the likelihood of an incorrect candidate variant call (e.g., a false positive).

[0068] At step 330, the score engine 235 scores the candidate variants based on the variant model 225 or corresponding likelihoods of true positives or quality scores.

[0069] At step 335, the processing system 200 outputs the candidate variants. In some embodiments, the processing system 200 outputs some or all of the determined candidate variants along with the corresponding scores. Downstream systems, e.g., external to the processing system 200 or other components of the processing system 200, can use the candidate variants and scores for various applications including, but not limited to, predicting presence of cancer, disease, or germline mutations.

[0070] In one embodiment, candidate variants are outputted for both cfDNA and / or ctDNA and wbcDNA. Herein, generally, candidate variants for wbcDNA are “normals” while candidate variants for cfDNA and / or ctDNA are “variants.” Various detection methodsand models can compare variants to normals to determine if the variants include signatures of cancer or any other disease. In various embodiments, normals and variants can be generated using any other process, any number of samples (e.g., a tumor biopsy or blood sample), or accessed from a database storing candidate variants.

[0071] In one embodiment, the outputted candidate variants are used in the methods described herein to generate an optimized sequencing panel assignment.III. C GENERATING AN OPTIMIZED SEQUENCING PANEL ASSIGNMENT

[0072] In FIG. 2, the sequencing panel generator 250 generates a targeted sequencing panel assignment using various features, scores, sequences, etc. determined by the processing system 200.

[0073] In some embodiments, the sequencing panel generate 250 is used to select at least a subset of identified feature values.

[0074] In one embodiment, the targeted sequencing panel generator 250 employs a machine learning model (e.g., a classifier model) to determine the targeted sequencing panel assignment based on the feature values from the sequencing data.

[0075] In one embodiment, the targeted sequencing panel generator 250 ranks the samples based on decreasing feature values. In another embodiments, the targeted sequence panel generator 250 ranks the samples based on the increasing feature values.

[0076] In some embodiments, the targeted sequencing panel includes about 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 or more samples. In one embodiment, the sequencing panel includes 16 samples. In another embodiment, the sequencing panel has no more than 16 samples.

[0077] In one embodiment, the method described herein includes a benchmarking sample. In one embodiments, the method described herein includes no more than one benchmarking sample. The benchmarking sample can include one or more known feature values. In some cases, the benchmarking sample can include one or more known feature values whose amounts in the benchmarking sample are known.

[0078] The targeted sequencing panel generator then employs a machine learning model that can employ different models to determine the targeted sequencing panel assignment based on the feature values from the sequencing data.

[0079] The machine learning model is selected from: a classifier model, a pre-specified algorithm, and a regression model. In one embodiment, the machine learning model is a classifier model.

[0080] In one embodiment, after the samples are ranked based on decreasing feature value or the feature values are ranked by decreasing (or increasing) value, the classifier model applies a greedy algorithm to add a next-highest ranked sample (or feature value) of the remaining ranked samples (or feature values) to a panel, wherein the panel to which the sample is sorted comprises the lowest value of feature values. This greedy algorithm is applied until a sequencing panel produces a limit of detection below a threshold.

[0081] In one embodiment, applying the classifier model includes: seeding the targeted sequencing panel with the subset of identified feature values; swapping an identified feature value in the selected subset of identified feature values for an identified feature value not in the selected subset of identified feature values, measuring the limit of detection after swapping; and assessing whether the optimized targeted sequencing panel produces a limit of detection below the threshold; optionally, iterating until a targeted sequencing panel is selected that produces a limit of detection below the threshold.

[0082] In some embodiments, the sequencing panel generator 250 generates the optimized targeted sequence panel assignment by applying a seed and swap approach to compiling a targeted sequencing panel. In some embodiments, the sequencing panel generator 250 iterates through samples to assign the feature values (and corresponding primers for amplifying the feature value) to a panel, determines mean of feature values for each targeted sequencing panel, swapping two feature values between two different targeted sequencing panels, and measuring deviation of mean feature value for each of the two different targeted sequencing panels following the swap. In one embodiment, if the change in feature value of the two different targeted sequencing panels is decreased by the swap, then it is accepted. In some embodiments, the targeted sequencing panel generator includes repeating these steps for a pre-specified number of swaps, thereby generating a targeted sequencing panel assignment based on the feature values from the sequencing data. In such cases, the repeating step is performed until the targeted sequencing panel selected is capable of producing a limit of detection below a threshold.IV. EXAMPLE IDENTIFICATION AND EXCLUSION OF VARIANTS WITH OUTLIER ALLELE FRACTIONS

[0083] As described herein, FIG. 4 shows an example workflow for sample by identifying and excluding a somatic variant comprising an outlier allele fraction, and then performing a TF estimation based on the identified somatic variants without relying on the one or more excluded . The workflow 400 can be executed by the system 200 or another similar system.

