Molecular counting of methylated cell-free DNA for therapeutic monitoring

Abstract

Aspects of the present disclosure include methods of quantifying methylation in a DNA sample, where the methods include processing the sample to encode for the presence or absence of DNA methylation, adding a synthetic molecule (e.g., a synthetic molecule) to the sample, the method includes generating a set of methylated molecules (e.g., a mass control template (QCT) molecule), generating a co-amplification mixture, sequencing the co-amplification mixture, and determining the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of and priority to U.S. Provisional Application No. 63 / 322,014, filed on March 21, 2022, and U.S. Provisional Application No. 63 / 439,492, filed on January 17, 2023, each of which is hereby incorporated by reference in its entirety.

[0003] Sequence Listing

[0004] This application contains a sequence listing with 2 sequences, which has been submitted by the Patent Center and is hereby incorporated by reference in its entirety. The XML copy was created on March 18, 2023, named 38227-55032-SequenceListing-WO.xml, and is 2,905 bytes in size. Background Art

[0005] While new cancer treatments continue to be developed at an unprecedented pace, evaluating whether a cancer treatment is working for a particular patient remains relatively cumbersome and qualitative. Imaging is the gold standard for measuring cancer status; however, PET / CT and CT scans give patients harmful doses of radiation, and these imaging facilities may be located far from where the patient lives. Imaging sessions are scheduled several months apart, which limits oncologists' ability to quickly understand how the cancer is responding to treatment. Qualitative measures such as physical exams and patient-reported symptoms are often confounded by treatment side effects and are therefore insufficient to accurately assess treatment efficacy. However, oncologists want to assess how well a patient's cancer is responding to treatment to inform treatment planning, such as whether the patient should continue with current therapy or switch to a new treatment plan.

[0006] Non-invasive liquid biopsies that measure cell-free DNA (cfDNA) have been developed to quantify the level of circulating tumor DNA (ctDNA) from blood samples. Several studies have found that ctDNA levels are useful and quantifiable information for longitudinal ctDNA measurements that accurately track tumor progression. Many assays use variant allele fractions (VAFs) to specifically quantify the abundance of somatic mutations (e.g., single nucleotide variants, copy number changes, insertions, deletions) and track these VAFs over time. However, assays that rely on quantifying somatic mutations have limitations. Since VAFs can be very small (~0.1-0.5%), there may be significant and unavoidable molecular sampling limitations. Variants that exist with low VAFs may present false negatives; a sample with an average of 2 variant molecules may actually have only 0 variant molecules 13.5% of the time due to Poisson sampling. Even if a variant is detected, the noise of the measurement is high due to the small number of variant molecules. For example, due to sampling noise, the standard deviation of 9 variant molecules in a sample is 3 variant molecules, resulting in a coefficient of variation (CV) of 33%. Small numbers of molecules can make VAF-based time series measurements noisy and difficult to interpret. However, understanding the extent of molecular sampling restriction requires an estimate of the number of variant molecules, information that many assays do not provide.

[0007] Therefore, absolute quantification is required to determine the number of variant molecules present. Digital droplet PCR has been shown to be a sensitive method for detecting ctDNA. However, due to the limited amount of initial sample, only a few genomic locations can be probed simultaneously.

[0008] Given the molecular sampling challenges associated with measuring somatic mutations, methylation has been explored as an alternative biomarker for ctDNA. Methylation has long been shown to be a powerful, consistent, and genomically widespread cancer biomarker. However, although methylation signals are far more abundant than signals from somatic mutations, it is a challenge to accurately and precisely quantify the amount of methylation. Methylation-specific qPCR is a commonly used method; however, due to the exponential nature of the assay, qPCR Ct measurements have high CVs (usually in the teens). Given the limited sample size, it is challenging to improve accuracy with multiple replicates. In addition, non-tumor cfDNA may provide background methylation signals, complicating the task of quantifying signals belonging to ctDNA. In addition, sampling multiple locations to measure methylation will help improve assay performance, but using qPCR to interrogate multiple loci is difficult and time-consuming. Summary of the invention

[0009] A methylation-based method is provided herein to more accurately and precisely assess treatment monitoring of, for example, cancer patients. The methylation-based method is a pan-cancer assay that provides a longitudinal assay of cancer burden via a methylation assessment. The assessment can identify multiple cancer methylation signatures based in part on methylated nucleotides (DNA) in a specific location in the cancer.

[0010] Accurate counting of methylated DNA molecules in a blood sample can provide information about the status of a patient's cancer. Because some methylated molecules are scarce (i.e., present at very low concentrations in the sample), they must be amplified to be detected. However, amplification can complicate the challenge of detecting methylated molecules because it can be noisy and methylated molecules amplify at different rates. Herein, the applicants have developed a methylation-based method that adds a quantitative counting template (QCT) prior to amplification, which provides a measurement tool that allows quantification after this step to determine how many methylated molecules were originally in the sample.

[0011] The methylation-based approach differs from other assays that rely on single nucleotide variation (SNV)-based ctDNA monitoring because methylated ctDNA is a global and additive marker that allows for a more robust and cumulative measurement of ctDNA. Typical SNV-based ctDNA monitoring assays such as tumor-informed MRD assays look at an average of 9 SNVs, while the approach described herein can typically quantify an average of 90 methylated loci; a 10-fold increase in signal.

[0012] The methylation-based method described is a next-generation sequencing (NGS)-based test designed to measure changes in methylated tumor molecules in cancer patients through a blood draw. In particular, the method quantifies methylated ctDNA (circulating tumor DNA) molecules isolated from cell-free DNA (cfDNA) at loci known to be hypermethylated in tumors compared to healthy tissue.

[0013] In one embodiment, plasma and buffy coat are separated from whole blood collected from the patient. Cell-free DNA (cfDNA) is extracted from plasma, and DNA (e.g., genomic DNA) is extracted from buffy coat. QCT (e.g., as described in U.S. Patent Publication Nos. 2020 / 0040380A1, 2019 / 0095577A1, 2019 / 0114389A1, and 2019 / 0211395A1, which are incorporated by reference) is used to quantify the number of methylated molecules in cfDNA and DNA (from buffy coat) at positions known to be highly methylated in cancer compared to non-cancerous tissues and blood in the genome> 500. Methylation (e.g., methylated DNA in buffy coat) measured in DNA is subtracted from cfDNA methylation (e.g., methylated molecules from plasma) to remove background from ctDNA signals. The remaining cfDNA methylated molecules at all highly methylated positions are added to quantify DNA methylation in the sample (e.g., calculating tumor methylation scores).

[0014] The quantified DNA methylation from the currently collected sample (e.g., the tumor methylation score in the sample) can be compared with the most recently reported quantified DNA methylation (e.g., the tumor methylation score) to determine an increase, decrease, or no change call. If the change in the quantified DNA methylation reaches a significance threshold, an increase or decrease can be reported. For baseline tests without any previous collection, no interpretive call (e.g., increase or decrease) is made for the change in quantified DNA methylation. The limit of detection (LOD) can detect a change of 0.2 percentage points in the tumor score with a dispersion of 3 standard deviations.

[0015] The present disclosure relates to a method for quantifying the number of methylated molecules present in a sample using a quantitative counting template (QCT). For example, a sample containing a DNA sequence (e.g., a mixture of methylated and unmethylated DNA sequences) is treated to convert unmethylated cytosine in the DNA sequence into uracil (i.e., treated to encode the presence or absence of DNA methylation in the DNA sequence). In some cases, DNA sequences containing methylated cytosine are sparsely present in the sample, and amplification is then required to detect and quantify these DNA sequences. As described above, amplification will complicate the ability to quantify methylated molecules. The addition and co-amplification of QCT will reduce these complexities. Therefore, QCT is added to the sample and co-amplified with the treated DNA sequence to produce a co-amplified mixture. The co-amplified mixture is then sequenced. Sequencing and subsequent analysis determine the number of QCT molecules, the number of methylated cytosines, and the number of unmethylated cytosines. Alternatively, sequencing and subsequent analysis determine the number of QCT molecules and the number of methylated sequencing reads (i.e., sequencing reads containing uracil). The number of methylated molecules in the sample is quantified based on the number of methylated reads and the number of reads from the QCT molecules. The quantified methylated molecules can then be used to facilitate diagnosis, treatment, or further evaluation of the subject.

[0016] This methylation-based approach can also be used to determine whether different assays performed on the same sample are reliable. For example, when two assays are performed on a sample: (1) an assay to determine the abundance of somatic mutations (e.g., a variant allele fraction (VAF) measurement) and (2) an assay to quantify DNA methylation using a QCT molecule as described herein, the latter can be used to provide information about the reliability of the former. In particular, quantifying the number of methylated molecules (based on the QCT molecule) as an indicator of the presence of cancer in a sample can be used to determine whether the variant allele fraction measurement of the sample (another indicator of the presence of cancer) is reliable. This is possible because the QCT molecule in the methylation assay provides an accurate and reliable way to determine the number of methylated molecules in a sample, thereby avoiding false negatives or false positives. In practice, this means that when a VAF measurement and methylation analysis are performed on material from the same sample, the accuracy and reliability of the methylation measurement can be attributed to the VAF measurement to give a VAF reliability score (e.g., a "true call" (i.e., when the VAF should be trusted) or a "no-call" (i.e., when the VAF should not be trusted)). For example, when the VAF is small (~0.1-0.5%), this can be reported as "negative" for the presence of cancer. If the corresponding quantification of methylated DNA (using the methods described herein) indicates the presence of cancer (e.g., the number of methylated DNA sequences is above a predetermined threshold), the VAF is a "false negative" or "no call", so the VAF measurement should not be trusted. Importantly, these results can indicate the need for additional evaluation of the subject, for example, new treatment options, genomic analysis, and / or a change in treatment regimen.

[0017] This methylation-based method can also be used to determine changes in the methylation profile over time in a subject. Changes in the methylation profile in a subject may indicate the presence of cancer or changes in an already existing cancer. The appearance of new methylation loci at subsequent time points, the increased contribution of a methylation locus to the overall methylation signal at subsequent time points, or a combination thereof is an indication of changes in the methylation profile in a subject. It is worth noting that, due to the use of QCT molecules, and the ability of QCT molecules to reliably and accurately quantify the number of methylated molecules at each time point, the methylation-based method described herein is able to accurately determine the methylation profile. This feature allows comparisons to be made between time points, so that the methylation profile in a subject can be determined over time. Other ctDNA assays lack accuracy and reliability between measurements taken at different time points, so these assays cannot measure the methylation profile in a subject over time. Therefore, the methylation-based method described herein has the unique ability to measure changes in the methylation profile, which can then provide information for the subject's oncologist to develop strategies for diagnosis, treatment, and additional evaluation.

[0018] In general, the present disclosure relates to a method for quantifying the number of methylated molecules present in a sample in a multiplexed manner using a quantitative counting template (QCT). Background subtraction using the methylation signal from the corresponding buffy coat of the sample will reduce noise in the assay and improve robustness against genomic DNA (gDNA) contamination. Background levels of methylation in healthy subjects vary from day to day, demonstrating the importance of background subtraction for these time series measurements. This method can be adapted to measure methylation signals in multiple cancer types using a single assay chemistry.

[0019] In one aspect, the present disclosure relates to a method for quantifying DNA methylation in a sample containing a DNA sequence, the method comprising: treating the sample to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match the target sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and the collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplification mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; and quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules.

[0020] In some embodiments, the sample is obtained from a blood draw comprising plasma and buffy coat, wherein the plasma comprises cfDNA and the buffy coat comprises genomic DNA (gDNA) sequences.

[0021] In some embodiments, the method further comprises extracting cfDNA from plasma from the sample and extracting gDNA sequences from the buffy coat from the sample prior to processing the sample to encode the presence or absence of DNA methylation.

[0022] In some embodiments, the method further comprises quantifying DNA methylation in a sample containing gDNA sequences from the buffy coat.

[0023] In some embodiments, the method includes: processing a sample containing a DNA sequence from a buffy coat to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match the target sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and the collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplified mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; and determining the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules, thereby quantifying DNA methylation in the gDNA sequence from the buffy coat.

[0024] In some embodiments, the method further comprises adding a spike-in of known sequence and amount to the sample prior to the processing step. In some embodiments, wherein processing the sample to encode the presence or absence of DNA methylation comprises bisulfite conversion or enzymatic conversion. In some embodiments, wherein the collection of synthetic molecules is a collection of quantitative counting templates (QCT). In some embodiments, each of the at least ten target loci is selected based on an expected increase in DNA methylation at each locus in cancer tissue compared to non-cancerous tissue.

[0025] In some embodiments, at least 100 target loci are amplified or at least 500 target loci are amplified.

[0026] In some embodiments, the co-amplified mixture is sequenced at a read depth of at least 1 read per molecule.

[0027] In some embodiments, the co-amplified mixture is sequenced at a read depth of 10 or more reads per molecule, 100 or more reads per molecule, or 1000 or more reads per molecule.

[0028] In some embodiments, the method further comprises aggregating the number of methylated molecules of at least two of the at least ten target loci to quantify DNA methylation in the sample.

[0029] In some embodiments, aggregating includes aggregating target loci of the at least ten target loci that contain less than a threshold amount of methylated molecules in the buffy coat sample.

[0030] In some embodiments, the sample further comprises at least one normalizing locus predicted to be hypermethylated in cfDNA in both cancerous and non-cancerous tissues.

[0031] In some embodiments, the co-amplification mixture comprises a collection of amplified synthetic molecules, a collection of amplified all or a subset of at least 10 target loci, and amplified at least one normalizing locus.

[0032] In some embodiments, the method further comprises normalizing the aggregated number of methylated molecules from at least two of the at least 10 target loci by the methylated molecules of at least one normalizing locus.

[0033] In some embodiments, the method further comprises subtracting background methylation.

[0034] In some embodiments, the method further comprises filtering out selected hypermethylated target loci.

[0035] In some embodiments, the method further comprises subtracting background methylation and filtering out selected hypermethylated loci.

[0036] In some embodiments, subtracting background methylation comprises subtracting the number of methylated molecules as measured in the buffy coat from the number of methylated molecules in the cfDNA.

[0037] In some embodiments, subtracting background methylation comprises subtracting the number of methylated molecules as measured in the buffy coat from the number of methylated molecules in the cfDNA on a per locus basis.

[0038] In some embodiments, filtering out selected hypermethylated target loci comprises filtering out target loci having hypermethylated cfDNA signals in non-cancerous tissues.

[0039] In some embodiments, filtering out selected hypermethylated target loci includes filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold. In some embodiments, the threshold is sample-specific. In some embodiments, the threshold is subject-specific. In some embodiments, the threshold comprises a predetermined average tumor methylation quantitative equivalent (QE), a predetermined maximum tumor methylation QE, or a combination thereof. In some embodiments, the threshold comprises a predetermined average tumor methylation QE.

[0040] In some embodiments, filtering out selected hypermethylated target loci is performed prior to determining the number of methylated molecules in the sample.

[0041] In some embodiments, subtracting background methylation is performed prior to determining the number of methylated molecules in a sample.

[0042] In some embodiments, filtering out hypermethylated loci and subtracting background methylation is performed prior to quantifying the number of methylated molecules in a sample.

[0043] In another aspect, the present disclosure relates to a method for determining a DNA methylation profile in a subject, the method comprising: processing a sample isolated from the subject to encode the presence or absence of DNA methylation in a DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match the target sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and an amplified collection of at least ten target loci from the DNA sequence; sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of methylated sequence reads in the sample and the number of reads from the collection of synthetic molecules; and determining the methylation profile in the subject based on the number of methylated molecules in the sample.

[0044] In another aspect, the present disclosure relates to a method for determining a DNA methylation profile in a subject over time, the method comprising: at a first time point: i) treating a sample isolated from the subject to encode the presence or absence of DNA methylation in a DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; ii) adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match the target sequence region of the endogenous target molecule; iii) generating a co-amplification mixture comprising the amplified collection of synthetic molecules and the amplified collection of at least ten target loci from the DNA sequence; i v ) sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; v) determining the number of sequence reads that are methylated sequence reads; vi) quantifying the number of methylated molecules in the sample based on the number of methylated sequence reads in the sample and the number of reads from the set of synthetic molecules; repeating steps i) to vi) at a second time point; and determining a DNA methylation profile in the subject based on the number of methylated molecules in the sample at the first time point and the number of methylated molecules in the sample at the second time point.

[0045] In some embodiments, the number of methylated molecules from each target locus in the sample is quantified based on the number of methylated sequence reads from each target locus in the sample and the number of reads from the collection of synthetic molecules.

[0046] In some embodiments, determining the methylation profile in a subject at a first time point and a second time point will identify a change in the methylation profile between the first time point and the second time point.

[0047] In some embodiments, the method further comprises: repeating steps i) to vi) at a third time point; and determining the methylation profile in the subject based on the number of methylated molecules for each target locus in the sample at the third time point.

[0048] In some embodiments, determining the methylation profile in a subject at a first time point, a second time point, and a third time point will identify changes in the methylation profile between the first and second time points, between the second and third time points, between the first and third time points, or a combination thereof.

[0049] In some embodiments, a change in the methylation profile indicates a change in a tumor in a subject. In some embodiments, the change in the tumor includes a change in tumor size, a change in the abundance of somatic mutations associated with the tumor, the presence of new somatic mutations associated with the tumor, a chromosomal abnormality associated with the tumor, or the tumor is resistant to treatment. In some embodiments, the change in the methylation profile is incorporated into the clinical recommendation for the subject.

[0050] In some embodiments, the method further comprises assigning a measure of: increase, decrease, or no change to the change in the methylation profile based on the comparison to a significance threshold. In some embodiments, the significance threshold is predetermined or dynamically calculated.

[0051] In some embodiments, the sample is obtained from a blood draw that contains plasma and buffy coat, wherein the plasma contains cfDNA and the buffy coat contains genomic DNA (gDNA) sequences.

[0052] In some embodiments, the method further comprises extracting cfDNA from plasma from the sample and extracting gDNA sequences from the buffy coat from the sample prior to processing the sample to encode the presence or absence of DNA methylation.

[0053] In some embodiments, the method further comprises quantifying DNA methylation in a sample containing gDNA sequences from the buffy coat.