[0084] The sequencing panel generator 250 obtains (retrieves) 410 sequencing data (e.g., test sequences) for a set of samples (e.g., here samples that meet a set of criteria described herein). The first sequencing data can be the CCGA indicator set but could be another set of genomic regions to be analyzed. The sequencing data is associated with a number of test sequences, and is associated with feature values (e.g., a presence or absence of a variant, a mean allele frequency, a total number of small variants, an allele frequency of true variants a GC content, an error rate, and a sequencing depth count,).

[0085] The sequencing panel generator 250 identifies 412 one or more feature values (e.g., somatic variant) from the sequencing data. In some embodiments, the feature value is a somatic variant. In some cases, the feature value can be GC content, sequencing depth, a mean allele frequency, a total number of small variants, and an allele frequency of true variants in the sequencing data. Other feature values are also possible.

[0086] In some embodiments, the sequencing panel generator 250 can access one or more additional sets of feature values (somatic variants) and apply a machine learning model to the samples based on the additional set of feature values (e.g., additional set of somatic variants). In such cases, the variant caller 240 can identify one or more additional subsets of feature values (e.g., somatic variants) for consideration from the additional set of feature values when identifying which variants having an outlier allele fraction to exclude. In one embodiments, the variant caller 240 can identify one or more additional subsets of feature values (e.g., somatic variants) for consideration when identifying which variants having an outlier allele fraction to exclude.

[0087] The shedding rate calculator 270 determines 414 the current shedding rate for all the identified somatic variants, for example, by determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

[0088] In some embodiments, the shedding rate calculator 270 assesses 416, for each identified somatic variant, an allele fraction, wherein if the allele fraction for the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant is excluded from a TF estimation.

[0089] In some embodiments, the shedding rate calculator 270 assesses 416 the allele fraction for each identified somatic variant by determining a significance metric compared to the current shedding rate. In some embodiments, determining the significance metric comprises calculating a p-value. In some embodiments, calculating the p-value comprisesusing a binomial distribution. In some embodiments, an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

[0090] In some embodiments, the method includes iterating through all or a portion of the steps until all or a portion of the identified somatic variants having a shedding rate that significantly deviate from the current shedding rate are excluded from the TF estimation.

[0091] In some embodiments, the method includes filtering and excluding germline variants from the identified somatic variants.

[0092] In some embodiments, the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals. In some embodiments, the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X. In some embodiments, the sequencing data is generated from a biological sample sequenced at a depth of at least 30X. In some embodiments, the sequencing data comprises from two or more biological samples.

[0093] In some embodiments, the sequencing data comprises a control sample. In some embodiments, the control sample comprises a cfDNA from a healthy patient.

[0094] In some embodiments, the sequencing data comes from two or more biological samples where at least one of the biological samples is a benchmarking sample.

[0095] In some embodiments, the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

[0096] In some embodiments, the somatic variants are, or are associated with, previously identified somatic variants.

[0097] In some embodiments, the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

[0098] In some embodiments, the tumor specific nucleotide variants are associated with previously tumor specific somatic variants.

[0099] In addition to filtering (i.e., identifying and excluding variants having outlier allele fractions, there are several filtering methods that can improve tumor fraction estimation. In a first example, the sequencing panel used to amplify the cfDNA can only derive feature values for genomic regions having variants in a threshold number of sequences in the sequencing data. In a second example, the sequencing panel used to amplify the cfDNA can duplicate, or remove duplications, of a genomic region from a panel to increase detection capability. In a third example, a system administrator can remove genomic regions from the analysis. In afourth example, a system administrator can remove samples from the sequencing panel. Finally, the sequencing panel generator can remove feature values from the panel based on a feature value blacklist. The feature value blacklist can include patented feature values, feature values known to cause false positives, or any other feature value that could decrease the detection capability of a panel.

[0100] Following identification and exclusion of the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant, a tumor fraction estimation is performed 418 where the TF estimation is based on the identified somatic variants without relying on the one or more excluded somatic variants.

[0101] In another embodiment, described herein is a method for detecting clonal hematopoiesis of indeterminate potential (CHIP) in a sample comprising cell-free DNA (cfDNA). In some embodiments, this method can be executed by the system 200 or another similar system.

[0102] In some embodiments, the method for detecting CHIP in sample includes: retrieving sequencing data; identifying somatic variants in the sequencing data; determining the current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; and determining, based on the allele fraction, if an identified somatic variant derives from CHIP, wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant the somatic variant is identified as deriving from CHIP.

[0103] In some embodiments, determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

[0104] In some embodiments, assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate. In some embodiments, determining the significance metric comprises calculating a p-value. In some embodiments, calculating the p-value comprises using a binomial distribution. In some embodiments, an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

[0105] In some embodiments, the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

[0106] In some embodiments, the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or100X. In some embodiments, the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX. In some embodiments, the sequencing data comes from two or more biological samples.