[0054] In some embodiments, the method includes: processing a sample containing a DNA sequence from a buffy coat layer to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample contains at least ten target loci from the DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target molecule containing at least one of the target loci, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; generating a co-amplification mixture, the co-amplification mixture comprising the collection of amplified synthetic molecules and the collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplified mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules; and aggregating the number of methylated molecules on at least two target loci to quantify DNA methylation in the sample containing a DNA sequence from the buffy coat layer, thereby quantifying DNA methylation in the buffy coat layer.

[0055] In some embodiments, the method further comprises adding a spike of known sequence and amount to the sample prior to the processing step.

[0056] In some embodiments, the collection of synthetic molecules is a collection of quantitative counting templates (QCTs).

[0057] In some embodiments, each locus of the at least ten target loci is selected based on a predicted increase in DNA methylation at the locus in cancer tissue compared to non-cancerous tissue.

[0058] In some embodiments, at least 100 target loci are amplified or at least 500 target loci are amplified.

[0059] In some embodiments, the sample further comprises at least one normalizing locus predicted to be hypermethylated in cfDNA in both cancerous and non-cancerous tissues.

[0060] In some embodiments, the co-amplification mixture comprises a collection of amplified synthetic molecules, a collection of amplified all or a subset of at least 10 target loci, and amplified at least one normalizing locus.

[0061] In some embodiments, the method further comprises aggregating the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in the sample.

[0062] In some embodiments, aggregating includes aggregating target loci of the at least ten target loci that contain less than a threshold amount of methylated molecules in the buffy coat sample.

[0063] In some embodiments, the method further comprises normalizing the aggregated number of methylated molecules from at least two of the at least 10 target loci by the methylated molecules of at least one normalizing locus.

[0064] In some embodiments, the method further comprises subtracting background methylation.

[0065] In some embodiments, the method further comprises filtering the hypermethylated loci.

[0066] In some embodiments, the method further comprises subtracting background methylation and filtering out selected hypermethylated loci.

[0067] In some embodiments, subtracting background methylation comprises subtracting the number of methylated molecules as measured in the buffy coat (e.g., as quantified herein) from the number of methylated molecules in the cfDNA.

[0068] In some embodiments, subtracting background methylation comprises subtracting the number of methylated molecules as measured in the buffy coat (e.g., as quantified herein) from the number of methylated molecules in the cfDNA on a per locus basis.

[0069] In some embodiments, filtering out selected hypermethylated loci, subtracting background methylation, or both is performed prior to quantifying the number of methylated molecules in a sample.

[0070] In some embodiments, filtering out selected hypermethylated target loci comprises filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold. In some embodiments, the threshold is sample specific. In some embodiments, the threshold is subject specific. In some embodiments, the threshold comprises a predetermined average tumor methylation QE, a predetermined maximum tumor methylation QE, or a combination thereof. In some embodiments, the threshold comprises a predetermined average tumor methylation QE.

[0071] In some embodiments, the method further comprises performing a treatment selection assay on the subject after detecting a change in the DNA methylation profile.In some embodiments, the treatment selection assay comprises genomic analysis to detect new somatic mutations, the abundance of somatic mutations, or both.

[0072] In another aspect, the present disclosure relates to a method for determining a confidence indication of the presence or absence of a somatic mutation in a sample containing a DNA sequence, the method comprising: determining the presence or absence of a somatic mutation in the sample; quantifying DNA methylation in the sample, comprising the steps of: treating the sample to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and the amplified a set of at least ten target loci from DNA sequences amplified from the sample; sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules; and returning such a confidence indication: when the number of methylated molecules in the sample is above a predetermined or dynamically calculated threshold, a true determination result for the presence or absence of a somatic mutation is returned; or when the number of methylated molecules in the sample is at or below a predetermined or dynamically calculated threshold, a no determination result for the presence or absence of a somatic mutation is returned, thereby a true determination identifies the confidence of the determination for the presence or absence of a somatic mutation.

[0073] In some embodiments, determining the presence or absence of a somatic mutation in a sample comprises determining a variant allele fraction of the somatic mutation.

[0074] In some embodiments, the sample is obtained from a blood draw that contains plasma and buffy coat, wherein the plasma contains cfDNA and the buffy coat contains genomic DNA (gDNA) sequences.

[0075] In some embodiments, the method further comprises extracting cfDNA from plasma from the sample and extracting gDNA sequences from the buffy coat from the sample prior to processing the sample to encode the presence or absence of DNA methylation.

[0076] In some embodiments, the method further comprises quantifying DNA methylation in a sample containing DNA sequences from the buffy coat.

[0077] In some embodiments, the method includes: processing a sample containing a DNA sequence from a buffy coat layer to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match the target sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and the collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplified mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; and quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules, thereby quantifying DNA methylation in the gDNA sequence from the buffy coat layer.

[0078] In some embodiments, the method further comprises adding a spike of known sequence and amount to the sample prior to the processing step.

[0079] In some embodiments, the collection of synthetic molecules is a collection of quantitative counting templates (QCTs).

[0080] In some embodiments, each locus of the at least ten target loci is selected based on a predicted increase in DNA methylation at each locus in cancer tissue compared to non-cancerous tissue.

[0081] In some embodiments, at least 100 target loci are amplified or at least 500 target loci are amplified.

[0082] In some embodiments, the sample further comprises at least one normalizing locus predicted to be hypermethylated in cfDNA in both cancerous and non-cancerous tissues.

[0083] In some embodiments, the co-amplification mixture comprises a collection of amplified synthetic molecules, a collection of amplified all or a subset of at least 10 target loci, and amplified at least one normalizing locus.

[0084] In some embodiments, the method further comprises aggregating the number of methylated molecules of at least two of the at least ten target loci to quantify DNA methylation in the sample.

[0085] In some embodiments, aggregating includes aggregating target loci of the at least ten target loci that contain less than a threshold amount of methylated molecules in the buffy coat sample.

[0086] In some embodiments, the method further comprises normalizing the aggregated number of methylated molecules from at least two of the at least 10 target loci by the methylated molecules of at least one normalizing locus.

[0087] In some embodiments, the method further comprises subtracting background methylation.

[0088] In some embodiments, the method further comprises filtering out selected hypermethylated loci.

[0089] In some embodiments, the method further comprises subtracting background methylation and filtering out selected hypermethylated loci.

[0090] In some embodiments, subtracting background methylation comprises subtracting the number of methylated molecules as measured in the buffy coat (e.g., as quantified herein) from the number of methylated molecules in the cfDNA.

[0091] In some embodiments, subtracting background methylation comprises subtracting the number of methylated molecules as measured in the buffy coat (e.g., as quantified herein) from the number of methylated molecules in the cfDNA on a per locus basis.

[0092] In some embodiments, filtering out selected hypermethylated loci, subtracting background methylation, or both is performed prior to quantifying the number of methylated molecules in a sample.

[0093] In some embodiments, filtering out selected hypermethylated target loci comprises filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold. In some embodiments, the threshold is sample specific. In some embodiments, the threshold is subject specific. In some embodiments, the threshold comprises a predetermined average tumor methylation QE, a predetermined maximum tumor methylation QE, or a combination thereof. In some embodiments, the threshold comprises a predetermined average tumor methylation QE.

[0094] In some embodiments, the method further comprises, upon returning a true call and in the event that the abundance of somatic mutations indicates the presence of cancer, performing a repeat treatment selection assay on the subject or repeating the treatment selection assay. In some embodiments, the treatment selection assay comprises genomic analysis to detect new somatic mutations, the abundance of somatic mutations, or both.

[0095] In some embodiments, the method further comprises, upon returning a no call, repeating the method described herein on a different sample taken from the subject.

[0096] In some embodiments, the method further comprises incorporating the confidence indication into a clinical recommendation to the subject. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] Figure 1A A flow chart illustrating one non-limiting embodiment of a method for quantifying DNA methylation in a sample.

[0098] Figure 1B A flow chart illustrating one non-limiting embodiment of a method for quantifying DNA methylation in a sample.

[0099] Figure 2A Data for a lung cancer detection assay are shown. Figure 2A Shown are total methylated molecules measured using the lung cancer assay for technical replicates of lung cancer contrived samples at different tumor fractions (0% tumor, 0.5% tumor, 1% tumor, 2.5% tumor, and 5% tumor). 0.5% prevalent methylation and 5% prevalent methylation were used as controls.

[0100] Figure 2B Shown are data from a lung cancer detection assay in human samples from two different lung cancer subjects. Figure 2B Total (sum) methylation QE at target loci (n=113) are shown. Artificial samples were taken from subjects 18-0901 and 20-0407. Figure 2B Artificial samples were assayed at 0% tumor fraction and 5% tumor fraction using the PCR volumes and thermocycler protocols indicated in .

[0101] Figure 3 against Figure 2A The concordance value for the overall QE is shown for each sample analyzed in .

[0102] Figure 4 against Figure 3 Each sample analyzed in shows methylation in a lung cancer artifact sample, with masking and background subtraction performed. Figure 4 Total normalized methylated molecules (y-axis) for lung cancer artificial samples and controls of different tumor fractions (0% tumor, 0.5% tumor, 1% tumor, 2.5% tumor, and 5% tumor) are shown, with 0.5% prevalent methylation and 5% prevalent methylation as controls. The data shown are background-subtracted and target loci with high background methylation are masked.

[0103] Figure 5Data from simulations of cfDNA tumor signals quantified using somatic mutation and methylation methods are shown. The top panel shows the variation of variant allele fractions (y-axis) quantified using somatic mutations with gDNA contamination multiples (x-axis). The middle panel shows the variation of methylated molecules (y-axis) quantified using methylation without buffy coat background methylation with gDNA contamination multiples (x-axis). The bottom panel shows the variation of methylated molecules (y-axis) quantified using methylation with buffy coat background methylation with gDNA contamination multiples (x-axis).

[0104] Figure 6 Methylation profiles of buffy coat isolated from different tubes, different tube types, different days, and different subjects from the same blood draw (i.e., 5000 genome equivalents (ge) of buffy coat) are shown. The data are shown as methylated molecules (y-axis) of different buffy coats (x-axis).

[0105] Figure 7 Hierarchical clustering of methylation profiles of buffy coat isolated from different tubes of the same blood draw, different tube types, different days, and different subjects (ie, 5000 ge of buffy coat) is shown. Cluster distances were calculated using the L1 norm.

[0106] Figure 8 Data showing the performance of the Pan-Cancer assay in human samples. Data are presented as normalized methylation QE (y-axis) for each human sample from different cancer types. Human samples were prepared at 0% tumor fraction and 5% tumor fraction.

[0107] Figures 9A-9F Data for a non-limiting example of loci used in the "Personalized Blacklist" are shown. Data are presented as normalized tumor QE (y-axis: "qe_norm") for each of the 12 samples indicated on the x-axis in cfDNA and buffy coat. Fig. 9A Tumor QE (normalized) of intron cg03134157_77 is shown. Fig. 9B Tumor QE (normalized) of introncg20907051_11 is shown. Fig. 9C Tumor QE (normalized) of intron_cg12880300_0 is shown. Fig.9D Tumor QE (normalized) for locus IRX4_cg13974394_0 is shown. Fig.9E Tumor QE (normalized) for locus EMID2_cg25290307_0 is shown. Fig.9F Tumor QE (normalized) of intron_cg11453719_0 is shown.

[0108] Fig.10Normalized total tumor molecules (y-axis) of the samples are shown on the x-axis based on the global blacklist thresholds described herein (e.g., (1) maximum tumor QE in any non-cancer sample>15; and (2) mean tumor QE (including 0s)>2).

[0109] Fig.11 The normalized total tumor molecules of the new blacklist (y-axis) are compared with the normalized total tumor molecules of the old blacklist. Each blacklist was applied to Fig.11 The methylation data of the 30 subjects listed in are plotted in the figure.

[0110] Figures 12A-12D Data are shown for 2 subjects (out of 30): Subject 8272 and Subject 6885, who had slightly different methylation results after applying the new blacklist. Data are presented as normalized total tumor molecules (y-axis) at day 0, day 63, and day 154 since first acquisition. Fig. 12A Data using the old blacklist for subject 8272 is shown. Fig. 12B Data using the new blacklist for subject 8272 is shown. Fig. 12C Data using the old blacklist for subject 6885 is shown. Fig.12D Data using the new blacklist for subject 6885 is shown.

[0111] Figures 13A-13E The five subjects (i.e., BTO-1 ( Fig.13A )、BTO-2( Fig. 13B )、BTO-3( Fig. 13C )、BTO-4( Fig.13D ) and BTO-5( Fig.13E ))’s data.

[0112] Fig.14 Shown are the response scores measured in triplicate at each tumor score for each tumor type.

[0113] Fig.15 Shown are the response scores measured in triplicate at each tumor score for each tumor type.

[0114] Fig.16 Response scores measured for 8 cancer subjects are shown, with samples from each subject from two time points. The 8 cancer subjects included three cancer types (lung cancer = 4 subjects, pancreatic cancer = 3 subjects, endometrial cancer = 1 subject). The colors in the legend indicate the decisions made based on changes in the response scores.

[0115] Fig.17The response scores of 12 healthy subjects are shown, and samples from each subject are from two time points. The cfDNA sample of subject HV006 at time point 1 did not meet the sample QC standard, so no judgment was made for this subject. No change or indeterminate judgments were made for the remaining 11 subjects.

[0116] Fig.18 The sensitivity of each tumor fraction plotted with 95% confidence intervals is shown. The dotted line indicates the 95% sensitivity, the limit of the detection sensitivity threshold. For each tumor fraction, the sensitivity was calculated by comparing each of the 16 replicates with each of the 16 0% tumor fraction replicates (a total of 256 comparisons).

[0117] Fig.19 Box plots of the median and interquartile range of response scores at each tumor score are shown. CVs were calculated using the standard deviation and mean for each tumor score.

[0118] Fig. 20 Simulated response scores for 1% and 2% tumor DNA samples are shown.

[0119] Fig.21 Methylation analysis of a 56-year-old male subject with stage IV pancreatic ductal adenocarcinoma at the time of initial collection is shown. The subject was treated with folfirinox ("folfiri"), followed by maintenance therapy with folfiri, and then with folo x . Fig. 22 Tumor QE (y-axis) is shown over days since last treatment (x-axis). The triangles indicate the cancer stage as determined by clinicians using imaging.

[0120] Fig. 22 Methylation analysis of a 50-year-old male subject with pancreatic adenocarcinoma stage IV at the time of first collection is shown. The subject was treated with 5000 mg of fluorouracil + 800 mg of leucovorin + 180 mg of oxaliplatin + 300 mg of irinotecan for 46 hours. Fig.23 Tumor QE (y-axis) is shown over days since last treatment (x-axis). The triangles indicate the cancer stage as determined by clinicians using imaging.

[0121] Fig.23 Methylation analysis of an 87-year-old male subject with colorectal adenocarcinoma stage IV at the time of first collection is shown. The subject received 6 cycles of fluorouracil 1100 mg + leucovorin 1100 mg + bevacizumab 500 mg. Fig.23 Tumor QE (y-axis) is shown over days since last treatment (x-axis). The triangles indicate the cancer stage as determined by clinicians using imaging.

[0122] Fig.24 Methylation analysis of a 56-year-old female subject with stage IV squamous cell carcinoma of the lung at the time of first collection is shown. The subject received 6 docetaxel treatments. Fig.25 Tumor QE (y-axis) is shown in days since last treatment (x-axis). Triangles indicate subjects entered hospice care.

[0123] Fig.25 Shown are the results of a treatment selection assay performed on one plasma sample collected at the same time point as the methylation assay for 40 different samples. Fig.26 Shown is the maximum variant allele fraction (VAF) for each sample (y-axis label: maximum VAF measured during hybrid capture) versus normalized total tumor QE (x-axis).

[0124] Fig.26 The methylation profile of subject 6885 measured at different time points over a period of about 294 days is shown. In particular, Fig.26 The normalized total tumor molecules sorted into the loci are shown by color (y-axis). Methylation profiles were measured at Day 0, Day 63, Day 145, Day 166, Day 229, and Day 294. The numbers in the legend indicate the total methylation percentage for each locus at that particular time point.

[0125] Fig. 27 The methylation profile of subject 5458 measured at different time points over a period of about 250 days is shown. In particular, Fig. 27 The normalized total tumor molecules (y-axis) that were sorted into the loci are shown by color. Methylation profiles were measured at Day 0, Day 35, Day 63, Day 119, Day 201, and Day 250. The numbers in the legend indicate the total methylation percentage for each locus at that particular time point.

[0126] Fig.28 Tumor methylation scores are shown, expressed as the normalized sum of methylated molecules at >500 loci hypermethylated in circulating tumor DNA (ctDNA) at each acquisition date. Days since first acquisition are indicated in parentheses. DETAILED DESCRIPTION

[0127] In one aspect, the present disclosure provides a method for quantifying DNA methylation in a sample containing a DNA sequence. The method may include: treating a sample to encode the presence or absence of DNA methylation in a DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising a target-related region having a nucleotide sequence matching a target sequence region of an endogenous target molecule and a variable region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule; generating a co-amplification mixture, the co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of at least ten target loci from amplified DNA sequences; sequencing the co-amplification mixture with a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying (determining) the absolute number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules; and optionally aggregating the number of molecules on at least two loci.

[0128] like Figure 1A As shown in , an embodiment of a method for quantifying DNA methylation in a sample may include: processing the sample to encode the presence or absence of DNA methylation 110; adding a collection of synthetic molecules (e.g., quality control template (QCT) molecules) to the sample, wherein each synthetic molecule comprises a target-associated region and a variable region) 112; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and the collection of amplified at least ten target loci from a DNA sequence 114; sequencing the co-amplification mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads 116; determining the number of sequence reads that are methylated sequence reads 118; quantifying (determining) the number of methylated molecules in the sample (e.g., an absolute number) based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules 120; and promoting diagnosis based on the amount of methylated molecules 122 and / or promoting treatment of one or more conditions 124.

[0129] In some embodiments, a spike of known sequence and amount is added to the sample prior to processing the sample to encode the presence or absence of DNA methylation.

[0130] In some embodiments, processing the sample to encode the presence or absence of DNA methylation is performed by bisulfite conversion or enzymatic conversion or any other suitable processing method to individually identify methylated DNA molecules.

[0131] In some embodiments, the collection of co-amplified synthetic molecules is a collection of quantitative counting templates (QCTs).

[0132] In some embodiments, the target locus is selected due to increased DNA methylation in cancer tissue compared to normal tissue.