[0107] In some embodiments, the two or more biological samples comprises a control sample. In some embodiments, the control sample comprises a cfDNA from a healthy patient. In some embodiments, the sequencing data comprises a benchmarking sample. In some embodiments, the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate. In some embodiments, the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

[0108] In another embodiments, this disclosure features a method for improving tumor fraction (TF) estimation in a sample by identifying and excluding a clonal hematopoiesis of indeterminate potential (CHIP) variant from the TF estimation. The method includes: retrieving sequencing data; identifying somatic variants in the sequencing data; determining the current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; determining, based on the allele fraction, if an identified somatic variant derives from CHIP, wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant is identified as deriving from CHIP and is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more exclude somatic variants identified as deriving from CHIP.

[0109] In some embodiments, determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

[0110] In some embodiments, assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate. In some embodiments, determining the significance metric comprises calculating a p-value.

[0111] In some embodiments, calculating the p-value comprises using a binomial distribution. In some embodiments, an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

[0112] In some embodiments, the method also includes: iterating through the steps until all or a portion of the identified somatic variants having a shedding rate that significantly deviates from the current shedding rate are excluded from the TF estimation.

[0113] In some embodiments, the method also includes filtering and excluding germline variants from the identified somatic variants.

[0114] In some embodiments, the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

[0115] In some embodiments, the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

[0116] In some embodiments, the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

[0117] In some embodiments, the sequencing data comprises from two or more biological samples. In some embodiments, the sequencing data comprises a control sample. In some embodiments, the control sample comprises a cfDNA from a healthy patient.

[0118] In some embodiments, the sequencing data comes from two or more biological samples where at least one of the biological samples is a benchmarking sample.

[0119] In some embodiments, the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

[0120] In some embodiments, the somatic variants are, or are associated with previously identified somatic variants.

[0121] In some embodiments, the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

[0122] In some embodiments, the somatic variants comprise tumor specific nucleotide variant.IV. A.1 SAMPLE SELECTION: SAMPLES FOR THE FRIST SET OF SEQUENCING DATA

[0123] As noted above, the steps for improving tumor fraction (TF) estimation include a consideration of the type of samples (e.g., target and control samples used to develop the method). Here, target samples selected for testing included one or more of (i) to be undetected by past classifiers (i.e., detectability), (ii) have available plasma tubes (i.e., available plasma), and (iii) contain current tumor fraction estimates less than 1%, if available (i.e., low estimated TF). As noted above, the pool of potential samples included samples from the Circulating Cell Genome Atlas 1 Study (CCGA1) and the Circulating Cell Genome Atlas 2 Study (CCGA2).

[0124] In some embodiments, the methods include benchmarking (control) samples. Benchmarking samples are contrived titrations with known tumor fractions (e.g., any of thetumor fraction values described herein). In some embodiments, the estimated tumor fraction (TF) of a benchmarking sample is at least 10%. In one embodiment, benchmarking samples can be used for benchmarking of variant calling. In one embodiment, benchmarking samples can also be used for TF estimation. Benchmarking samples were selected from CCGA1 participants that included WBC WGS data.IV.A.2 PANEL CONSIDERATIONS: NUMBER OF PARTICIPANTS INCLUDED IN THE SEQUENCING DATA

[0125] The number of participants (e.g., samples) per panel can be determined empirically. In some embodiments, determining the number of participants per panel includes the sequence cost and panel ordering costs. When sequencing panel assignments include a fixed number of participants, where each participant is assigned to a panel, the number of participants per panel is a tradeoff between the cost of ordering more panels and the increased cost of “wasted” sequencing on participants (i.e., samples) where the requisite information has already been derived.

[0126] In one embodiment, the number of participant (i.e., samples) per panel minimizes total cost as a function of the number of panels ordered (with fixed number of participants, number of panels is equivalent to participants per panel). Therefore, minimizing total costs can be represented as:

[0127] where N = number of panels, costrp = cost per read pair, depth = target raw sequencing depth, sites = number of variants per participant, nsampies = total number of participants. Here, the samples are assumed to have the same number of genomic regions per participant. In addition, the per panel cost is assumed to be constant.

[0128] As such, the cost of sequencing can be represented as: (Cost rp * depth * sites * N2samples)

[0129] Here, cost as a function of N is convex for N > 0. Solving for N, thereby determining the number of participants per panel can be represented as:IV.B IDENTIFYING SOMATIC VARIANTS (FEATURE VALUES)

[0130] In one embodiment, a feature value is a variant (e.g., a somatic variant). Somatic variants can be determined, for example, as described in Section III. B.

[0131] As noted above, the aim of the methods described herein is to improving tumor fraction (TF) estimation by identifying and excluding variants having an outlier allele fraction. Non-limiting examples of identifying somatic variants are as described in SectionIII.B.IV. C TARGETED SEQUENCING PANEL CONSIDERATION: NU BER OF FEATURE VALUESTO INCLUDE IN PANEL

[0132] As noted above, the steps for improving tumor fraction (TF) estimation include a consideration of the number of feature values (e.g., somatic variants) to include in the method. In some cases, the number of feature values (e.g., somatic variants) to include in the targeted sequencing panel can be determined empirically.

[0133] In one embodiment, the number of feature values (e.g. somatic variants) to include in the targeted sequencing panel is determined by calculating the estimated cTAF for targeted sequencing panels designed with varying numbers of somatic variants and selecting the number of somatic variants based on which panels produced a cTAF below a threshold. FIG. 8 shows a plot of targeted sequencing panels designed to include varying numbers of somatic variants and the corresponding cTAF for each panel. The targeted sequencing panels to the left of the line indicate those that have a cTAF below the desired LoD threshold.

[0134] In some embodiments, determining the number of variants per participant includes (i) confidence of the variant call, (ii) error rate of the site, and (iii) ease of sequencing / availability of cfDNA. In one embodiment, the number of variants per participant is less than 500 variants. In one embodiment, the number of variants per participant is 500 variants or greater.

[0135] In one embodiment, determining (i) confidence of the variant call includes a log-likelihood of a true call versus noise. This can be represented as:Binomfalt, tot, 0.5) LLR = log(— - - T2-Binom(alt, tot, noise)

[0136] In one embodiment, determining (ii) “error of the site” (i.e., site of the variant) depends, at least in part, on the conversion type of the variant (e.g., SNP). In one embodiment, greater error rates exist for conversion type including for A>G, T>C; OA,G>T; and OT, G>A. The variants having the lowest error rates met the criteria for “error rate of the site.”

[0137] In one embodiment, determining (iii) ease of sequencing / availability of cfDNA depends, at least in part, on the GC content. A skilled artisan would appreciate that other factors contribute to ease of sequencing. Here, relative coverage of genomic regions is reproducible across samples. As such, the consistent (or inconsistent) coverage is used in determining ease of sequencing.IV.D ASSAY PERFORMANCE

[0138] Analysis of previously sequence data showed that many samples with previously very low TF estimates now receive very high (>10%) TF estimates (see FIG. 5 and FIG. 6). This was likely driven by small number of variants with much higher AF (although too low for germline). This data highlighted a need for by identifying and excluding a somatic variant comprising an outlier allele fraction from a TF estimation.

[0139] One example filter is a binominal p-value filter, which includes calculating k = sum(alt counts), N = sum(total counts) across all sites -> p = k / N; and calculating P(alt >= X | p) for a binomial distribution at each site. A p-value threshold that removes sites with outlier allele fractions is p-value < le-6. When applying the binomial p-value filter to the data generated using the previously described sequencing panels shows that for most participants, the number of sites removes is between 0-5 (FIGs. 7A-7B). Application of the binomial p- value filter can be performed until convergence is reached. In some cases, samples with many variants removed are high mean AF (> 10%) such that “smaller” relative variations are flagged. For example, observing 2800 / 5000 with an expected AF of 0.5 results in a filtered site.

[0140] In an effort to determine the source of the variants with outlier allele fractions several hypothesis were advance.

[0141] One explanation is that the variants with outlier allele fractions were the results of technical artifacts (e.g., contamination, alignment artifacts, missed collapsing, among others). With respect to contamination, this explanation was determined as being unlikely because the same variants observed in WGBS and WGS of tumor. With respect to alignment artifacts, this explanation was determined as being unlikely for similar reason: observed in WGBS and WGS despite bisulfite and aligner differences. For example, using a stricter variant selection for sequencing panels removed variants near indels, thereby supporting the assertion that alignment artifacts are unlikely to be the source of variants withoutlier allele fractions (AFs). With respect to “missed collapsing”, typical bag size is 5-10 and outlier variant AFs can be orders of magnitude higher than the bulk of the variants, so this also seems unlikely. Manual checking in a genome browser view provided further support that this was an unlikely source of the variants with the outlier AFs.

[0142] Another explanation was that the outlier Afs were a product of biological noise. For example, CHIP in infiltrating leukocytes (clonal hematopoiesis of indeterminate potential) presented as a strong possibility. Another example of biological noise is “subclones.” Subclones are a possible source of the outlier AFs but less likely than CHIPs because the filtered variants have similar AFs in tumor as unfiltered variants, so this would require similar sampling of clones in tumor sequencing, but there are orders of magnitude difference in shedding so this may not be the source of the variants with outlier AFs.

[0143] With CHIP emerging as a likely source of the variants with outlier AFs, an analysis of typical CHIP mutations was performed.

[0144] As shown in FIG. 8 and FIG. 9, most CHIPs increase with age in an individual (FIG. 8) and are unique to individual patients (FIG. 9). Additional analysis of CHIP variants revealed concentration in particular genes (FIG. 10). Notably, 49% of CHIP variants were found in the top 15 genes in FIG. 10.