[0133] In some embodiments, at least 100 target loci are amplified.

[0134] In some embodiments, at least 800 target loci are amplified.

[0135] In some embodiments, the co-amplified mixture is sequenced at a read depth of at least 1 read per molecule.

[0136] In some embodiments, the co-amplified mixture is sequenced at a read depth of 10 or more reads per molecule, 100 or more reads per molecule, or 1000 or more reads per molecule.

[0137] In some embodiments, the co-amplified mixture is sequenced at a read depth of at least 1000 sequencing reads per genomic position.

[0138] In some embodiments, the number of molecules at a polymerized locus is also included, wherein the locus is a cancer-specific locus.

[0139] In some embodiments, quantification of DNA methylation is performed in a cell-free DNA sample.

[0140] In some embodiments, cell-free DNA in plasma and cell-free DNA in buffy coat are extracted from the same patient prior to any spiking step and the samples are processed to encode the presence or absence of DNA methylation.

[0141] In some embodiments, further comprising aggregating the number of molecules at a locus in the cell-free DNA sample, wherein the locus is a locus containing less than a threshold amount of methylated molecules in the buffy coat sample.

[0142] In some embodiments, the generated co-amplification mixture comprises a collection of amplified spike molecules, a collection of amplified synthetic molecules, and a collection of amplified at least 10 target loci from a plurality of DNA sequences and at least one locus that exhibits hypermethylation in cell-free DNA.

[0143] In some embodiments, further comprising aggregating the number of methylated molecules at the target loci and at least one locus that exhibits hypermethylation in the cell-free DNA.

[0144] In some embodiments, further comprising normalizing the polymerase number of the target methylated molecule by the polymerase number of the molecule with high methylation in the cell-free DNA.

[0145] In some embodiments, a background subtraction step is also included.

[0146] In some embodiments, a step of filtering hypermethylated loci is further included.

[0147] In some embodiments, a background subtraction step and a step of filtering hypermethylated loci are also included.

[0148] In some embodiments, the background subtraction step comprises subtracting the number of methylated molecules (or methylation signal) as measured in the buffy coat from the number of methylated molecules (or methylation signal) in the cfDNA.

[0149] In some embodiments, further comprising subtracting the number of methylated molecules (or methylation signals) as measured in the buffy coat from the number of methylated molecules (or methylation signals) in the cfDNA on a per locus basis.

[0150] In some embodiments, the step of filtering hypermethylated loci comprises filtering target loci with high background methylation prior to quantifying the absolute number of methylated molecules in the sample.

[0151] In some embodiments, the filtered target loci include target loci for which the number of methylated molecules (or methylation signal) in the buffy coat methylation is above a threshold.

[0152] In some embodiments, the threshold is sample-specific. In some embodiments, the threshold is subject-specific. In such a case, this is referred to as a "personalized" blacklist. In some embodiments in which the threshold is sample-specific or subject-specific, the threshold is determined by evaluating the buffy coat methylation signal of the locus with a high methylation cfDNA signal in a non-cancer subject.

[0153] In some embodiments, the threshold comprises a predetermined mean tumor methylation QE, a predetermined maximum tumor methylation QE, or a combination thereof.

[0154] In some embodiments, the predetermined average tumor methylation QE is equal to or greater than 0.5, equal to or greater than 1.0, equal to or greater than 1.5, equal to or greater than 2.0, equal to or greater than 2.5, equal to or greater than 3.0, equal to or greater than 3.0, equal to or greater than 3.5, equal to or greater than 4.0, equal to or greater than 4.5, or equal to or greater than 5.0.

[0155] In some embodiments, the predetermined maximum tumor methylation QE in any non-cancer sample is equal to or greater than 5, equal to or greater than 10, equal to or greater than 15, equal to or greater than 20, equal to or greater than 25, equal to or greater than 30, equal to or greater than 35, equal to or greater than 40, equal to or greater than 45, or equal to or greater than 50.

[0156] In one embodiment, the thresholds include (1) maximum tumor QE in any non-cancer sample>15; and (2) mean tumor QE (including 0s)>2.

[0157] In some embodiments, the target locus is selected to be differentially methylated in at least two cancer tissues of origin.

[0158] In some embodiments, at least 5 loci are selected as target loci per cancer tissue of origin.

[0159] In some embodiments, the origin of the cancer tissue is determined based on the abundance of methylated molecules at the target loci.

[0160] One aspect of the present disclosure provides a method for quantifying the amount of tumor DNA in a sample containing a DNA sequence, the method comprising: i. adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target molecule, and a variant region having a nucleotide sequence that does not match the sequence region of the endogenous target molecule; ii. generating a co-amplification mixture, the co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of at least 10 amplified target loci from the DNA sequence; iii. sequencing the co-amplification mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; iv. quantifying the absolute number of molecules containing somatic mutations in the sample based on the number of reads from the sample and the number of reads from the collection of synthetic molecules; and v. determining a lower limit of the variant allele fraction based on the total number of molecules present at each locus in the sample before amplification.

[0161] One aspect of the present disclosure provides a method for quantifying methylation in a DNA sample, the method comprising: adding a spike of known sequence to the sample; processing the sample to encode the presence or absence of methylation into the DNA sequence itself; adding a collection of QCT or spike molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; simultaneously amplifying a mixture of the sample and QCT / spike molecules at 10 or more different loci; sequencing the amplified mixture (at a read depth of at least one read per molecule in the sample); determining the number of methylated reads based on the read sequence; calculating the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the QCT / spike at each locus; and aggregating the number of molecules at the loci.

[0162] In some embodiments, the method includes adding a constant amount of a spike to each sample prior to methylation conversion so that the efficiency of the conversion process can be monitored.

[0163] In some embodiments, processing the sample to encode the presence or absence of methylation into the DNA sequence itself includes performing a bisulfite conversion or an enzymatic conversion. Bisulfite conversion is described in U.S. Pat. No. 8,257,950, which is incorporated herein by reference in its entirety. Bisulfite conversion and enzymatic conversion convert unmethylated cytosine to uracil. Therefore, it can be determined whether each read is from a methylated molecule by determining the number of methylated reads based on the read sequence.

[0164] In some embodiments, QCT can be used to quantify each locus. Alternatively, in other embodiments, a spike molecule can be used to quantify each locus.

[0165] In some embodiments, a mixture of sample and QCT / spiking molecules is amplified at 10 or more different loci simultaneously. In some embodiments, these loci are selected based on being more methylated in tumors compared to normal tissue of the same tissue type. In some embodiments, more than 100 loci are amplified. In some embodiments, more than 800 loci are amplified. In some embodiments, these loci have increased methylation in cancer.

[0166] In some embodiments, the method includes sequencing the amplified mixture (e.g., at a read depth of at least one read per molecule in the sample). A read depth qualifier may be optional. In some embodiments, the read depth qualifier is at least 1000 reads per locus (e.g., based on an expectation of 1000-2000 sample molecules per genomic position per tube of blood).

[0167] In some embodiments, treating the sample to encode the presence or absence of methylation into the DNA sequence itself comprises bisulfite conversion or enzymatic conversion.

[0168] In some embodiments, the method includes calculating the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the QCT / spiking at each locus. For the spike, the method includes dividing the number of sample reads by the number of spike reads multiplied by the number of spikes added to the sample.

[0169] In some embodiments, the method includes aggregating the number of molecules at the loci.In certain embodiments, subsets of the loci are summed (e.g., only lung cancer-specific loci are summed).

[0170] One aspect of the present disclosure includes a method for quantifying methylation in a cell-free DNA sample, the method comprising extracting cell-free DNA from plasma and extracting DNA from a buffy coat from the same individual; adding a spike of known sequence to each DNA sample; treating each DNA sample to encode the presence or absence of methylation into the DNA sequence itself; adding a collection of QCT or spike molecules to each DNA sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; simultaneously amplifying a mixture of the sample and QCT / spike molecules at 10 or more different loci; sequencing the amplified mixture (at a read depth of at least one read per molecule in the sample); determining the number of methylated reads based on the read sequence; and computationally determining the number of methylated molecules in the sample based on the number of reads from the sample and the number of reads from the QCT / spike at each locus; and aggregation.

[0171] In some embodiments, the method includes aggregating the number of molecules at the loci in the cell-free DNA sample while excluding loci that have a significant amount of methylated molecules in the buffy coat sample. To determine significance, if any, the method includes determining a threshold number of molecules.

[0172] For example, cfDNA has background methylation signals that vary from patient to patient. Since most cell-free DNA is derived from leukocytes, one aspect of the method can use the methylation signal from the buffy coat as a baseline. In some embodiments, the method includes minimizing the cfDNA background signal by ignoring any loci with significant methylation in the buffy coat (e.g., can be used for minimal residual disease (MRD) detection applications). In other embodiments, the method avoids background subtraction by selecting loci with low methylation background in most people.

[0173] One aspect of the present disclosure includes a method for quantifying methylation in a DNA sample, the method comprising: adding a spike of known sequence to the sample; treating the sample to encode the presence or absence of methylation into the DNA sequence itself; adding a collection of QCT or spike molecules to the sample, the collection of molecules comprising: a) a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and b) a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; simultaneously amplifying a mixture of the sample and QCT / spike molecules at at least 10 different target loci and at at least one locus known to be highly methylated in cfDNA; sequencing the amplified mixture (at a read depth of at least one read per molecule in the sample); determining the number of methylated reads based on the read sequence; computationally determining the number of methylated molecules in the sample based on the number of reads from the sample and the number of reads from the QCT / spike at each locus; aggregating the number of methylated molecules at the target loci and the highly methylated loci; and normalizing the aggregated number of target methylated molecules by the aggregated number of the highly methylated loci. For example, normalizing the aggregate number of target methylated molecules by the aggregate number of highly methylated loci may include 100 methylated molecules for 5000 input molecules.

[0174] For example, in some embodiments, the method can monitor the amount of ctDNA over time. Therefore, the measurement of methylated molecules may need to be consistent relative to the total input amount of DNA. For example, if the patient's cancer is stable, but the amount of cfDNA collected at the second time point is twice that of the first time point, then although the cancer remains the same size, it may appear that the size of the cancer has doubled. In one aspect of the method of the present invention, whether the cfDNA is from a tumor or not, the method of the present invention can view the loci that are highly methylated in the cfDNA and standardize the target loci according to these highly methylated control loci.

[0175] One aspect of the present disclosure includes a method of quantifying methylation in a DNA sample, the method comprising: processing the sample to encode the presence or absence of methylation into the DNA sequence itself; simultaneously amplifying a mixture of the sample and QCT / spiking molecules at 10 or more different target loci and at at least one locus known to be highly methylated in cfDNA (e.g., at least one locus, at least two loci, at least three loci, at least four loci, at least five loci, at least six loci, at least seven loci, at least eight loci, at least nine loci, at least ten loci, etc.); sequencing the amplified mixture (at a read depth of at least one read per molecule in the sample); determining the number of methylated reads based on the read sequence; computationally determining the average number of methylated reads in the target loci and the average number of methylated reads in the highly methylated loci; aggregating the number of methylated reads at the target loci and the highly methylated loci; and normalizing the aggregated number of target methylated reads by the aggregated number of methylated reads at the highly methylated loci.

[0176] One aspect of the present disclosure includes a method of quantifying methylation in a DNA sample, the method comprising: adding a spike of known sequence to the sample; treating the sample to encode the presence or absence of methylation into the DNA sequence itself; adding a collection of QCT or spike molecules to the sample, the collection of molecules comprising: a) a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and b) a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; simultaneously amplifying the mixture of sample and QCT / spike molecules at 10 or more different loci, wherein the loci are selected to be differentially methylated in cancer tissues of at least two origins; sequencing the amplified mixture (at a read depth of at least one read per molecule in the sample); determining the number of methylated reads based on the read sequence; computationally determining the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the QCT / spike at each locus; and determining the origin of the tissue based on the abundance of methylated molecules at the loci.

[0177] For example, methylation patterns are different in tumor tissues of different origins. In one aspect of the invention, the method includes determining the origin of the tumor tissue based on which positions have methylated molecules in the cfDNA sample.

[0178] In some embodiments, a mixture of sample and QCT / spike molecules is amplified at 5 or more loci (eg, loci for each cancer tissue of origin).

[0179] In some embodiments, determining the origin of the tissue based on the abundance of methylated molecules at the loci includes assigning a score to each tissue type.

[0180] In one aspect of the present disclosure, a method for quantifying the amount of tumor DNA in a DNA sample is provided herein, the method comprising: adding a collection of QCT or spike molecules to the sample, the collection of molecules comprising: a) a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and b) a variant region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule; simultaneously amplifying a mixture of the sample and the QCT / spike molecules at 10 or more different loci; sequencing the amplified mixture (at a read depth of at least one read per molecule in the sample); calculating and determining the number of molecules containing somatic mutations in the sample based on the number of reads from the sample and the number of reads from the QCT / spike at each locus; and determining a lower limit on the variant allele fraction based on the total number of molecules initially present at each locus in the sample.

[0181] In another aspect, the present disclosure relates to a method for determining a DNA methylation profile in a subject, the method comprising: processing a sample isolated from the subject to encode the presence or absence of DNA methylation in a DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and an amplified collection of at least ten target loci from the DNA sequence; sequencing the co-amplification mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying (determining) the number of methylated molecules for each target locus in the sample based on the number of methylated sequence reads from each target locus in the sample and the number of reads from the collection of synthetic molecules; and determining the methylation pattern in the subject based on the number of methylated molecules for the target loci in the sample.

[0182] In another aspect, the present disclosure relates to a method for determining a DNA methylation profile in a subject over time, the method comprising: at a first time point: i) treating a sample isolated from the subject to encode the presence or absence of DNA methylation in a DNA sequence; ii) adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a nucleotide sequence corresponding to an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; iii) generating a co-amplification mixture comprising the collection of amplified synthetic molecules and amplified at least ten targets from the DNA sequence; a collection of target loci; iv) sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; v) determining the number of sequence reads that are methylated sequence reads; vi) quantifying (determining) the number of methylated molecules for each target locus in the sample based on the number of methylated sequence reads from each target locus in the sample and the number of reads from the collection of synthetic molecules; repeating steps i) to vi) at a second time point; and determining the methylation pattern in the subject based on the number of methylated molecules for each target locus in the sample at the first time point and the number of methylated molecules for each target locus in the sample at the second time point.

[0183] In some embodiments of determining a DNA methylation profile in a subject over time, the method includes repeating steps i) to vi) at a third time point, a fourth time point, or a fifth time point; and determining the methylation pattern in the subject based on the number of methylated molecules for each target locus in the sample at the third time point, the fourth time point, or the fifth time point.

[0184] In some embodiments of determining a DNA methylation profile in a subject, after detecting a change in the DNA methylation profile, a treatment selection assay is performed on the subject. In some embodiments, the treatment selection assay comprises genomic analysis to detect new somatic mutations, the abundance of somatic mutations, or both.

[0185] In some embodiments, the method detects the methylation signal at each locus, thereby creating a methylation profile at each time point, which comprises an aggregation of the methylation signals at each locus (see, e.g., Example 10). The methylation profile can then vary over time, wherein each time point is represented as an aggregation of the methylation signals at each locus (see, e.g., Example 10).

[0186] Quantifying the methylation profile over time can allow monitoring changes in the methylation profile, which is an indication of changes in disease progression. For example, a substantial increase in the percentage contribution of one or more loci to the methylation profile from one time point to another (i.e., a substantial increase in the composition of the methylation pattern associated with one or more loci) may indicate an increase in disease progression (e.g., an increase in tumor size).

[0187] In another aspect, the present disclosure relates to a method for quantifying the abundance of somatic mutations in a sample containing a DNA sequence, the method comprising: determining the abundance of somatic mutations in the sample; quantifying DNA methylation in the sample, comprising the steps of: treating the sample to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; generating a co-amplification mixture, the co-amplification mixture comprising amplification A method of preparing a method for preparing a co-amplified sample of a nucleic acid sequence of at least ten target loci from a DNA sequence amplified from a sample of a nucleic acid sequence of at least ten target loci amplified from a DNA sequence; sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying (determining) the absolute number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules; and returning a true determination result for the abundance of somatic mutations when the number of methylated molecules is above a predetermined or dynamically calculated threshold; and returning a no determination result for the abundance of somatic mutations when the number of methylated molecules is at or below a predetermined or dynamically calculated threshold.

[0188] In some embodiments, the somatic mutation is selected from one or more of the following: single nucleotide variation, copy number variation, insertion, and deletion.

[0189] In some embodiments, determining the abundance of somatic mutations comprises using the variant allele fraction (VAF).

[0190] In some embodiments, the abundance of somatic mutations indicates the presence of cancer when the abundance of somatic mutations is at or above a predetermined threshold.

[0191] In some embodiments, the method includes, when a true call is returned and if the abundance of somatic mutations indicates the presence of cancer, performing a treatment selection assay on the subject. In some embodiments, the treatment selection assay includes genomic analysis to detect new somatic mutations, the abundance of somatic mutations, or both.

[0192] In some embodiments, the method further comprises, upon returning a no call, repeating the method of determining the abundance of a somatic mutation and quantifying DNA methylation on a different sample taken from the subject.

[0193] Embodiments of the methods described herein, for example Figure 1A As shown in , this can be done in any suitable way. Figure 1A Embodiments of the methods shown in may include every combination and arrangement of various system components and various method processes (including any variations (e.g., embodiments, variations, examples, specific examples, figures, etc.)), wherein portions of the embodiments of the methods and / or processes described herein may be performed asynchronously (e.g., sequentially), simultaneously (e.g., in parallel), or in any other suitable order by and / or using one or more instances, elements, components and / or other aspects of system 200 and / or other entities described herein.

[0194] Any variants described herein (e.g., embodiments, variations, examples, specific examples, figures, etc.) and / or any portion of the variants described herein may be additionally or alternatively combined, aggregated, excluded, used, performed sequentially, performed in parallel, and / or applied in other ways.

[0195] like Figure 1A Portions of the embodiments of the methods described in the can be at least partially implemented and / or realized as a machine configured to receive a computer-readable medium storing computer-readable instructions. The instructions can be executed by a computer-executable component that can be integrated with the system. The computer-readable medium can be stored on any suitable computer-readable medium, such as RAM, ROM, flash memory, EEPROM, optical device (CD or DVD), hard drive, floppy disk drive, or any suitable device. The computer-executable component can be a general-purpose or specific application processor, but any suitable dedicated hardware or hardware / firmware combination device can alternatively or additionally execute the instructions.