[0145] With the aim of improving tumor fraction (TF) estimation in a sample by identifying and excluding a somatic variant comprising an outlier allele fraction from the TF estimation the method described herein was tested. The method, as noted above, included retrieving sequencing data; identifying somatic variants in the sequencing data; determining the current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction, wherein if the allele fraction for the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more excluded somatic variants.

[0146] The variants with outlier AFs identified and filtered were analyzed. FIG. 11 shows count number of filtered variants in lOOkb bins across the genome. Notably, if filtered variants were CHIP variants, may have expected concentration in implicated genes (e.g.DNMT3 A, TET2). But maybe not surprising given that variants were selected to be private to each individual, and common population variants were filtered. The filtered variants showed weak correlation of fraction of filtered variants with age (FIG. 12).

[0147] In addition to analyzing the indirect signals of CHIP (e.g., genomic region and age dependence) WBC sequencing for the participants was analyzed.

[0148] Looking at 91 participants that also have WGS of WBC and WGS of tumor biopsy, 211 filtered variants were identified. This number was reduced to 122 filtered variants after removing filtered variants with 0 alternate alleles in targeted sequencing, which was the product current implementation: thresholding on p(alt > X) instead of p(alt >= X)). This resulted in “observedWbc” if at least one alt allele observed in WBC (~40X coverage). Overall, this data showed that filtered variants appear more often in WBC (see FIGs. 13A- 13B).

[0149] An analysis of Alt allele counts (e.g., the total count minus the ref count) in WBC showed consistency with cfDNA (see FIGs. 14A-14B). For this analysis, P(alt == 0) calculated from Binomial(N = WBC cov, p = AF in patient-specific panel). When alt allele counts are observed in WBC, AF was consistent with AF in patient-specific panel. Additionally, when alt alleles were not observed in WBC, which is generally consistent with the expected results. FIG. 14A shows a plot of allele fraction (AF) in patient-specific panels versus AF in WBC. The data shows that 40 of the filtered variants has a WBC alt! = 0. FIG. 14B shows a plot of count versus P(WBC alt == 0) (from FIG. 14 A), showing that 82 filtered variants has WBC alt == 0.

[0150] Notably, other filtered variants explained by low MAPQ (FIGs. 15A-15B). By default, the Pecan pipeline included fragments with MAPQ >= 15. For this analysis, the MAPQ can be restricted to MAPQ == 60 to reduce spurious counts without any significant loss of coverage (see FIGs. 15A-15B).

[0151] Assessment of healthy WBC samples was performed to see if deconvolution of 23 healthy WBC samples (Conversant) would produce expected results. Note, the WBC samples were not used in reference construction. As shown in FIG. 16, deconvolved fractions generally match reference ranges for WBC cell types. In line with expectations, non-blood cell types are assigned (close to) zero fraction (see FIG. 16).

[0152] Comparing the findings described herein to the literature, reveals several advantages of the present methods.

[0153] A first difference between the literature methods and the methods described include that the somatic variants were first selected using WGBS of tumor and cfDNA for germline filtering. In particular, ~30X tumor sequencing, which means the method wasn’t expected to select any CHIP variants present at low AF in tumor biopsy. Additionally, CHIPvariants present at similar AF in plasma cfDNA as in buffy coat gDNA would expect to have been removed with most of the high AF CHIP variants as “germline.” Thus, the putative CHIP variants identified herein due to surprisingly high AF compared to other tumor somatic variants. Lastly, the methods described herein enable “pre-selected” variants with high AF in tumor and low AF in WBC / plasma because the plasma cfDNA for is germline filtered prior to identifying varaints with outlier AFs.

[0154] Specific data showing that that methods described herein are able to improve TF estimation by identifying and excluding a variant having an outlier allele fraction include a survey of the 122 “fail” filter variants (i.e., variants that were identified and excluded based on the methods described herein) from 91 participants. This data provides evidence for the presence of CHIP variants, for example, the “fail filter” category was enriched for variants in WBC compared to pass filter; variants observed in WBC have similar AF as targeted; variants not observed in WBC have generally lower AF in targeted data / low WBC coverage; and relationship of fraction of variants failing filter with age.

[0155] Lastly, additional analysis revealed that CHIP variants could also explain discordant TF estimates. For example, samples with low panel-based TF estimates (<le-4) but high WGS-based TF estimates (>le-3) (see FIG. 17A) are explainable in the context of CHIP variants (FIG. 17B). In particular, filing WGS-based variants with any alt allele observed in WBC, recalculates “TF” as sum(alt) / sum(tot) (FIG. 17B).V. EXEMPLARY EMBODIMENTS

[0156] Embodiment 1. A method for improving tumor fraction (TF) estimation in a sample by identifying and excluding a somatic variant comprising an outlier allele fraction from the TF estimation, the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining a current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction, wherein if the allele fraction for the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more excluded somatic variants.