[0196] Those skilled in the art will recognize from the preceding detailed description and from the accompanying drawings and claims that embodiments of the method (e.g., Figure 1A Modifications and changes may be made to the invention as shown in the accompanying drawings and / or variations without departing from the scope defined in the claims.

[0197] 6.1 Generating synthetic molecules and adding them to samples

[0198] In some embodiments, a method for quantifying DNA methylation in a sample may include: adding a collection of synthetic molecules (e.g., QCT molecules) to a treated sample, the collection of synthetic molecules comprising: a target-associated region having sequence similarity to a target sequence region of an endogenous target molecule (e.g., associated with a genetic disorder; etc.), and a variable region having sequence dissimilarity to the sequence region of the endogenous target molecule (e.g., comprising an embedded molecular identifier (EMI) region comprising a collection of variable "N" bases, wherein each "N" base is selected from any one of an "A" base, a "G" base, a "T" base, and a "C" base).

[0199] In some embodiments, a method can include generating a collection of QCT molecules that can function to generate molecules to be used (e.g., added, processed, sequenced, etc.) in one or more stages (e.g., steps, phases, periods, time periods, etc.) of at least one of sequencing library preparation and sequencing (e.g., high-throughput sequencing, etc.), such as for facilitating downstream computational processing (e.g., determining the number of QCT sequence reads to facilitate quantification of the number (or percentage) of methylated molecules).

[0200] In some embodiments, a synthetic molecule (e.g., a QCT molecule) comprises a target-associated region (e.g., each QCT molecule comprises one or more target-associated regions; etc.). Figure 1B As shown in, the target-associated region comprises sequence similarity (e.g., full sequence similarity; sequence similarity that satisfies a threshold condition; sequence similarity of a particular number of bases; etc.) with one or more target sequence regions of one or more target molecules (e.g., endogenous target molecules; target molecules corresponding to one or more biological targets; etc.), but may additionally or alternatively comprise any suitable correlation with any suitable component of the one or more target molecules. The target-associated region preferably enables co-amplification of the corresponding QCT molecule (e.g., comprising the target-associated region, etc.) and the nucleic acid molecule (e.g., nucleic acid, nucleic acid fragment, etc.) comprising the target sequence region, which may facilitate improved accuracy of molecule counting (e.g., determining molecule counting parameters; by taking into account amplification bias; etc.), but may additionally or alternatively enable any suitable process associated with sequencing library preparation, sequencing, and / or portions of the embodiments of the methods described herein. In one example, the methods described herein may include co-amplifying a collection of QCT molecules and a nucleic acid molecule comprising a biological target based on sequence similarity of a target-associated region and a target sequence region of a biological target, and wherein quantifying DNA methylation in a sample may include determining a target molecule count (e.g., a methylated molecule count) describing the number of methylated molecules associated with sequencing based on a collection of synthetic molecules (e.g., QCT sequence read clusters).

[0201] In some embodiments, the synthetic molecule (e.g., QCT molecule) can omit the target-associated region. For example, the QCT molecule can be used with components of a sample comprising a biological target without target relevance (e.g., no predetermined similarity to a target sequence region of the biological target) and / or without corresponding co-amplification with components of the sample (e.g., nucleic acid molecules comprising the target sequence region; etc.). In some examples, the QCT molecule can be pre-treated to be suitable for sequencing, such as where the pre-treated QCT molecule can be added to a sample treated for sequencing so as to be co-sequenced without co-amplification (e.g., to improve user-friendliness). QCT molecules that omit target-associated regions are preferably used to facilitate contamination parameter determination, but can be used in addition or alternatively to facilitate any suitable sequencing-related parameter determination. In one specific example, the set of QCT molecules can be adapted for subsequent sequencing (e.g., high-throughput sequencing, such as NGS; etc.), wherein generating the set of QCT molecules can include amplifying a first subset of QCT molecules of the set of QCT molecules (e.g., each QCT molecule comprises a first common QCT identifier region; etc.); and amplifying a second subset of QCT molecules of the set of QCT molecules (e.g., each QCT molecule comprises a second common QCT identifier region; etc.), wherein the QCT molecule sequencing reads are derived from sequencing corresponding to: a QCT mixture generated based on the first subset of QCT molecules and a sample comprising a biological target (e.g., comprising a first target molecule corresponding to the biological target; etc.), and an additional QCT mixture generated based on the second subset of QCT molecules and an additional sample comprising a biological target (e.g., comprising a second target molecule corresponding to the biological target; etc.), wherein the sample and the additional sample correspond to a first sample chamber and a second sample chamber of the sample chamber, respectively. However, the target-associated region and / or the QCT molecules omitting the target-associated region can be configured in any suitable manner.

[0202] In some embodiments, a synthetic molecule (e.g., a QCT molecule) comprises one or more variable regions (e.g., one or more variable regions per QCT molecule; adjacent variable regions; separated variable regions; etc.). Figure 1BAs shown in , the variable region preferably comprises sequence dissimilarity (e.g., full sequence dissimilarity; dissimilarity of a specific number of bases; partial sequence dissimilarity; etc.) to one or more sequence regions of the target molecule (e.g., different from the target sequence region; etc.). The variable region may additionally or alternatively include one or more EMI regions. In one variation, the EMI region may comprise a set of variable "N" bases (e.g., one or more variable "N" bases, etc.), wherein each "N" base is selected from (e.g., randomly selected; selected according to a predetermined statistical distribution and / or probability; etc.) any one of an "A" base, a "G" base, a "T" base, and a "C" base. In one variation, the EMI region may comprise a synthetic region (e.g., on a microarray; using silicon-based synthesis; etc.), the synthetic region comprising one or more specific bases (e.g., designed and synthesized bases; etc.), such as a synthetic region designed to facilitate QCT sequence read cluster determination (e.g., by maximizing the pairwise Hamming distance between EMI regions; etc.). In one variation, the QCT molecule may additionally or alternatively comprise multiple EMI regions (e.g., a variation region comprises multiple EMI regions; adjacent EMI regions; separated EMI regions; an EMI region comprising variable "N" bases; an EMI region comprising a synthetic region; etc.). For example, each variation region of the set of QCT molecules may comprise an embedded molecular identifier region comprising a set of variable "N" bases, wherein each "N" base is selected from any of an "A" base, a "G" base, a "T" base, and a "C" base, wherein each QCT molecule of the set of QCT molecules further comprises an additional EMI region comprising an additional set of variable "N" bases, wherein the additional EMI region is separated from the EMI region by a sequence region of the QCT molecule, such as wherein the variable "N" base is selected from any of an "A" base, a "G" base, a "T" base, and a "C" base. The set of "N" bases and the additional set of variable "N" bases can each contain a determined (e.g., predetermined) number of "N" bases (e.g., greater than three "N" bases, greater than any suitable number of "N" bases, an exact number of "N" bases; etc.), and wherein determination of sequencing-related parameters (e.g., contamination parameters) can be based on QCT sequence read clusters, which are obtained based on the EMI region and the additional EMI region of the set of QCT molecules (e.g., based on different EMI sequence reads corresponding to pairs of the EMI region and the additional EMI region; etc.).

[0203] In some embodiments, such as Figure 1BAs shown in , a QCT molecule can include a QCT identifier region that identifies the QCT molecule (and / or other suitable QCT molecules), such as a common QCT identifier region (e.g., a common sequence region having dissimilarity to one or more sequence regions of a target molecule, etc.) that identifies QCT molecules belonging to a set of QCT molecules (e.g., where different QCT identifier regions are unique to different sets of QCT molecules, etc.). In one example, the variation region of each QCT molecule of a first set of QCT molecules can include a first EMI region that is separated from a second EMI region by at least the first QCT identifier region, wherein each additional QCT molecule of a second set of QCT molecules can include a first additional EMI region that is separated from a second additional EMI region by at least the second QCT identifier region. In one example, the first EMI region, the second EMI region, the first additional EMI region, and the second additional EMI region may include a set of variable "N" bases, and wherein each "N" base is selected from any one of an "A" base, a "G" base, a "T" base, and a "C" base, and wherein computationally determining the set of QCT sequence read clusters may include determining the set of QCT sequence read clusters based on the first QCT identifier region and the second QCT identifier region and based on the first EMI region, the second EMI region, the first additional EMI region, and the second additional EMI region. In one example, for each QCT molecule of the first set of QCT molecules, the corresponding QCT molecule sequence is characterized by a full sequence similarity to a first sequence template of a biological target excluding the first QCT identifier region, the first EMI region, and the second EMI region; and wherein, for each additional QCT molecule of the second set of QCT molecules, the corresponding additional QCT molecule sequence is characterized by a full sequence similarity to a second sequence template excluding the second QCT identifier region, the first additional EMI region, and the second additional EMI region. In one embodiment, the QCT molecule sequence can be identical to the target molecule sequence (e.g., one or more regions of the target molecule sequence; etc.) except for two separated 5N sequence segments interrupted by different, previously determined QCT identifier regions (e.g., unique identifier sequences, etc.). In one embodiment, the QCT identifier region (e.g., Figure 1B) to enable the use of multiple QCT libraries, which can be added at one stage for internal controls, or added at different stages to track loss of input biological targets or other user errors. Additionally or alternatively, the QCT identifier regions can be configured in any suitable manner. However, the QCT molecules can include any suitable combination of regions of any suitable type (e.g., where different QCT molecules include the same or different types and / or numbers of regions; have any suitable sequence similarity and / or dissimilarity to sequence regions of the target molecule; etc.).

[0204] In some embodiments, the method can additionally or alternatively include generating one or more QCT libraries (e.g., each QCT library comprising QCT molecules, etc.), such as where the QCT library can comprise multiple sets of QCT molecules, such as where each set of QCT molecules can be identified by a different QCT identifier region. In one example, generating the QCT library can include amplifying different sets of QCT molecules (e.g., for preparation for sequencing, such as where the QCT molecules are amplified and then added to one or more components of a sample to generate a QCT mixture; etc.). In an example, generating the QCT library can include determining the number of QCT molecules to be included in the QCT library. In one specific example, a solution to the birthday problem can be used to determine the maximum number of unique QCT molecules that should be included in each sample given a particular diversity of QCT molecules, such as where, for 4^10 sequences that can be generated by 10 variable N bases in the QCT molecules, up to 1200 QCT molecules can be used, the probability of a single valid EMI collision is ˜0.5 (exp(−1200*1199 / 2 / 4^10) ˜0.5), and where the probability of a single valid collision is N2% with 200 QCT molecules. In one specific example, generating a QCT library can include generating a QCT library suitable for configuring (e.g., in a single stage of at least one of sequencing library preparation and high-throughput sequencing, etc.) less than 0.00001 nanograms (and / or other suitable amounts) of amplifiable QCT molecules for each sample in a collection of samples. However, determining the number of QCT molecules to be included in the QCT library and generating the QCT library can be performed in any suitable manner.

[0205] In one embodiment, a QCT library can be generated by synthesizing the complementary strand of a single-stranded oligonucleotide sequence containing a variable "N" sequence. In a specific example, a double-stranded QCT library can be generated by resuspending and annealing a QCT ultramer with a complementary primer sequence, extending the sequence using a Klenow fragment (exo-), and treating with exonuclease I. The final product can be purified to remove unused single-stranded DNA molecules, and the QCT library can be quantified using a fluorescent assay such as the Qubit HS assay, whereby the number of QCT molecules to be added to each sample can be calculated using the expected molecular weight of the double-stranded QCT molecules. However, generating QCT molecules can be performed in any suitable manner.

[0206] In some embodiments, synthetic molecule libraries (e.g., QCT libraries) can be added at different sequencing library preparation stages (e.g., sample preparation stages) and / or sequencing stages to track sample loss. In one embodiment, if a first set of QCT molecules (e.g., QCT1 molecules; a first QCT molecule comprising a first consensus QCT identifier region; etc.) is dispensed at the point of sample collection, and an equal amount of a second set of QCT molecules (e.g., QCT2 molecules; a second QCT molecule comprising a second consensus QCT identifier region; etc.) is dispensed after sample purification (or after treating the sample to encode the presence or absence of DNA methylation), the purification yield (or treatment efficiency yield) can be assessed via comparison of the molecule counts of the first set of QCT molecules to the second set of QCT molecules (e.g., QCT1 molecule counts compared to QCT2 molecule counts, etc.). In one example, a synthetic molecule library can be used to track the efficiency of treating a sample to encode the presence or absence of DNA methylation (i.e., measuring bisulfite conversion efficiency).

[0207] In some embodiments, the set of synthetic molecules includes: a target-related region with a nucleotide sequence that matches the target sequence region of an endogenous target, and a variable region with a nucleotide sequence that does not match the target sequence region of an endogenous target molecule. As used herein, the term "match" refers to sufficient complementarity between two nucleotide sequences, wherein the sequence binds and forms the basis of a nucleic acid extension reaction (e.g., amplification). Sufficient complementarity can refer to a match of up to 5 mismatches between two nucleotide sequences (e.g., a target-related region on a synthetic molecule and a target sequence region on an endogenous target). As used herein, the expression "mismatch" refers to a lack of complementarity between two nucleotide sequences, so that it is impossible to achieve enough Watson-Crick base pairing to enable the two nucleotide sequences to combine, and amplification does not occur or rarely occurs when put into an amplification reaction. Lack of complementarity can refer to a match of more than 5 mismatches between two nucleotide sequences (e.g., a target-related region on a synthetic molecule and a target sequence region on an endogenous target).

[0208] 6.2 Co-amplification and sequencing of the co-amplification mixture

[0209] In some embodiments, generating a co-amplification mixture comprising a set of amplified synthetic molecules and a set of amplified at least ten target loci from a DNA sequence comprises amplifying synthetic molecules (e.g., QCT molecules) and nucleic acid molecules (e.g., nucleic acids encoding the presence or absence of DNA methylation) comprising a target-associated region. For example, the target-associated region preferably enables co-amplification of corresponding QCT molecules (e.g., comprising a target-associated region, etc.) and nucleic acid molecules (e.g., nucleic acids, nucleic acid fragments, etc.) comprising a target sequence region, which can facilitate improved accuracy of molecule counting (e.g., determining molecule counting parameters; by considering amplification bias; etc.), and can also additionally or alternatively enable any suitable process related to sequencing library preparation, sequencing, and / or portions of the embodiments of the method. In one example, co-amplification of a set of QCT molecules and nucleic acid molecules comprising methylated DNA is based on sequence similarity of the target-associated region to the target sequence region of the methylated DNA.

[0210] In some embodiments, the sample comprises a DNA sequence having a locus that is predicted to have increased DNA methylation in cancer tissue compared to non-cancerous tissue. The loci that are predicted to have increased DNA methylation in cancer tissue compared to non-cancerous tissue are selected based on, for example, but not limited to, population-based surveys of methylation loci, patient-specific metrics (e.g., methylation patterns from tumors), and / or literature-based studies identifying methylation loci.

[0211] In some embodiments, the loci that DNA methylation is expected to increase in cancer tissue compared to non-cancerous tissue are amplified and analyzed for the presence of DNA methylation. In some embodiments, the loci that DNA methylation is expected to increase in cancer tissue compared to non-cancerous tissue include at least 10, 20, 50, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 or more loci, wherein compared to non-cancerous tissue, each locus in cancer tissue is expected to increase DNA methylation. In some embodiments, the loci that DNA methylation is expected to increase in cancer tissue compared to non-cancerous tissue also include loci (e.g., standardized loci) in which DNA sequences are methylated (e.g., highly methylated) in cancer tissue and non-cancerous tissue. In such a case, generating a co-amplification mixture results in a co-amplification mixture comprising: a collection of amplified synthetic molecules, a collection of amplified loci (e.g., at least ten target loci predicted to have increased DNA methylation in cancer tissue compared to non-cancerous tissue), and a collection of amplified at least one normalizing locus (e.g., a locus that is methylated (e.g., highly methylated) in both cancer tissue and non-cancerous tissue).

[0212] In some embodiments, sequencing associated with one or more embodiments of method 100 (e.g., associated with 116) preferably includes high-throughput sequencing, which may include any one or more of the following and / or be associated with any one or more of the following: NGS, NGS-related technologies, massively parallel signature sequencing, Polony sequencing, 454 pyrophosphate sequencing, Illumina sequencing, SOLiD sequencing, Ion Torrent semiconductor sequencing, DNA nanoball sequencing, Heliscope single molecule sequencing, single molecule real-time (SMRT) sequencing, nanopore DNA sequencing, sequencing technology of any generation (e.g., second generation sequencing technology, third generation sequencing technology, fourth generation sequencing technology, etc.), amplicon-related sequencing (e.g., targeted amplicon sequencing), metagenomic-related sequencing, synthetic sequencing, tunneling current sequencing, hybridization sequencing, mass spectrometry sequencing, microscopy-based technology, and / or any suitable technology involving high-throughput sequencing. In some embodiments, sequencing may include any suitable sequencing technology (e.g., Sanger sequencing, capillary sequencing, etc.).

[0213] 6.3 Identification and Quantification of Methylated Molecules

[0214] In some embodiments, determining the number of methylated molecules in a sample based on the number of methylation reads from the sample and the number of reads from the QCT / spiked-in at each locus includes one or more of: target molecule counts (e.g., absolute molecule counts of methylated molecules, such as in the original sample; absolute counts of endogenous target molecules, such as in the original sample; etc.); reference molecule counts (e.g., absolute counts of endogenous reference molecules; such as in the original sample; etc.); QCT molecule counts (e.g., corresponding to the number of valid QCT sequence read clusters; corresponding to the number of different QCT molecules added to the sample components; etc.); related ratios (e.g., correction factors; ratios between molecule counts and related number of sequence reads; etc.); and / or any other suitable parameters related to methylated molecule counts.

[0215] like Figure 1A As shown in , methylated molecule counts are preferably used to facilitate one or more diagnoses 122, but may additionally or alternatively be used in any suitable portion (e.g., as input) of an embodiment of method 100, including facilitating treatment of one or more conditions 124.