[0157] Embodiment 2. The method of embodiment 1, wherein determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

[0158] Embodiment 3. The method of embodiment 1 or 2, wherein assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate.

[0159] Embodiment 4. The method of embodiment 3, wherein determining the significance metric comprises calculating a p-value.

[0160] Embodiment 5. The method of embodiment 4, wherein calculating the p-value comprises using a binomial distribution.

[0161] Embodiment 6. The method of embodiment 4 or 5, wherein an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

[0162] Embodiment 7. The method of any one of embodiments 1-6, further comprising: iterating through the steps of embodiment 1 until all or a portion of the identified somatic variants having a shedding rate that significantly deviate from the current shedding rate are excluded from the TF estimation.

[0163] Embodiment 8. The method of any one of embodiments 1-7, further comprising filtering and excluding germline variants from the identified somatic variants.

[0164] Embodiment 9. The method of any one of embodiments 1-8, wherein the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

[0165] Embodiment 10. The method of any one of embodiments 1-9, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

[0166] Embodiment 11. The method of embodiment 10, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

[0167] Embodiment 12. The method of any one of embodiments 1-11, wherein the sequencing data comprises from two or more biological samples.

[0168] Embodiment 13. The method of embodiment 12, wherein the sequencing data comprises a control sample.

[0169] Embodiment 14. The method of embodiment 13, wherein the control sample comprises a cfDNA from a healthy patient.

[0170] Embodiment 15. The method any one of embodiments 1-14, wherein the sequencing data comes from two or more biological samples where at least one of the biological samples is a benchmarking sample.

[0171] Embodiment 16. The method of embodiment 15, wherein the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

[0172] Embodiment 17. The method of any one of embodiments 1-16, wherein the somatic variants are, or are associated with, previously identified somatic variants.

[0173] Embodiment 18. The method of any one of embodiments 1-17, wherein the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

[0174] Embodiment 19. The method of any one of embodiments 1-18, wherein the tumor specific nucleotide variants are associated with previously tumor specific somatic variants.

[0175] Embodiment 20. A method for detecting clonal hematopoiesis of indeterminate potential (CHIP) in a sample comprising cell-free DNA (cfDNA), the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining a current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; and determining, based on the allele fraction, if an identified somatic variant derives from CHIP, wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant the somatic variant is identified as deriving from CHIP.

[0176] Embodiment 21. The method of embodiment 20, wherein determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

[0177] Embodiment 22. The method of embodiment 20 or 21, wherein assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate.

[0178] Embodiment 23. The method of embodiment 22, wherein determining the significance metric comprises calculating a p-value.

[0179] Embodiment 24. The method of embodiment 23, wherein calculating the p-value comprises using a binomial distribution.

[0180] Embodiment 25. The method of embodiment 23 or 24, wherein an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p- value < le-6.

[0181] Embodiment 26. The method of any one of embodiments 20-25, wherein the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

[0182] Embodiment 27. The method of any one of embodiments 20-26, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

[0183] Embodiment 28. The method of embodiment 27, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

[0184] Embodiment 29. The method of any one of embodiments 20-28, wherein the sequencing data comes from two or more biological samples.

[0185] Embodiment 30. The method of embodiment 29, wherein the two or more biological samples comprises a control sample.

[0186] Embodiment 31. The method of embodiment 30, wherein the control sample comprises a cfDNA from a healthy patient.

[0187] Embodiment 32. The method any one of embodiments 20-31, wherein the sequencing data comprises a benchmarking sample.

[0188] Embodiment 33. The method of embodiment 32, wherein the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

[0189] Embodiment 34. The method of any one of embodiments 20-33, wherein the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

[0190] Embodiment 35. A method for improving tumor fraction (TF) estimation in a sample by identifying and excluding a clonal hematopoiesis of indeterminate potential (CHIP) variant from the TF estimation, the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining a current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; determining, based on the allele fraction, if an identified somatic variant derives from CHIP, wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant is identified as deriving from CHIP and is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more exclude somatic variants identified as deriving from CHIP.

[0191] Embodiment 36. The method of embodiment 35, wherein determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

[0192] Embodiment 37. The method of embodiment 35 or 36, wherein assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate.

[0193] Embodiment 38. The method of embodiment 37, wherein determining the significance metric comprises calculating a p-value.

[0194] Embodiment 39. The method of embodiment 38, wherein calculating the p-value comprises using a binomial distribution.

[0195] Embodiment 40. The method of embodiment 38 or 39, wherein an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p- value < le-6.

[0196] Embodiment 41. The method of any one of embodiments 35-40, further comprising: iterating through the steps of embodiment 35 until all or a portion of the identified somatic variants having a shedding rate that significantly deviates from the current shedding rate are excluded from the TF estimation.