[0216] In some embodiments, determining a methylated molecule count parameter (e.g., methylated molecule count; etc.) can be based on a correction factor ratio, such as determined by multiplying the number of methylated molecule sequence reads by the correction factor ratio, which is determined based on the QCT molecule count (e.g., the number corresponding to QCT sequence read clusters, such as the number of valid QCT sequence read clusters; etc.) and the QCT molecule sequence reads (e.g., the number of QCT molecule sequence reads corresponding to the QCT sequence read clusters; etc.). In a specific example, the number of valid non-contaminating QCT sequence read clusters (e.g., QCT sequence read clusters remaining after discarding QCT sequence read clusters with 2 or fewer reads and / or with any suitable number or fewer reads; etc.) can be indicative of the QCT molecule count (e.g., the number of QCT molecules for a particular sample chamber; the number of QCT molecules for a particular sample; the number of QCT molecules for a particular sample identifier; etc.). In a specific example, the correction factor can be obtained by dividing the QCT molecule count by the sequencing reads obtained from the corresponding QCT molecule, such as where multiplying the correction factor by the sequencing reads belonging to the target molecule (e.g., in a specific sample chamber; from a specific sample; associated with a specific sample identifier; etc.) can obtain the target molecule count (e.g., the absolute number of initial biological target molecules accessible to the amplification assay; etc.). In one example, the average QCT sequencing depth used to determine the absolute count of the endogenous target molecule and the absolute count of the endogenous reference molecule is determined separately by their corresponding QCT.

[0217] Alternatively, in one embodiment, a read depth threshold for discarding QCT sequence read clusters (e.g., for determining a molecular count parameter and / or a suitable sequencing-related parameter; etc.) can be adaptively determined based on a characteristic of the depth distribution of QCT molecular sequence reads (e.g., EMI sequence reads). For example, a threshold can be set for each index sample by calculating the average EMI read depth within each sample, calculating the square root of the average read depth, and discarding QCT sequence read clusters with a read depth below the square root of the average read depth. Additionally or alternatively, the read depth threshold for discarding QCT sequence read clusters can be calculated in any suitable manner. However, determining the methylated molecular count can be performed in any suitable manner.

[0218] In some embodiments, variant calling includes determining the variant allele frequency (VAF) of a DNA variant. As used herein, "variant allele frequency" or "variant allele fraction" refers to the percentage of sequence reads observed to match a specific DNA variant divided by the overall sequence at that locus.

[0219] In some embodiments of (eg, quantifying DNA methylation), Figure 1AThe methods can be used to a) determine parameters for use in an algorithm for determining a diagnostic result of an assay, b) track loss of input methylated DNA (e.g., at different stages of an assay), c) return a no-call result when the number of methylated molecules is too low (e.g., determining when an assay is unreliable, etc.), d) return a no-call result when the corresponding quantification of the abundance of somatic mutations in a sample is too low (e.g., determining when an assay is unreliable), e) design an assay for detecting changes in methylation profiles over time, and / or f) assist in treatment decisions and clinical decisions based on the results of a diagnostic assay.

[0220] Embodiments may additionally or alternatively determine the fraction of biological material accessible to the assay, such as by quantifying biological targets (e.g., methylated molecules) based on the use of QCT molecules, which can improve the measurement of total genomic material available and the calculation of expected biological target (e.g., methylated molecule) concentrations, since not all targets are accessible to the assay. In one specific example, this may be caused by shearing of DNA into a short size distribution, such as for circulating free DNA, which is measured in liquid biopsy applications for measuring circulating tumor DNA.

[0221] The methods provided herein (see e.g. Figure 1A ) can be used in connection with one or more conditions (e.g., used in connection with characterizing, diagnosing, treating, and / or performing a process associated with one or more conditions; etc.), wherein the conditions may include one or more of and / or otherwise be associated with one or more of: cancer (e.g., by analysis associated with any suitable oncogene, cancer biomarker, and / or other cancer-related target; by analysis associated with a liquid biopsy) and / or any other suitable condition. In one example, ( Figure 1A and / or as described herein) methods may include determining a methylated molecule count (e.g., corresponding to the number of methylated molecules in a sample; based on the use of a QCT molecule; etc.) for facilitating a diagnosis associated with a liquid biopsy.

[0222] In some embodiments, the number of methylated molecules of cfDNA and the number of methylated molecules of gDNA are quantified in the same workflow. For example, the methylated molecules of cfDNA and the methylated molecules of gDNA can be indexed (e.g., using a defined index sequence) so that the source of the methylated molecules (e.g., cfDNA vs. gDNA) can be distinguished. This enables cfDNA and gDNA to be amplified and / or sequenced (e.g., multiplexed) in the same reaction.

[0223] 6.4 Processing methylation sequencing reads

[0224] The methods provided herein (see e.g. Figure 1A) include a step of processing methylation sequence reads. In some embodiments, the processing of methylation sequence reads includes filtering out selected hypermethylated target loci; and / or subtracting background methylation. The processing step (e.g., filtering out selected hypermethylated target loci and / or subtracting background methylation) can be performed before determining the number of methylated molecules in the sample based on the number of methylation reads from the sample and the number of sequence reads from the set of synthetic molecules. The processing step (e.g., filtering out selected hypermethylated target loci and / or subtracting background methylation) can be performed before determining the number of molecules in the number of methylated molecules in the sample based on the number of methylation reads from the sample and the number of sequence reads from the set of synthetic molecules, but before the polymerization step.

[0225] In some embodiments, the step of subtracting background methylation includes subtracting the number of methylated molecules measured in the buffy coat (ie, gDNA) from the number of methylated molecules in cfDNA. In such a case, both cfDNA and gDNA are quantified for methylated molecules. For example, the methylation measured in gDNA is then subtracted from the cfDNA methylation to remove the background from the cfDNA methylation signal. In some embodiments, the background subtraction step includes subtracting the number of methylated molecules measured in the buffy coat (ie, gDNA) from the number of methylated molecules in cfDNA on a per locus basis. For example, the methylation measured in gDNA for the same target locus is subtracted from the cfDNA methylation measured for a specific target locus to remove the background from the cfDNA methylation signal on a per locus basis.

[0226] In some embodiments, the step of filtering out selected hypermethylated target loci includes filtering out target loci whose total number of methylated molecules in the buffy coat is higher than a threshold value (e.g., a predetermined number of methylated molecules quantified). In some embodiments, the step of filtering out selected hypermethylated target loci may include filtering out target loci with high background methylation before quantifying the number of methylated molecules in the sample. In some embodiments, the selected hypermethylated target loci filtered out include target loci with hypermethylated cfDNA signals in non-cancer subjects. In some embodiments, the selected hypermethylated target loci are selected from a global blacklist (see, e.g., Example 6). In some embodiments, the selected hypermethylated target loci are selected from a personal blacklist (see, e.g., Example 6).

[0227] In some embodiments, the selected high methylation target loci filtered out include target loci whose number (e.g., total number) of methylated molecules in the buffy coat is higher than a threshold. In some embodiments, the threshold is sample-specific (e.g., the number of methylated molecules in the buffy coat is determined for each sample). In some embodiments, the threshold is subject-specific (e.g., the number of methylated molecules in the buffy coat is determined for each subject). In some embodiments, the threshold includes a predetermined average tumor methylation quantitative equivalent (QE), a predetermined maximum tumor methylation QE, or a combination thereof. Quantitative equivalent (QE) is an estimate of the number of genomic equivalents of a locus based on QCT analysis. In some embodiments, the threshold includes the maximum tumor QE>15 in any non-cancer sample, wherein the high methylation target loci with a maximum tumor QE greater than 15 in any non-cancer sample are filtered out. In some embodiments, the threshold includes an average tumor QE (including 0s)>2, wherein the high methylation target loci with an average tumor QE>2 are filtered out.

[0228] 6.5 Aggregation and Normalization

[0229] The methods provided herein (see e.g. Figure 1A ) include a step of aggregating the number of methylated molecules on all target loci or a subset of target loci (e.g., a subset of the at least ten target loci) to quantify DNA methylation in the sample. In some embodiments, the method includes aggregating the number of methylated molecules on at least two target loci (e.g., at least two of the at least ten target loci) to quantify DNA methylation in the sample. In some embodiments, the method includes aggregating the methylated molecules on at least two loci (e.g., at least two of the at least ten target loci) in a cfDNA (plasma) sample. In some embodiments, the method includes aggregating the methylated molecules on at least two loci (e.g., at least two of the at least ten target loci) in a buffy coat (gDNA) sample.

[0230] In some embodiments, aggregating the number of methylated molecules on all target loci or a subset of target loci (e.g., all or a subset of the at least ten target loci) includes aggregating target loci containing less than a threshold amount of methylated molecules in the buffy coat sample. The threshold value for the amount of methylated molecules in the buffy coat sample can be predetermined or dynamically calculated. A dynamically calculated threshold value is not a fixed threshold value, but varies based on certain conditions or inputs. With respect to the threshold amount of methylated molecules, the threshold value can be dynamically calculated based on factors such as the size of the system being studied, the concentration of the molecules present, or the sensitivity of the detection method used to measure the molecules. For example, in a sample (e.g., a blood draw), the threshold amount of methylated molecules may vary with the type of cancer or the stage of cancer.

[0231] The methods provided herein (see e.g. Figure 1A ) include a step of standardizing the polymerization number of methylated molecules to help quantify DNA methylation in a sample. In one embodiment, the method includes standardizing the polymerization number of target methylated molecules by the methylated molecules in the standardized locus (e.g., the methylated molecules of at least one standardized locus). In some embodiments, the method includes standardizing the polymerization number of target methylated molecules by the polymerization number of methylated molecules in the standardized locus. In such a case, the method includes aggregating the number of methylated molecules on at least two standardized loci. In some embodiments, the method includes standardizing the polymerization number of methylated molecules from at least two of at least 10 target loci by the methylated molecules of at least one standardized locus.

[0232] The aggregation and / or normalization steps enable quantification of a tumor methylation score that represents the normalized sum of the methylated molecules that are hypermethylated (eg, at the at least ten target loci) in the sample.

[0233] The methods provided herein (see e.g. Figure 1A ) can be used to quantify the abundance of tumor DNA in a sample (see, e.g., Section 6.8).

[0234] The methods provided herein (see e.g. Figure 1A ) can be used to determine a DNA methylation profile in a subject (see, e.g., Section 6.9).

[0235] The methods provided herein (see e.g. Figure 1A ) can be used to quantify the abundance of somatic mutations in a sample (see, e.g., Section 6.10).

[0236] Additionally or alternatively, the data described herein (e.g., sequencing-related parameters, identifiers, read depths, sequence reads, sequence region determination, QCT molecule design, primer design, etc.) can be associated with any suitable time indicator (e.g., seconds, minutes, hours, days, weeks, time periods, time points, timestamps, etc.), including one or more of: a time indicator that indicates when the data was collected, determined, transmitted, received, and / or otherwise processed; a time indicator that provides context for the content described by the data, such as a time indicator that indicates the order of stages of sequencing library preparation and / or sequencing; changes in a time indicator (e.g., changes in the data over time; changes in the data; data patterns; data trends; data extrapolations and / or other predictions, etc.); and / or any other suitable indicator associated with time.

[0237] Additionally or alternatively, parameters, metrics, inputs, outputs, and / or other suitable data described herein can be associated with a type of value, including any one or more of the following: a score, a binary value, a classification, a confidence level, an identifier (e.g., a sample identifier, a QCT molecule identifier, etc.), a value along a spectrum, and / or any other suitable type of value. For any suitable component associated with embodiments of method 100 and / or system 200, any suitable type of data described herein can be used as input, generated as output, and / or manipulated in any suitable manner.

[0238] In some cases, such as Figure 1B The embodiments described in may additionally or alternatively include a sample processing network configured to generate molecules (e.g., QCT molecules; QCT libraries; etc.), process biological samples and / or perform other suitable processes; a sequencing system configured to sequence processed genetic material from a mixture generated based on the biological sample and the QCT molecules; a computing system (e.g., a remote computing system; a local computing system; etc.) configured to analyze sequence reads, determine QCT sequence read clusters, determine sequencing-related parameters, facilitate diagnosis, facilitate treatment and / or perform other suitable processes (e.g., computing processes); and / or any other suitable components. Figure 1B The components of the system in the embodiment may be physically and / or logically integrated in any manner (e.g., with any suitable functional distribution of the components, such as with respect to Figure 1A ; etc.). However, Figure 1A Methods and Figure 1B The system may be configured in any suitable manner.

[0239] 6.6 Facilitating diagnosis, treatment or other evaluation

[0240] In one aspect, the present disclosure relates to a method that may additionally or alternatively include facilitating diagnosis 122 ( Figure 1A ), which can be used to assist, determine, provide and / or otherwise facilitate one or more diagnoses of one or more conditions.

[0241] Facilitating one or more diagnoses may include any one or more of the following: determining one or more diagnoses (e.g., based on the number of methylated molecules; etc.); providing one or more diagnoses (e.g., to one or more users; to one or more care providers, such as for use by one or more care providers in providing a medical diagnosis to a patient; etc.); assisting one or more diagnoses (e.g., providing one or more sequencing-related parameters and / or other suitable parameters to one or more care providers and / or other suitable entities for determining a diagnosis, such as in combination with other data; etc.); and / or any suitable process related to a diagnosis. For example, assisting a diagnosis may include providing a quantification of DNA methylation in a sample from a patient (e.g., to a user, to a care provider; etc.), wherein the quantification is suitable for use in determining a diagnostic result of an assay associated with a liquid biopsy. In one example, quantifying DNA methylation may include quantifying the number of methylated molecules in a sample (e.g., the absolute number of methylated molecules in a sample) to facilitate a diagnosis associated with a liquid biopsy.

[0242] In some embodiments, facilitating diagnosis (e.g., cancer diagnosis) can include facilitating diagnosis based on the amount of methylated molecules in a sample (see Figure 1A , 122). For example, determining the absolute number of methylated molecules may include determining the absolute number of methylated molecules associated with a DNA sequence based on dividing the total read counts of the methylated molecules by the average QCT sequencing depth, wherein determining the absolute number of methylated molecules may include determining the absolute number of methylated molecules associated with a second DNA sequence that is not expected to be methylated based on dividing the total read counts of the methylated molecules by the average QCT sequencing depth, and wherein facilitating diagnosis (e.g., cancer diagnosis) may include facilitating diagnosis (e.g., cancer diagnosis) based on comparison of methylated molecules; etc.

[0243] In some embodiments, facilitating diagnosis (e.g., cancer diagnosis) can include facilitating diagnosis based on the amount of methylated molecules in a sample (see Figure 1A , 122). For example, determining the absolute number of methylated molecules may include determining the absolute number of methylated molecules based on dividing the total read counts of methylated molecules by the average QCT sequencing depth, wherein determining the absolute counts of methylated molecules may include determining the absolute counts of methylated molecules in a sample that is not expected to have a methylated sequence at the same corresponding DNA sequence based on dividing the total read counts of methylated molecules by the average QCT sequencing depth, and wherein facilitating diagnosis (e.g., cancer diagnosis) may include facilitating diagnosis (e.g., cancer diagnosis) of methylated molecules based on comparison.

[0244] In one aspect, the present disclosure relates to a method that may additionally or alternatively include promoting treatment 124 ( Figure 1A). For example, the methods described herein, including: quantification of DNA methylation, determination of DNA methylation profiles, and / or quantification of the abundance of somatic mutations, can be used to assess whether a cancer treatment is effective for a particular patient. The results provided by the methods described herein can help provide information to an oncologist to facilitate future treatment decisions. In one example, when the method for quantifying DNA methylation in a sample results in the discovery of increased DNA methylation in a subject, the subject's oncologist can use the quantified increase in DNA methylation as a partial basis for adjusting the subject's treatment regimen. In another example, when the method for determining the methylation profile in a sample results in the discovery of changes in the DNA methylation profile that indicate changes in the subject's cancer, the subject's oncologist can use the changes in the DNA methylation profile as a partial basis for adjusting the subject's treatment regimen. In another example, when the method for quantifying the abundance of somatic mutations in a sample results in the discovery of somatic mutations that indicate changes in cancer and / or cancer, the subject's oncologist can use the quantification of the abundance of somatic mutations as a partial basis for adjusting the subject's treatment regimen.

[0245] In one embodiment, the present disclosure relates to a method that may additionally or alternatively include a recommendation for an additional assessment of a subject, based on which the method may be used to assist, determine, provide and / or otherwise promote diagnosis or treatment. For example, the methods described herein, including: quantification of DNA methylation, determination of DNA methylation profiles and / or quantification of the abundance of somatic mutations, may be used as a basis for additional assessment of a subject or recommendation for additional assessment of a subject. In some cases, performing an additional assessment includes performing a treatment selection assay on a subject. In some cases, performing an additional assessment includes performing a genomic analysis assay on a subject (i.e., an assay to assess a subject's mutation spectrum).

[0246] In one embodiment, the DNA methylation quantified in a sample collected from a subject according to the methods described herein can be incorporated into a clinical recommendation for the subject. The clinical recommendation can include a plan for further testing or treatment: for example, based on the DNA methylation quantified in the sample and the clinical significance of the DNA methylation, a plan for further testing or treatment can be formulated. In some cases, this may involve additional testing to confirm the diagnosis, monitor disease progression or recurrence, or prescribe targeted therapies specific to the subject's DNA methylation profile.

[0247] 6.7 Quantification of DNA methylation in samples

[0248] The present disclosure also relates to methods of quantifying DNA methylation in a sample comprising a DNA sequence.

[0249] In some embodiments, the method comprises: treating the sample to encode the presence or absence of DNA methylation in the DNA sequence. In other embodiments, the sample has been treated to encode the presence or absence of DNA methylation in the DNA sequence.

[0250] In some embodiments, a sample (e.g., a sample containing a mixture of methylated and unmethylated DNA sequences) may include at least ten target loci from a DNA sequence (e.g., a locus that is amplified and evaluated to quantify methylation). In some cases, each locus in the at least ten target loci is selected due to an expected increase in DNA methylation at each locus in cancer tissue compared to normal tissue. The sample may also include a standardized locus (e.g., a locus that is highly methylated in both cancer tissue and normal tissue (e.g., non-cancerous tissue)). In some cases, the standardized locus is included in the target locus (e.g., the at least ten target loci). The method also includes adding a collection of synthetic molecules to the sample, the synthetic molecules comprising a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule having at least one of the target loci (wherein the endogenous target molecule comprises at least one of the target loci), and a variant region having a sequence that does not match a sequence region of an endogenous target molecule. The method also includes generating a co-amplification mixture, which includes a set of amplified synthetic molecules and a set of amplified at least ten target loci from a DNA sequence. In some embodiments, the co-amplification mixture includes at least one standardized locus, which is expected to have high methylation in cfDNA in both cancerous tissue and non-cancerous tissue. The method includes sequencing the co-amplification mixture to generate sequence reads, and determining the number of sequence reads as methylated sequence reads. The method also includes quantifying (determining) the number of methylated molecules in the sample (for example, each locus) based on the number of methylated reads from the sample (for example, from each locus) and the number of reads from the set of synthetic molecules (for example, absolute number). In one embodiment, the method includes quantifying (determining) the number of methylated molecules in each locus in the sample based on the number of methylated reads from each target locus of the sample and the number of reads from the set of synthetic molecules (for example, absolute number). In some embodiments, the method also includes polymerizing the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in the sample.