[0197] Embodiment 42. The method of any one of embodiments 35-41, further comprising filtering and excluding germline variants from the identified somatic variants.

[0198] Embodiment 43. The method of any one of embodiments 35-42, wherein the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

[0199] Embodiment 44. The method of any one of embodiments 35-43, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

[0200] Embodiment 45. The method of embodiment 44, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

[0201] Embodiment 46. The method of any one of embodiments 35-45, wherein the sequencing data comprises from two or more biological samples.

[0202] Embodiment 47. The method of embodiment 46, wherein the sequencing data comprises a control sample.

[0203] Embodiment 48. The method of embodiment 47, wherein the control sample comprises a cfDNA from a healthy patient.

[0204] Embodiment 49. The method any one of embodiments 35-48, wherein the sequencing data comes from two or more biological samples where at least one of the biological samples is a benchmarking sample.

[0205] Embodiment 50. The method of embodiment 49, wherein the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

[0206] Embodiment 51. The method of any one of embodiments 35-50, wherein the somatic variants are, or are associated with previously identified somatic variants.

[0207] Embodiment 52. The method of any one of embodiments 35-51, wherein the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

[0208] Embodiment 53. The method of any one of embodiments 35-52, wherein the somatic variants comprise tumor specific nucleotide variants.

[0209] Embodiment 54. A non-transitory computer-readable medium storing one or more programs, the one or more programs including instructions which, when executed by an electronic device including a processor, cause the device to perform any of the methods of the preceding embodiments.

[0210] Embodiment 55. An electronic device, comprising: one or more processors; memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods of the preceding embodiments.VI. INCORPORATION BY REFERENCE

[0211] All publications, patents, patent applications and other documents cited in this application are hereby incorporated by reference in their entireties for all purposes to the same extent as if each individual publication, patent, patent application or other document were individually indicated to be incorporated by reference for all purposes. Specifically, U.S. Provisional Application No. 63 / 535,450, filed August 30, 2024, is incorporated by reference in its entiretyVII. ADDITIONAL CONSIDERATIONS

[0212] The foregoing description of the embodiments of the invention has been presented for the purpose of illustration; it is not intended to be exhaustive or to limit theinvention to the precise forms disclosed. Persons skilled in the relevant art can appreciate that many modifications and variations are possible in light of the above disclosure.

[0213] Some portions of this description describe the embodiments of the invention in terms of algorithms and symbolic representations of operations on information. These algorithmic descriptions and representations are commonly used by those skilled in the data processing arts to convey the substance of their work effectively to others skilled in the art. These operations, while described functionally, computationally, or logically, are understood to be implemented by computer programs or equivalent electrical circuits, microcode, or the like. Furthermore, it has also proven convenient at times, to refer to these arrangements of operations as modules, without loss of generality. The described operations and their associated modules can be embodied in software, firmware, hardware, or any combinations thereof.

[0214] Any of the steps, operations, or processes described herein can be performed or implemented with one or more hardware or software modules, alone or in combination with other devices. In one embodiment, a software module is implemented with a computer program product including a computer-readable non-transitory medium containing computer program code, which can be executed by a computer processor for performing any or all of the steps, operations, or processes described.

[0215] Embodiments of the invention can also relate to a product that is produced by a computing process described herein. Such a product can include information resulting from a computing process, where the information is stored on a non-transitory, tangible computer readable storage medium and can include any embodiment of a computer program product or other data combination described herein.

[0216] Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application based hereon. Accordingly, the disclosure of the embodiments of the invention is intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims.

Claims

WHAT IS CLAIMED IS:

1. A method for improving tumor fraction (TF) estimation in a sample by identifying and excluding a somatic variant comprising an outlier allele fraction from the TF estimation, the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining a current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction, wherein if the allele fraction for the identified somatic variant significantly deviates from the current shedding rate the identified somatic variant is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more excluded somatic variants.

2. The method of claim 1, wherein determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

3. The method of claim 1 or 2, wherein assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate.

4. The method of claim 3, wherein determining the significance metric comprises calculating a p-value.

5. The method of claim 4, wherein calculating the p-value comprises using a binomial distribution.

6. The method of claim 4 or 5, wherein an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

7. The method of any one of claims 1-6, further comprising: iterating through the steps of claim 1 until all or a portion of the identified somatic variants having a shedding rate that significantly deviate from the current shedding rate are excluded from the TF estimation.

8. The method of any one of claims 1-7, further comprising filtering and excluding germline variants from the identified somatic variants.

9. The method of any one of claims 1-8, wherein the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

10. The method of any one of claims 1-9, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

11. The method of claim 10, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

12. The method of any one of claims 1-11, wherein the sequencing data comprises from two or more biological samples.

13. The method of claim 12, wherein the sequencing data comprises a control sample.

14. The method of claim 13, wherein the control sample comprises a cfDNA from a healthy patient.

15. The method any one of claims 1-14, wherein the sequencing data comes from two or more biological samples where at least one of the biological samples is a benchmarking sample.