[0251] In another embodiment, the present disclosure also relates to a method for quantifying DNA methylation in a sample containing a cell-free DNA (cfDNA) sequence. In some embodiments, the quantified DNA methylation is a tumor methylation score. The method includes processing a sample (e.g., wherein the sample is taken from a blood draw including plasma with cfDNA and a buffy coat layer with genomic DNA (gDNA)) to encode the presence or absence of DNA methylation in a DNA sequence (e.g., cfDNA and / or gDNA sequence). A sample (e.g., a sample containing a mixture of methylated and unmethylated DNA sequences) may include at least ten target loci from a DNA sequence (e.g., a locus that is amplified and evaluated for the presence of methylation). In some cases, each locus in the at least ten target loci is selected due to an expected increase in DNA methylation at each locus in a cancer tissue compared to a normal tissue (e.g., non-cancerous tissue). The sample may also include a standardized locus (e.g., a locus that is highly methylated in both cancer tissue and normal (e.g., non-cancerous) tissue). A collection of synthetic molecules is added to the sample, wherein the collection of molecules comprises: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule. The sample containing the synthetic molecules is amplified to generate a co-amplification mixture, which comprises a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from a DNA sequence, wherein the at least ten target loci optionally include a standardization locus. The method comprises sequencing the co-amplification mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads and determining the number of sequence reads that are methylation sequence reads. The method also includes processing the methylated sequence reads using one or more of the following: filtering out selected highly methylated target loci (e.g., filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold); and / or subtracting background methylation (e.g., subtracting background methylation includes subtracting the number of methylated molecules as measured in the buffy coat from the number of methylated molecules in cfDNA (e.g., on a per locus basis)). The method also includes quantifying (determining) the number of methylated molecules in the sample (e.g., an absolute number) based on the number of methylated reads from the sample and the number of sequence reads from a collection of synthetic molecules. In one embodiment, the method includes quantifying (determining) the number of methylated molecules per locus in the sample (e.g., an absolute number) based on the number of methylated reads from each target locus of the sample and the number of reads from a collection of synthetic molecules. In some embodiments, the method also includes aggregating the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in the DNA sample. In some embodiments, the sample is obtained from a blood draw that contains plasma and buffy coat, wherein the plasma contains cfDNA sequences and the buffy coat contains gDNA.

[0252] In one embodiment, the method includes determining (quantifying) the number of methylated molecules in cfDNA and / or gDNA at the at least ten target loci using QCT molecular counting technology (e.g., as described in U.S. Patent Publication No. 2019 / 0211395A1, which is incorporated by reference), wherein each of the at least ten target loci is selected due to an expected increase in DNA methylation at each locus in cancer tissue compared to normal (e.g., non-cancerous) tissue. In such an embodiment, the number of methylated molecules of cfDNA and the number of methylated molecules of gDNA are each quantified (e.g., in separate workflows) by: processing a sample to encode the presence or absence of DNA methylation in a DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of the endogenous target molecule; generating a co-amplification mixture comprising the collection of amplified synthetic molecules and the collection of amplified at least ten target loci (or a subset thereof) from the DNA sequence; sequencing the co-amplified mixture (e.g., at a read depth of at least one read sequence per molecule in the sample) to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; and quantifying (determining) the number of methylated molecules in the sample (e.g., an absolute number) based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules.

[0253] In some embodiments, quantifying DNA methylation in a sample includes quantifying the number of methylated molecules of cfDNA and the number of methylated molecules of gDNA, which can be quantified in the same workflow. In one embodiment, the methylated molecules of cfDNA and the methylated molecules of gDNA can be indexed (e.g., using a defined index sequence) to be able to distinguish the source of the methylated molecules (e.g., cfDNA vs. gDNA), thereby enabling cfDNA and gDNA to be amplified and / or sequenced (e.g., multiplexed) in the same reaction.

[0254] In one embodiment, quantifying DNA methylation in a sample containing a gDNA sequence from a buffy coat layer includes: processing a sample containing methylated and unmethylated gDNA sequences from a buffy coat layer to encode the presence or absence of DNA methylation in the gDNA sequence. In such a case, the sample contains at least ten target loci from the gDNA sequence, wherein each of the at least ten target loci is selected due to an expected increase in DNA methylation at each locus in cancer tissue compared to normal tissue and is the same target locus analyzed in cfDNA extracted from plasma. The method includes adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence matching a target sequence region of an endogenous target, and a variant region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule. The method also includes generating a co-amplification mixture, the co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of at least ten target loci from a DNA sequence amplified, and sequencing the co-amplification mixture to generate sequence reads. In addition, the method includes determining the number of sequence reads that are methylated sequence reads; and quantifying (determining) the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules. Optionally, the method includes aggregating the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in a sample containing DNA sequences from the buffy coat.

[0255] In some embodiments, the method for quantifying DNA methylation in a sample includes adding a spike of known sequence and amount to the sample prior to the processing step. The spike comprises a known sequence having unmethylated cytosines that are converted to uracils after being subjected to the processing step. The initial number of spike molecules and the percentage of cytosine bases converted to thymine / uracil bases can be calculated. This enables calculation of the bisulfite conversion yield and conversion efficiency, thereby establishing a bisulfite conversion QC metric, and thus determining whether the sample has not passed the bisulfite conversion step (i.e., a step of processing a sample containing a DNA sequence to encode the presence or absence of DNA methylation in the DNA sequence).

[0256] In some embodiments, the method for quantifying DNA methylation in a sample includes a step of processing methylated sequence reads by one or more of the following: filtering out selected hypermethylated target loci (e.g., filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold); or subtracting background methylation (e.g., subtracting the number of methylated molecules as measured in the buffy coat from the number of methylated molecules in cfDNA, which can be performed on a per-locus basis). In some embodiments, processing is performed prior to the step of quantifying the number (e.g., absolute number) of methylated molecules in the sample. In some embodiments, filtering hypermethylated target loci, subtracting background methylation, or both are performed prior to quantifying the number of methylated molecules in the sample.

[0257] In some embodiments, the method for quantifying DNA methylation in a sample includes further comprising the step of aggregating the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in the sample. Aggregation may include aggregating target loci containing less than a threshold amount of methylated molecules in the buffy coat sample.

[0258] In some embodiments, the method for quantifying DNA methylation in a sample includes aggregating the number of methylated molecules on the at least ten target loci (or a subset thereof, such as at least two target loci); and normalizing the number of methylated molecules from all or a subset of the at least 10 target loci by the methylated molecules of the at least one normalizing locus. In some embodiments, this results in the quantification of DNA methylation (e.g., a tumor methylation score) representing the normalized sum of the methylated molecules (e.g., at the at least ten target loci) that are hypermethylated in the sample.

[0259] In some embodiments, aggregating the number of methylated molecules on the at least ten target loci (or a subset thereof, such as at least two target loci) comprises aggregating at least one target locus that exhibits high methylation in the sample containing the DNA sequence.

[0260] In one embodiment, the method also includes determining the origin of the cancer tissue by quantifying methylation in the DNA sample, wherein determining the origin of the cancer tissue includes: determining the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules at each locus; and determining the origin of the tissue based on the abundance of methylated molecules at the locus.

[0261] 6.8 Quantification of Tumor DNA in Samples

[0262] The present disclosure also relates to a method for quantifying the amount of tumor DNA in a sample containing a DNA sequence. The method includes: adding a collection of synthetic molecules to a sample (e.g., wherein the sample is taken from a blood draw containing plasma with cfDNA and a buffy coat layer with genomic DNA (gDNA)), the collection of molecules comprising: a target-related region having a nucleotide sequence matching a target sequence region of an endogenous target (e.g., comprising at least one target locus), and a variant region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule. The method includes generating a co-amplification mixture, the co-amplification mixture comprising a collection of amplified synthetic molecules, a collection of at least 10 target loci from a DNA sequence amplified, and one or more standardized loci that are optionally amplified; sequencing the co-amplification mixture with a read depth of at least one read sequence per molecule in the sample to generate sequence reads; quantifying (determining) the number of molecules containing somatic mutations in the sample based on the number of reads from the sample containing somatic mutations and the number of reads from the collection of synthetic molecules.

[0263] 6.9 Determining the Methylation Profile in a Subject Over Time

[0264] The disclosure also relates to methods for determining a DNA methylation profile in a subject over time. Using continuous quantification of DNA methylation to determine a methylation profile in a subject over time will enable early indications of the response or progression (or recurrence) of cancer to treatment. As described herein, the methods of the present invention uniquely enable continuous quantification of this type of DNA methylation to be achieved.

[0265] In one embodiment, a method for determining a DNA methylation profile in a subject over time includes quantifying DNA methylation at a first time point and a second time point; and determining whether methylation (which is an alternative tumor measurement) has changed over time (e.g., increased or decreased compared to the measurement at the first time point). In some cases, the change in DNA methylation (i.e., the Tumor Methylation Score) exceeds a significance threshold and is reported as an increase or decrease. In some embodiments, a method for determining a DNA methylation profile in a subject over time can distinguish a change of 0.2 percentage points (or less (e.g., a change of 0.1%, 0.05%, 0.01%, 0.05%, 0.001%) in DNA methylation (e.g., the number of methylated molecules) with a discreteness of 3 standard deviations.

[0266] In some embodiments, a method for determining a DNA methylation profile in a subject over time comprises: i) processing a sample isolated from the subject to encode the presence or absence of DNA methylation in a DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence, wherein each of the at least ten target loci is selected based on an expected increase in DNA methylation at each locus in cancer tissue compared to non-cancerous tissue and the sample optionally comprises a normalizing locus (e.g., a locus that is highly methylated in both cancer tissue and normal tissue (e.g., non-cancerous tissue)); ii) adding to the sample a collection of synthetic molecules (e.g., QCT molecules), the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a target-associated region having a sequence region that does not match an endogenous target molecule. nucleotide sequence; iii) generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of at least ten target loci from amplified DNA sequences; iv) sequencing the co-amplification mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; v) determining the number of sequence reads that are methylated sequence reads; vi) quantifying (determining) the number of methylated molecules in the sample (e.g., for each target locus) based on the number of methylated sequence reads in the sample (e.g., from each target locus) and the number of reads from the collection of synthetic molecules; repeating steps i) to vi) at a second time point; and determining the methylation pattern in the subject based on the number of methylated molecules in the sample (e.g., for each target locus) at the first time point and the number of methylated molecules in the sample (e.g., for each target locus) at the second time point. In some embodiments, the sample has been processed to encode the presence or absence of DNA methylation in the DNA sequence, and therefore step i) is not performed.

[0267] In some embodiments, determining the methylation profile in a subject at a first time point and a second time point will identify a change in the methylation profile.

[0268] In one embodiment, a method for determining a DNA methylation profile in a subject over time comprises repeating steps i) to vi) at a third time point; and determining the methylation profile in the subject based on the number of methylated molecules (e.g., of each target locus) in the sample at the third time point.

[0269] In some embodiments, determining the methylation profile in a subject at a first time point, a second time point, and a third time point will identify changes in the methylation profile between the first and second time points, between the second and third time points, between the first and third time points, or a combination thereof.

[0270] In one embodiment, a method for determining a DNA methylation profile in a subject over time comprises quantifying the number of methylated molecules in cfDNA (plasma) and gDNA (buffy coat) at the at least ten target loci using a QCT molecule counting technique (e.g., as described in U.S. Patent Publication No. 2019 / 0211395A1, which is incorporated by reference), wherein each of the at least ten target loci is selected due to an expected increase in DNA methylation at each locus in cancer tissue compared to normal (e.g., non-cancerous) tissue. In such an embodiment, the number of methylated molecules of cfDNA is quantified as described herein (see Section 6.7), and the number of methylated molecules of gDNA is quantified as described herein (see Section 6.7).

[0271] In some embodiments, a method for determining a DNA methylation profile in a subject over time comprises adding a spike of known sequence and amount to the sample prior to the processing step at each time point. The spike comprises a known sequence having unmethylated cytosines that are converted to uracils after being subjected to the processing step, which enables calculation of bisulfite conversion yield and conversion efficiency, thereby establishing a bisulfite conversion QC metric that can be used to determine whether a sample has failed the bisulfite conversion step.

[0272] In some embodiments, the method for determining the DNA methylation profile in a subject over time includes a step of processing the methylated sequence reads by one or more of the following: filtering out selected hypermethylated target loci (e.g., filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold); or subtracting background methylation (e.g., subtracting the number of methylated molecules as measured in the buffy coat from the number of methylated molecules in cfDNA, which can be performed on a per-locus basis). In some embodiments, processing is performed prior to the step of quantifying the number (e.g., absolute number) of methylated molecules in the sample. In some embodiments, filtering hypermethylated target loci, subtracting background methylation, or both are performed prior to quantifying the number of methylated molecules in the sample.

[0273] In some embodiments, the method for quantifying DNA methylation in a sample includes further comprising the step of aggregating the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in the sample. Aggregation may include aggregating target loci containing less than a threshold amount of methylated molecules in the buffy coat sample.

[0274] In some embodiments, the method for determining the DNA methylation profile in a subject over time includes aggregating the number of methylated molecules on at least ten target loci (or a subset thereof, such as at least two target loci); and standardizing the number of methylated molecules from all or a subset of the at least 10 target loci by the methylated molecules of the at least one standardized locus. In some embodiments, aggregation and standardization result in a quantification of DNA methylation (e.g., a tumor methylation score) representing the standardized sum of methylated molecules (e.g., at the target loci of at least ten hypermethylations in a sample at a given time point). Comparison across time points enables determination of the DNA methylation profile in a subject (over time). In some cases, the DNA methylation profile identifies changes in the methylation profile from a first time point to a second time point, from a second time point to a third time point, or from a first time point to a third time point.

[0275] In some embodiments, a method for determining a DNA methylation profile in a subject over time comprises assigning a measure of change in the methylation profile as follows: increase, decrease, or no change based on a comparison to a significance threshold. The significance threshold can be predetermined or dynamically calculated. For example, but not limited to, a method for determining a DNA methylation profile in a subject over time indicates a change, wherein the change is an increase because the DNA methylation profile (e.g., a tumor methylation score TM) exceeds a predetermined significance threshold by about 15% compared to a previous time point. For example, but not limited to, a method for determining a DNA methylation profile in a subject over time indicates a change, wherein the change is a decrease because the DNA methylation profile (e.g., a tumor methylation score TM) exceeds a predetermined significance threshold by about 15% compared to a previous time point. TM ) exceeds a predetermined significance threshold by about 15% compared to a previous time point. In some embodiments, a change in a DNA methylation profile (e.g., a tumor methylation score) compared to a previous time point is considered to be an increase or decrease of 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, or 1% above a predetermined significance threshold.

[0276] In some embodiments, a change in the methylation pattern indicates a change in a tumor in a subject. Changes in a tumor include changes in tumor size, changes in the abundance of somatic mutations associated with a tumor, the presence of new somatic mutations associated with a tumor, chromosomal abnormalities associated with a tumor, or a tumor that is resistant to treatment. In some embodiments, a DNA methylation profile (e.g., a tumor methylation score) is used to determine whether the tumor is resistant to treatment. TM ) compared to the DNA methylation profile at a previous time point by at least 1.5-fold (e.g., 2-fold, 3-fold, 4-fold, or 5-fold) indicates that a change has occurred in the tumor. In some embodiments, the DNA methylation profile (e.g., tumor methylation score TM) is at least 2-fold greater than the DNA methylation profile at a previous time point, indicating that a change has occurred in the tumor. In some embodiments, when the DNA methylation in the sample includes at least 10% (e.g., at least 15%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, or at least 90%) of DNA methylation originating from newly methylated target loci (e.g., target loci that were not found to be methylated at a previous time point), the DNA methylation profile identifies that a change has occurred in the tumor (e.g., a tumor associated with the sample). In some embodiments, when the DNA methylation in the sample includes at least 40% of the DNA methylation in the sample (e.g., methylation at the target loci in cfDNA) originating from newly methylated target loci (e.g., target loci that were not found to be methylated at a previous time point), the DNA methylation profile identifies that a change has occurred in the tumor (e.g., a tumor associated with the sample). In some embodiments, a DNA methylation profile (e.g., a tumor methylation score) is a DNA methylation score that is greater than or equal to 10%. TM ) increased by at least 2-fold compared to the DNA methylation profile at the previous time point and at least 40% of the DNA in the sample was methylated (e.g., tumor methylation score TM 40% of the total methylation rate (eg, 40%) is from a newly methylated target locus (eg, a target locus that was not found to be methylated at a previous time point) indicates that a change has occurred in a tumor (eg, a tumor associated with the sample).

[0277] In one embodiment, the DNA methylation profile determined using the methods described herein can be incorporated into clinical recommendations. Clinical recommendations can include plans for further testing or treatment: For example, based on the quantified DNA methylation in the sample and the clinical significance of the DNA methylation, a plan for further testing or treatment can be formulated. In some cases, this may involve additional testing to confirm the diagnosis, monitor disease progression or recurrence, or prescribe targeted therapies specific to the subject's DNA methylation profile.

[0278] 6.10 Quantifying the abundance of somatic mutations in a sample

[0279] The present disclosure also relates to methods for determining a confidence indication of the presence or absence of a somatic mutation in a sample, wherein the confidence indication of the presence or absence of a somatic mutation in a sample is assigned a "true call" or "no call" based in part on the quantification of DNA methylation in the same sample from which the presence or absence of the somatic mutation was determined. In one example, the consistency between the presence or absence of a somatic mutation and DNA methylation indicates that methylation (i.e., methylation analyzed and identified according to the methods described herein) can be used to provide information for the confidence level of the presence or absence of a somatic mutation, which in turn can be used for clinical recommendations to subjects.

[0280] In one embodiment, a method for determining a confidence indication of the presence or absence of a somatic mutation in a sample containing a DNA sequence comprises one or more of the following steps. In some embodiments, the method includes a step of determining the presence or absence of a somatic mutation in the sample. In other embodiments, the method does not include a step of determining the presence or absence of a somatic mutation in the sample because the processing step has been completed before using the method provided herein. In such a case, the method includes quantifying DNA methylation in a sample for determining the presence or absence of a somatic mutation, but determining the presence or absence of a somatic mutation is performed before using the method described herein.

[0281] In some embodiments, a method for determining a confidence indication of the presence or absence of a somatic mutation in a sample containing a DNA sequence includes: quantifying DNA methylation in a sample, which includes steps as described herein (see, e.g., Section 6.7). In some cases, the method includes optionally treating the sample to encode the presence or absence of DNA methylation in a DNA sequence, wherein the sample comprises at least ten target loci from a DNA sequence, wherein each locus of the at least ten target loci is selected based on an expected increase in DNA methylation at each locus in a cancer tissue compared to a non-cancerous tissue, and the sample optionally comprises a standardized locus (e.g., a locus that is highly methylated in both cancer tissue and normal (e.g., non-cancerous) tissue). In some embodiments, the sample is treated to encode the presence or absence of DNA methylation in a DNA sequence before performing the method described herein. In some embodiments, the method includes adding a collection of synthetic molecules to the sample, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches the corresponding nucleotide sequence of an endogenous target, and a variant region having a nucleotide sequence that does not match the sequence region of the endogenous target molecule. The method also includes generating a co-amplification mixture, the co-amplification mixture comprising a set of amplified synthetic molecules and a set of amplified at least ten target loci from a DNA sequence; and sequencing the co-amplified mixture (e.g., deep sequencing with at least one read sequence per molecule in the sample to generate sequence reads). The method also includes determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules.

[0282] After determining the presence or absence of somatic mutations and the amount of DNA methylation in the sample, the method returns such a confidence indication (i.e., the confidence of the determination of the presence or absence of somatic mutations): when the number of methylated molecules in the sample is higher than a predetermined or dynamically calculated threshold, a true determination result of the presence or absence of somatic mutations is returned; or when the number of methylated molecules in the sample is at or below a predetermined or dynamically calculated threshold, a no determination result of the presence or absence of somatic mutations is returned, thereby truly determining the confidence of the determination of the presence or absence of somatic mutations.

[0283] In one embodiment, the method for determining the confidence indication of the presence or absence of somatic mutations in a sample includes using QCT molecular counting technology (e.g., as described in U.S. Patent Publication 2019 / 0211395A1, which is incorporated by reference) to quantify the number of methylated molecules in cfDNA (plasma) and gDNA (buffy coat) at the at least ten target loci, wherein each locus in the at least ten target loci is selected due to the expected increase in DNA methylation at each locus in cancer tissue compared to normal (e.g., non-cancerous) tissue. In such an embodiment, the number of methylated molecules of cfDNA is quantified as described herein (see Section 6.7), and the number of methylated molecules of gDNA is quantified as described herein (see Section 6.7). The quantification of the number of methylated molecules provides a mechanism by which the determination of the presence or absence of somatic mutations can be given a true determination (e.g., a correct determination) or no determination (e.g., DNA is insufficient to give sufficient confidence to the abundance measurement). For example, quantifying DNA methylation in the same sample used to determine the presence or absence of somatic mutations will return a confidence indication for the determination of somatic mutations. When the number of methylated molecules in the sample is above a predetermined or dynamically calculated threshold, a true determination result for the presence or absence of somatic mutations is returned. When the number of methylated molecules in the sample is at or below a predetermined or dynamically calculated threshold, a no determination result for the presence or absence of somatic mutations is returned. No determination indicates that sufficient confidence cannot be given to the determination of the presence or abundance of somatic mutations to say that the measurement is correct.

[0284] In some embodiments, a method for determining a confidence indication of the presence or absence of a somatic mutation in a sample comprises adding a spike of known sequence and amount to the sample prior to the processing step. The spike comprises a known sequence having unmethylated cytosines that are converted to uracil after being subjected to the processing step. As described above, this enables calculation of the bisulfite conversion yield and conversion efficiency, thereby establishing a bisulfite conversion QC metric that can be used to determine whether a sample has failed the bisulfite conversion step.

[0285] In some embodiments, the method for determining a confidence indication of the presence or absence of a somatic mutation in a sample includes a step of processing the methylated sequence reads by one or more of the following: filtering out selected hypermethylated target loci (e.g., filtering out target loci where the total number of methylated molecules in the buffy coat is above a threshold); or subtracting background methylation (e.g., subtracting the number of methylated molecules as measured in the buffy coat from the number of methylated molecules in cfDNA, which can be performed on a per-locus basis). In some embodiments, processing is performed prior to the step of quantifying the number (e.g., absolute number) of methylated molecules in the sample. In such a case, filtering hypermethylated loci, subtracting background methylation, or both are performed prior to quantifying the absolute number of methylated molecules in the sample.

[0286] In some embodiments, when a true call is returned and in the case where the presence of somatic mutations indicates the presence of cancer, the method includes performing a repeat treatment selection assay on the subject or repeating the treatment selection assay (e.g., the treatment selection assay includes genomic analysis to detect new somatic mutations, the abundance of somatic mutations, or both). In some embodiments, when a no call is returned, the method is repeated at least in part on a different sample taken from the same subject.

[0287] In some embodiments, the method for determining a confidence indication of the presence or absence of a somatic mutation in a sample includes further comprising the step of aggregating the number of methylated molecules on at least two of the at least ten target loci to quantify DNA methylation in the sample. Aggregation may include aggregating target loci containing less than a threshold amount of methylated molecules in the buffy coat.

[0288] In some embodiments, a method for determining a confidence indication of the presence or absence of a somatic mutation in a sample comprises aggregating the number of methylated molecules at the at least ten target loci (or a subset thereof, such as at least two target loci); and normalizing the number of methylated molecules from all or a subset of the at least 10 target loci by the methylated molecules of the at least one normalizing locus.

[0289] 6.11 Kit

[0290] In another aspect, the disclosure relates to a kit for quantifying DNA methylation in a sample using the methods described herein. In some embodiments, a kit for quantifying DNA methylation in a sample using the methods described herein. In some embodiments, the kit may include one or more of the following: conversion reagents required to encode the presence or absence of DNA methylation in a DNA sequence, synthetic molecules (e.g., QCT molecules) added to the sample at various points in sample processing, reagents required to amplify the DNA sample, and reagents required to prepare a sequencing library (such as those required for indexing PCR).

[0291] 7. Examples

[0292] 7.1 Materials and Methods for Examples 1-10

[0293] 7.1.1 Tumor Hypermethylation Target Selection

[0294] The Cancer Genome Atlas (TCGA) was queried for human methylation data on tumors and normal tissues of the same tissue type of the subject. The Infinium HumanMethylation450BeadChip (Illumina, San Diego, California) was used to collect data from TCGA and provide beta values ​​representing the methylation scores at specific CpG positions in the genome. In addition, methylation data of leukocytes collected from patients with similar age to cancer patients (average = 63.9 years, sd = 13.3 years) were obtained from GEO accession number GSE40279 (see Hannum, G. et al., (2013). Molecular Cell, 49 (2), 359-367.doi.org / 10.1016 / j.molcel.2012.10.016).

[0295] Tumor hypermethylation at each CpG site was calculated by subtracting normal tissue beta from tumor beta. To avoid selecting CpG sites with spurious hypermethylation, the average hypermethylation was calculated for each CpG island, and CpG islands were ranked in order of highest hypermethylation. Additionally, CpG islands were filtered so that the average leukocyte beta was < 0.2 to minimize the effect of background signal from the buffy coat on cfDNA. Several CpG positions were selected from each selected CpG island for primer design.

[0296] Lung cancer-specific assays were designed using TCGA data from subjects with lung adenocarcinoma and lung squamous cell carcinoma. Pan-cancer assays were also designed using TCGA data from many cancer types (e.g., lung adenocarcinoma, lung squamous cell carcinoma, colon adenocarcinoma, breast invasive carcinoma, pancreatic adenocarcinoma, liver hepatocellular carcinoma, bladder urothelial carcinoma, esophageal cancer, kidney renal cell carcinoma, kidney renal papillary cell carcinoma, prostate adenocarcinoma, thyroid cancer, and uterine corpus endometrial carcinoma).

[0297] 7.1.2 Selection of highly methylated targets

[0298] In addition to selecting targets with high methylation in tumors compared to normal tissues, control loci in the buffy coat that were highly methylated in both tumors and normal tissues were selected. Acquisition and analysis of methylation data were similar to those for high methylation targets, but were sorted in order of highest leukocyte methylation signal and filtered to have leukocyte beta > 0.9, and tumor and normal tissues of six cancer types (lung adenocarcinoma, lung squamous cell carcinoma, colon adenocarcinoma, breast invasive carcinoma, pancreatic adenocarcinoma, liver hepatocellular carcinoma) were selected.

[0299] 7.1.3 Primer and QCT design

[0300] Primer pairs targeting the genomic positions of the selected CpG positions were designed in silico using Primer3 and using a bisulfite-converted reference genome. This process was performed under the assumption that the reference hg19 human genome was fully methylated, with all Cs converted to Ts except for CpGs. The ideal annealing temperature was set to 60°C.

[0301] Single-stranded QCTs were designed for each amplicon. The predicted amplicon sequence for each primer pair was determined using Bowtie and the transformed human genome, and 17 bases of flanking genomic sequence were added to the 5' and 3' ends. Eleven bases in the insert region of the QCT were replaced with N, allowing a small number of unique QCTs to be added to each PCR reaction.

[0302] 7.1.4 Additional Amplicons

[0303] In addition to tumor hypermethylation and highly methylated genomic targets, other amplicons are also included in the assay. For A>G or G>A SNP common according to the single nucleotide polymorphism database (dbSNP) of NCBI, 15 genotyping amplicons are designed. The genotype at these positions can be used to check the sample exchange that may occur during laboratory processing or in the clinic. Include three amplicons for genomic positions on the Y chromosome to determine the gender of the patient.

[0304] Two artificial amplicons are included to check the bisulfite conversion yield and conversion efficiency. These synthetic oligonucleotides are spiked into cfDNA and gDNA samples in a fixed amount immediately before bisulfite conversion. These oligonucleotides are then amplified in multiple reactions together with all other amplicons. During the analysis, the initial number of spike-in molecules and the percentage of cytosine bases converted to thymine bases can be calculated (these oligonucleotides are synthesized without any methylation). For multiple PCR no-template control, the same amount of bisulfite conversion spike-in is added after bisulfite conversion, so that the overall bisulfite conversion yield can be estimated. Calculating the bisulfite conversion yield and conversion efficiency allows the establishment of bisulfite conversion QC metrics and therefore determines whether the sample has not passed the bisulfite conversion step of the process.

[0305] 7.1.5 Sample Source

[0306] To obtain a large number of tumor samples covering multiple cancer types, banked flash-frozen tumors and buffy coats from the same subjects were obtained from Spectrum Health (Grand Rapids, MI).

[0307] To test whether the assay can detect methylation changes that are consistent with clinical outcomes, samples were collected retrospectively and prospectively from cancer patients. Retrospective samples consisting of banked plasma and buffy coat samples were obtained through collaborations with the University of California, San Diego and the University of Florida. Prospective sample collection was performed by a contract research organization and its partner clinics; patients diagnosed with cancer but who had not yet started treatment were recruited. Blood was collected using Streck tubes before treatment and at subsequent time points after treatment. Clinical outcomes are provided when available.

[0308] To test the effect of the assay on healthy individuals, blood was collected from healthy volunteers at different time points.

[0309] 7.1.6 Sample preparation

[0310] Blood collected in EDTA tubes was centrifuged within one hour of collection, and plasma and buffy coat were separated. Blood collected in Streck tubes was allowed to stand overnight, then centrifuged and plasma and buffy coat were separated. The plasma volume was recorded for standardization during analysis.

[0311] cfDNA was extracted using the QIAamp circulating nucleic acid kit (Qiagen), and gDNA was extracted from tumor and buffy coat samples using the DNeasy blood and tissue kit (Qiagen).

[0312] Artificial samples that mimic cell-free DNA (cfDNA) samples from cancer patients were created by mixing tumor genomic DNA (gDNA) into buffy coat gDNA at various tumor fractions.

[0313] 7.1.7 Bisulfite conversion and next-generation sequencing library preparation

[0314] Samples were bisulfite converted using Diagenode Premium Bisulfite Kit (Catalog No. C02030030) or Zymo EZ-96 DNA Methylation-Lightning MagPrep Kit (Catalog No. D5046). When the cfDNA or gDNA sample volume was greater than the recommended input volume, the sample was divided into two halves, converted separately, and then re-merged. Enzymatic conversion was also tested. As a positive control for detecting methylation, universally methylated genomic DNA (Catalog No. S7821, Sigma-Aldrich) was diluted in the buffy coat with various tumor fractions.

[0315] Primer mixes were created by combining all primer pairs and iteratively removing and / or rebalancing the concentration of each primer pair to optimize balanced coverage across target amplicons. QCT was diluted to 200 molecules per PCR reaction at each amplicon. The lung cancer assay included a total of 113 target amplicons and the pan-cancer assay included a total of 679 target amplicons.

[0316] Multiplex PCR was performed on bisulfite-converted samples using Q5U polymerase (NEB, Ipswich, MA). Subsequently, PCR indexing was performed using Q5 polymerase (NEB, Ipswich, MA) to allow sequencing of multiple samples with dual indexing in the same sequencing run. Pooled libraries were bead-cleaned and loaded onto a NextSeq 2000 (Illumina, San Diego, CA) sequencer using P3100 cycling reagents and 5% PhiX for unidirectional sequencing.

[0317] 7.1.8 Calculating the number of methylated molecules

[0318] Fastq files were adapter trimmed on the 3’ end using BBDuk and then mapped using BWA-MEM to a custom gene set consisting of the target hypermethylated amplicon and the highly methylated amplicon and QCT sequences.

[0319] For each amplicon, reads mapped to the target amplicon (e.g., target hypermethylated amplicon or highly methylated amplicon) are binned based on sequence. The number of CpGs contained in each sequence is calculated, and each sequence is classified as belonging to a methylated read if the number of CpGs is greater than or equal to the maximum possible number of CpGs for the amplicon minus one. Reads mapped to the corresponding QCT sequences are treated separately and binned based on the random N sequence of the QCT. The average number of reads for each QCT molecule is calculated assuming that the sequence of each QCT molecule is unique in each reaction. The total number of methylated molecules for the amplicon can then be calculated by dividing the total number of methylated reads by the average reads for each QCT molecule.

[0320] When measurements are obtained for paired cfDNA and buffy coat samples, background methylation levels are reduced by subtracting the buffy coat methylation signal from the cfDNA methylation signal on a per-locus basis. Based on methylation pattern estimates, approximately 55% of cfDNA is of leukocyte origin (see Moss, J. et al., (2018), Nature Communications, 9(1), doi.org / 10.1038 / s41467-018-07466-6), which suggests that subtracting all buffy coat methylation is a conservative approach to removing background methylation signals. Any negative calculated molecular number after background subtraction is set to zero.

[0321] In order to perform comparable subtraction between the buffy coat and cfDNA samples, the sample must first be standardized based on the input amount of the genome equivalent. The average value of the methylated molecules of the highly methylated loci is used to estimate the total number of genome equivalents in the sample overall. The number of methylated molecules at each locus is standardized based on the estimated genome equivalent of the sample. The standardized number of methylated molecules is then used for background subtraction.

[0322] 7.1.9 Filtering of Hypermethylated Loci

[0323] Based on the test to healthy subjects, it is found that some high methylation loci still have significant methylation signals even after the buffy coat is subtracted. In order to minimize the amount of false cancer signals, some loci are filtered out from the analysis. In some cases, this is referred to as "blacklist" in this article. Any loci that are found to continuously contribute moderate methylation or occasionally contribute high methylation in healthy subjects are added to the list of loci to be ignored. These loci usually contain high methylation in the buffy coat.

[0324] After filtering the hypermethylated loci, a noise floor was established using the results from healthy subjects, below which the methylation signal was uninterpretable because it was comparable to the amounts observed in healthy subjects.

[0325] 7.1.10 Determination of methylation changes

[0326] When data from two time points are available, a determination is made to see whether the amount of methylated molecules has increased, decreased, or remained unchanged. The determination is made by modeling the measured values ​​of each time point as a normal distribution using the mean value of the number of methylated molecules measured (standardized based on the total number of input molecules). A standard deviation is assigned to a normal distribution based on the total number of methylated molecules; this can be achieved by interpolating based on previous studies, where the coefficient of variation is measured from an artificial sample (see, e.g., Arm 3 of Example 7). The normal distribution of the first time point is subtracted from the normal distribution of the second time point and standardized based on the mean value of the normal distribution of the first time point, thereby creating a standardized difference normal distribution. If the mean value of the distribution is ≥15% and the log2 likelihood ratio of the difference ≥15% to <15% is greater than 3, an "increase" determination is made. If the mean value of the distribution is ≤-15% and the log2 likelihood ratio of the difference ≤-15% to >-15% is greater than 3, a "decrease" determination is made. If the amplitude of the relative difference is not large enough or the statistical likelihood is not strong enough, a "no change" determination is made. If any time point is below the noise floor, the mean value of that time point is set to the noise floor, the standard deviation is still calculated based on the total number of methylated molecules, and a determination is still made. If the methylation signals from both time points are below the noise floor, an "uncertain" determination is made. Any other suitable statistical method can be used to make a determination, including but not limited to mean, median, standard deviation, likelihood ratio, expectation maximization, statistical significance test.

[0327] 7.1.11 Methylation profile

[0328] To take advantage of the fact that the assay measures methylation at hundreds to thousands of hypermethylated positions, the methylation profile is determined at each time point. One method of constructing the methylation profile is to classify the loci based on the time point at which the locus was first found to have an informative methylation signal. An informative methylation signal is defined as greater than 2 molecules after background subtraction. Classifying the loci by the first informative time point reveals when the methylation profile changed and how much it changed.