16. The method of claim 15, wherein the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

17. The method of any one of claims 1-16, wherein the somatic variants are, or are associated with, previously identified somatic variants.

18. The method of any one of claims 1-17, wherein the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

19. The method of any one of claims 1-18, wherein the tumor specific nucleotide variants are associated with previously tumor specific somatic variants.

20. A method for detecting clonal hematopoiesis of indeterminate potential (CHIP) in a sample comprising cell-free DNA (cfDNA), the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining a current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; and determining, based on the allele fraction, if an identified somatic variant derives from CHIP,wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant the somatic variant is identified as deriving from CHIP.

21. The method of claim 20, wherein determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

22. The method of claim 20 or 21, wherein assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate.

23. The method of claim 22, wherein determining the significance metric comprises calculating a p-value.

24. The method of claim 23, wherein calculating the p-value comprises using a binomial distribution.

25. The method of claim 23 or 24, wherein an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

26. The method of any one of claims 20-25, wherein the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

27. The method of any one of claims 20-26, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

28. The method of claim 27, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

29. The method of any one of claims 20-28, wherein the sequencing data comes from two or more biological samples.

30. The method of claim 29, wherein the two or more biological samples comprises a control sample.

31. The method of claim 30, wherein the control sample comprises a cfDNA from a healthy patient.

32. The method any one of claims 20-31, wherein the sequencing data comprises a benchmarking sample.

33. The method of claim 32, wherein the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

34. The method of any one of claims 20-33, wherein the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

35. A method for improving tumor fraction (TF) estimation in a sample by identifying and excluding a clonal hematopoiesis of indeterminate potential (CHIP) variant from the TF estimation, the method comprising: retrieving sequencing data; identifying somatic variants in the sequencing data; determining a current shedding rate for all the identified somatic variants; assessing, for each identified somatic variant, an allele fraction; determining, based on the allele fraction, if an identified somatic variant derives from CHIP, wherein if the allele fraction for an identified somatic variant significantly deviates from the current shedding rate, the identified somatic variant is identified as deriving from CHIP and is excluded from a TF estimation; and performing the TF estimation based on the identified somatic variants without relying on the one or more exclude somatic variants identified as deriving from CHIP.

36. The method of claim 35, wherein determining the current shedding rate comprises determining p = k / N, wherein k = # of alternate allele observations across all variants for this individual and N = # of total observations across all variants for this individual.

37. The method of claim 35 or 36, wherein assessing the allele fraction for each identified somatic variant comprises determining a significance metric compared to the current shedding rate.

38. The method of claim 37, wherein determining the significance metric comprises calculating a p-value.

39. The method of claim 38, wherein calculating the p-value comprises using a binomial distribution.

40. The method of claim 38 or 39, wherein an identified somatic variant significantly deviates from the current shedding rate when it is assigned a p-value < le-6.

41. The method of any one of claims 35-40, further comprising: iterating through the steps of claim 35 until all or a portion of the identified somatic variants having a shedding rate that significantly deviates from the current shedding rate are excluded from the TF estimation.

42. The method of any one of claims 35-41, further comprising filtering and excluding germline variants from the identified somatic variants.

43. The method of any one of claims 35-42, wherein the sequencing data is obtained from sequencing cell-free DNA existing in biological samples obtained from a plurality of individuals.

44. The method of any one of claims 35-43, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least IX, 5X, 10X, 20X, 30X, 40X, 50X, 60X, 70X, 80X, 90X, or 100X.

45. The method of claim 44, wherein the sequencing data is generated from a biological sample sequenced at a depth of at least 3 OX.

46. The method of any one of claims 35-45, wherein the sequencing data comprises from two or more biological samples.

47. The method of claim 46, wherein the sequencing data comprises a control sample.

48. The method of claim 47, wherein the control sample comprises a cfDNA from a healthy patient.

49. The method any one of claims 35-48, wherein the sequencing data comes from two or more biological samples where at least one of the biological samples is a benchmarking sample.

50. The method of claim 49, wherein the benchmarking sample comprises variants having an allele fraction that significantly deviates from the current shedding rate.

51. The method of any one of claims 35-50, wherein the somatic variants are, or are associated with previously identified somatic variants.

52. The method of any one of claims 35-51, wherein the somatic variants comprise clonal hematopoiesis of indeterminate potential (CHIP) variants.

53. The method of any one of claims 35-52, wherein the somatic variants comprise tumor specific nucleotide variants.

54. A non-transitory computer-readable medium storing one or more programs, the one or more programs including instructions which, when executed by an electronic device including a processor, cause the device to perform any of the methods of the preceding claims.

55. An electronic device, comprising: one or more processors; memory; andone or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods of the preceding claims.

Images