[0329] 7.2 Example 1. Lung cancer assay can detect significant methylation in 0.5% tumor fraction artificial samples

[0330] The detection limit of lung cancer assay was evaluated using artificial samples. Without wishing to be bound by theory, the preparation of artificial samples simulates clinical samples as much as possible. For these experiments, artificial samples included sheared tumor gDNA, which was sheared to an average fragment length of ~170bp using an ultrasonicator to simulate the size distribution of cfDNA. Artificial samples made with universal methylation were used as positive controls (e.g., replacing tumor DNA as a positive control). The positive control was sheared in the same manner as artificial lung cancer samples. A total of 1.35E9 sequencing reads were obtained, with an average of 11E6 average reads per sample. There were an average of 62 reads per QCT in all samples and amplicons. Each artificial sample was created using 5000 genome equivalents (ge).

[0331] Technical replicates of human lung cancer samples were tested at 0%, 0.5%, 1%, 2.5% and 5% tumor fractions (see Figure 2A ). The assay was able to distinguish samples with 0.5% tumor fraction from samples with 0% tumor fraction, indicating a detection limit of at least 0.5%.

[0332] The lung cancer assay was also used to measure methylation in additional lung tumor samples (see Figure 2B ). Figure 2B The use of this lung cancer assay to detect methylation in human samples from two additional lung cancer subjects is shown. For these experiments, different PCR volumes and different thermal cycling protocols were used (see Figure 2B ).

[0333] Additional analyses using the lung cancer assay included measurement of Figure 1A The consistency value (CV) of the samples analyzed in ( Figure 3 ). CV is used to determine whether the amount of ctDNA and, accordingly, the tumor is truly increased or decreased or whether the increase or decrease is a false positive or false negative. Without any additional post-processing, the CV of the methylation measurement is around 10% or less. This indicates that the methylation measurements generated using the method described herein reliably identify a 30% change in the signal within 3 standard deviations. According to the RECIST guidelines, a partial response is defined as a reduction of at least 30% in the longest diameter of the target tumor lesion, and progressive disease is defined as an increase of at least 20% in the longest diameter of the target tumor lesion. In view of these RECIST guidelines, the assay described herein is capable of assessing the response of a tumor to treatment in the context of current standard care. In fact, quantifying treatment response according to the RECIST guidelines is limited by the selection of a limited number of tumors that can be identified by imaging, while methylation assays will provide measurements of all tumors present in the body that release DNA into the bloodstream.

[0334] 7.3 Example 2. Background subtraction to improve signal-to-background ratio

[0335] Optimization is performed to improve the signal to background ratio. Without wishing to be bound by theory, improving the signal to background ratio is particularly useful for detecting smaller tumor signals. For example, one medical application is for minimal residual disease (MRD) detection, where the difference between having any tumor signal and zero signal can be the difference between relapse and remission.

[0336] Although targets with low methylation in the buffy coat and normal tissues are filtered during target selection, some targets may still have significant amounts of background methylation. One approach to reduce the impact of background methylation is to mask the signal from target loci with buffy coat methylation above a certain threshold. Another parallel approach is to subtract the methylation signal as measured in the buffy coat from the methylation signal measured in cfDNA (in plasma) on a per-locus basis.

[0337] Figure 4 Total normalized methylated molecules for technical replicates of lung cancer artificial samples at different tumor scores are shown. In particular, Figure 4 Data from a lung cancer assay is shown, which includes background subtraction (i.e., subtracting the methylation signal as measured in the buffy coat from the methylation signal measured in cfDNA) and masking target loci with high background methylation (i.e., masking the signal from the target loci with buffy coat methylation above a certain threshold). For example, the background methylation signal is calculated by using the buffy coat methylation signal for each locus, normalizing relative to the average of the highly methylated loci methylated molecules in each buffy coat sample (e.g., the average of the standardized loci), and averaging across all buffy coat samples. 55% of the background signal is subtracted from the methylation signal for each cfDNA artificial sample, normalized relative to the average of the highly methylated loci methylated molecules in each artificial sample. For every 1000 highly methylated loci methylated molecules, highly methylated target loci with more than 10 methylated molecules are masked. Any other suitable threshold may be determined.

[0338] By masking loci and subtracting the buffy coat methylation signal from the cfDNA signal of the artifact sample, the background signal was reduced from an average of 277 methylated molecules to 111 methylated molecules in the 0% tumor fraction artifact sample. More stringent masking of hypermethylated target loci is expected to further reduce the background signal.

[0339] Overall, the data suggest that using background subtraction (i.e., subtracting the methylation signal as measured in the buffy coat from the methylation signal measured in cfDNA) while masking target loci with high background methylation (i.e., masking signals from target loci where buffy coat methylation is above a certain threshold) can improve the signal-to-background ratio.

[0340] 7.4 Example 3. Unlike somatic mutation-based methods, hypermethylation-based measurements are robust to gDNA contamination

[0341] A common problem with cfDNA assays is gDNA contamination. This can occur during the plasma separation process of centrifuged whole blood when the buffy coat accidentally separates with the plasma fraction. gDNA contamination can adversely affect the accuracy of cfDNA tumor quantification, especially when using somatic variant allele fractions for cfDNA tumor quantification. This is because the additional buffy coat affects the denominator. Figure 5 As shown in (top panel), 1x gDNA contamination (e.g., contamination that doubles the total number of molecules) reduces the variant allele fraction by half. This can lead to false positive signals. When positive signals are associated with treatment efficacy, such results have the potential to change medical decisions.

[0342] This experiment shows that the detection of tumor hypermethylation is less susceptible to gDNA contamination than the somatic mutation ctDNA assay. This is at least because of the selection of targets with low buffy coat methylation and the use of background subtraction methods. If the background methylation signal in the buffy coat is zero, the absolute number of methylated molecules is completely unaffected by gDNA contamination ( Figure 5 ; middle panel). Even though the buffy coat does contribute a small amount of methylation background signal ( Figure 5 ; Bottom panel), in the case of 1x amount of gDNA contamination, measuring buffy coat methylation and subtracting the signal from the cfDNA signal can also keep the total methylation signal within a 20% increase. This will minimize the impact of gDNA contamination on tumor methylation measurements and any potential downstream medical decisions to be made based on tumor methylation measurements.

[0343] In addition to target selection and background subtraction, there are additional biochemical methods to minimize gDNA contamination, including bead purification to select the size of DNA fragments in cfDNA samples. However, since these methods can also fail, it would be helpful to adopt an orthogonal method to achieve robustness to gDNA contamination.

[0344] 7.5 Example 4. Day-to-day variability of buffy coat methylation in healthy subjects is comparable to subject-to-subject variability

[0345] The extent of day-to-day variability in methylation in healthy subjects was assessed as part of the assay's ability to make accurate time-series measurements. Because a significant portion of the cfDNA signal is of buffy coat origin, methylation profiles were measured in buffy coat from healthy subjects from different tubes of the same blood draw, different tube types, different days, and different subjects. Figure 6The data in Figure 2 show the methylation profiles of buffy coat (i.e., 5000 ge of buffy coat) isolated from different tubes, different tube types, different days, and different subjects from the same blood draw (about Figure 6 Sample identifiers in , see Table 1).

[0346]

[0347]

[0348]

[0349]

[0350]

[0351] For example, the two tubes collected from subject 5 on the same day using the same tube type were very similar, which was expected (see Figure 6 ). Tube type did not cause significant differences, as the EDTA tube was very similar to both Streck tubes. The data from the Streck tube collected on Day 2 was similar to the data from the Streck tube collected on Day 1, but there were several loci with noteworthy differences. There was significant subject-to-subject variability, as Subject 10 had several methylated loci that were not methylated in any tube from Subject 5.

[0352] Figure 7 Hierarchical clustering of methylation profiles from buffy coat isolated from different tubes, different tube types, different days, and different subjects from the same blood draw is shown. Clustering is based on the methylated molecules at each target hypermethylated locus, normalized relative to the amount of methylated molecules measured at the hypermethylated locus. Cluster distances are calculated using the L1 norm.

[0353] Hierarchical clustering revealed that tubes collected on the same day clustered most closely together, regardless of tube type (see Figure 7 ). Examples of tubes from the same day that clustered together include Subject 1 Day 1, Subject 3 Day 1, Subject 3 Day 2, and Subject 10 Day 1. However, tubes collected from the same subject but on different days were clustered just as far apart as tubes collected from different subjects. This suggests that background subtraction is optimal if performed using the methylation profile of the buffy coat collected at each unique time point.

[0354] 7.6 Example 5. Pan-Cancer Assay Detects Methylation Signals in Multiple Cancer Types

[0355] A pan-cancer assay was designed based on the methods described herein.

[0356] Figure 8 Data showing methylation signals in a pan-cancer assay in artificial samples of several different cancer types. For these experiments, a 5% tumor fraction artificial sample was made by mixing 5% tumor gDNA by mass with 95% buffy coat from the same subject. The 0% tumor fraction artificial sample was pure buffy coat. Samples were labeled with their cancer type (BRCA: breast invasive carcinoma, COAD: colon adenocarcinoma, LIHC: liver hepatocellular carcinoma, LUNG: lung cancer, PAAD: pancreatic adenocarcinoma). A universally methylated sample was included as a positive control for the methylation assay.

[0357] like Figure 8 As shown in , the Pan-Cancer assay is able to detect methylation signals in 5% tumor fraction artificial samples of several cancer types using the same chemistry for all samples. There is a noticeable variability in the methylation signals across samples, which may be due to several reasons. First, the methylation profile of a tumor may vary from tumor to tumor. The profile may or may not match the target positions included in the assay. In the assay described in this article, increasing the number of target positions can help minimize the effects of biological variability in methylation across tumors. This variability may also be due to technical limitations in making artificial samples from tumor samples, as the tumor itself contains an unknown proportion of normal tissue. Therefore, comparisons of measurements made on artificial samples of the same tumor fraction across samples should be made with caution. Although the exact tumor mass is unknown, the assay is still able to distinguish between 5% tumor fraction artificial tumor samples and pure white film layer samples from the same subject.

[0358] Overall, these data demonstrate that the approach described herein includes pan-cancer detection capabilities.

[0359] 7.7 Conclusions from Examples 1-5

[0360] These data demonstrate an innovative approach for cancer treatment monitoring by quantifying the absolute number of methylated molecules. Existing ctDNA assays have poor accuracy for treatment monitoring because the number of variants that can be detected is very limited and therefore the total number of variant molecules is very small. Whole exome sequencing of tumor biopsies can be performed to identify additional variants and overcome molecular sampling limitations, but this requires tumor biopsies, which may be difficult to obtain from patients or may not have sufficient clinical surplus material. By addressing several technical shortcomings of existing treatment monitoring methods, we believe that this methylation-based, tumor-agnostic, multiplexed quantitative assay will be of great benefit to oncologists and their patients in guiding treatment decisions.

[0361] 7.8 Example 6. Filtering of Hypermethylated Loci

[0362] The lung cancer assay described in Example 2 uses background subtraction (which includes subtracting the methylation signal as measured in the buffy coat from the methylation signal measured in cfDNA) while masking target loci with high background methylation (i.e., masking the signal from target loci with buffy coat methylation above a certain threshold) to improve the signal-to-background ratio. However, masking target loci with high background methylation has proven to be an effective but blunt tool that requires further optimization. An attempt to further improve the signal-to-background ratio using masking of target loci is described below.

[0363] Attempts to improve the signal-to-background ratio using masking of target loci involved analysis of 16 samples from 10 independent healthy subjects with the goal of identifying loci that, when masked (filtered out), would reduce the signal-to-background ratio.

[0364] The pan-cancer assay described herein measures methylation at loci that are typically hypermethylated in cancer. One limitation of this assay is that not all methylation signals are cancer signals. This may be due to loci with non-zero amounts of background methylation (e.g., methylation in non-cancer subjects).

[0365] As mentioned above, limiting the amount of background methylation has an impact on accurate treatment monitoring. Without wishing to be bound by theory, randomly present methylation may introduce false signals, which may lead to incorrect clinical interpretations. A large amount of background methylation (e.g., equal to or greater than 20 molecules) will also increase the noise floor (e.g., the total amount of methylation observed in subjects without cancer). A large amount of background methylation will also affect the ability to detect relative changes in cancer over time, because the background is likely to exist at each time point.

[0366] Previous studies established a blacklist of amplicons that contribute significant background methylation (see Example 2). The blacklist was generated by analyzing 4 samples from 4 healthy subjects.

[0367] Although effective in limiting the amount of background methylation, a more comprehensive analysis of non-cancer samples is still needed to better establish an updated amplicon blacklist. As described above, 16 samples from 10 independent subjects that had previously been subjected to methylation assays were analyzed to determine an updated amplicon blacklist. The updated blacklist was applied to the methods described herein and the samples were evaluated for significant changes in clinical interpretation after analysis using the amplicon blacklist.

[0368] Methods. The training set consisted of 16 samples from the 10 independent subjects listed above. The analysis included 8 samples from batch 2, from subjects who did not have cancer but had undergone liver surgery. These 8 samples were collected from four subjects at 2 times. The subjects were 38, 39, 83, and 84 years old. The analysis also included 8 samples from 6 healthy subjects: 2 subjects at 2 time points each and the remaining subjects at 1 time point each.

[0369] To determine the updated amplicon blacklist, two approaches were used: patient- or sample-specific personalized amplicon blacklist and a global amplicon blacklist.

[0370] 7.8.1 Personalized Blacklist

[0371] The personalized approach enables blacklisting of only loci that are determined to be likely to contribute to background methylation signal in a particular sample. A non-limiting example of a "personalized blacklist" or "sample-specific blacklist" includes evaluating the buffy coat methylation signal of loci with high methylated cfDNA signals in non-cancer subjects (see Figures 9A-9F ).

[0372] Other factors in identifying loci to be included in a "personalized blacklist" include determining the age and sex of the patient from whom the sample was collected.

[0373] Fig. 9A An example of a locus used in a "personalized blacklist" is shown: intron cg03134157_77 ("intron 0313"). Fig. 9A As shown in , the tumor QE of intron 0313 shows significantly higher cfDNA methylation than the buffy coat methylation only when there is significant buffy coat methylation. Tumor QE is calculated by subtracting blue from red. The error bar indicates one standard deviation as estimated by the molecular count noise based on the original QE. Overall, the consistency of methylation at intron0313 or almost no methylation at intron 0313 means that the buffy coat methylation threshold for this locus is about 20 tumor QEs. This means that a buffy coat methylation threshold of about 20 tumor QEs is required to correctly exclude this locus. Note that a buffy coat threshold that is too small may risk blacklisting too many loci (essentially becoming a global blacklist).

[0374] The use of the second locus in a “personalized blacklist” was analyzed. Fig. 9BMethylation of introncg20907051_11 ("intron 2090") is shown. Specifically, intron 2090 showed no significant difference in cfDNA methylation in the buffy coat methylation, regardless of the presence or absence of significant buffy coat methylation (see Fig. 9B ). Since the vast majority of healthy samples tested had non-zero tumor QE, intron 2090 is an example of a blacklist locus for all samples.

[0375] Figures 9C-9F Additional loci analyzed for inclusion in the “personalized blacklist” are shown in . Fig. 9C The tumor QE for intron_cg12880300_0 was shown to have an average tumor QE of 59 and 9 of the 16 samples had a tumor QE greater than 2. Since buffy coat methylation was consistently high and cfDNA methylation was variable, this locus is an example of a blacklisted locus for all samples.

[0376] Fig.9D Tumor QE is shown for the locus IRX4_cg13974394_0 ("IRX4"). For the IRX4 locus, significant tumor QE (red greater than blue) was observed in only 1 of the 16 samples. This sample was not flagged during QC. In fact, for this cfDNA sample, the reads per QCT were approximately 4.9, which is lower than the other samples, but the difference is not significant. The data suggest that IRX4 is best assessed on a sample-by-sample basis. Additionally, the data suggest that the high gDNA signal makes any actual tumor QE measurement at this locus relatively imprecise. Furthermore, in this locus, methylation appears to be age-related. BTO-23 and BTO-24 are in their 80s, BTO-21 and BTO-22 are nearly 40, and the vast majority of healthy volunteers are younger.

[0377] Figures 9E-9F Tumor QEs for loci EMID2_cg25290307_0 ("EMID2") and intron_cg11453719_0 are shown, respectively. Fig.9E As shown in , the 10QE (dashed line) filter for gDNA will filter out BTO-21_2, BTO-23_1, and BTO-23_2, but not BTO-24_2. Fig.9F As shown in , the 10QE (dashed line) filter for gDNA will filter out all samples except BTO-24_1 and Healthy_009_1. In these two loci, the personalized blacklist approach requires a relatively stringent 10QE threshold, but even this cannot ignore false methylation signals in several samples. Therefore, it may be best to include these loci on the global blacklist.

[0378] 7.8.2 Global Blacklist

[0379] To determine the criteria for including amplicons on the global blacklist, an empirical approach was used with thresholds set for the mean tumor methylation QE and the maximum tumor methylation QE (described below). The thresholds were applied to determine the blacklist and calculate the new total tumor QE in non-cancer samples. The thresholds were also applied to previous analyses (e.g., UCSD and UF results) to determine whether these results are still clinically valid.

[0380] This analysis focused in part on the top 20 amplicons based on mean tumor methylation QE (see Table 2). In Table 2, the columns indicate the number of samples with tumor QE greater than 2 for that locus (n_samples_gt2), the mean tumor QE (mean_tumor_norm) and maximum tumor QE (max_tumor_norm) for all 16 samples, and the mean buffy coat methylation QE (meanbg_norm).

[0381]

[0382] Based in part on the analysis of the top 20 amplicons as shown in Table 2, the thresholds for the global blacklist included: (1) maximum tumor QE in any non-cancer sample>15; and (2) average tumor QE (including 0s)>2.

[0383] In one non-limiting example, applying the thresholds (1) maximum tumor QE in any non-cancer sample > 15; and (2) average tumor QE (including 0s) > 2 resulted in a global blacklist of 32 amplicons. Compared to the previous blacklist (see Table 3), there were 16 amplicons that overlapped, meaning that 6 amplicons were removed and 16 amplicons were added.

[0384]

[0385]

[0386] The new global blacklist generated based on the above thresholds (i.e., (1) maximum tumor QE in any non-cancer sample > 15; and (2) average tumor QE (including 0s) > 2) was applied to the 16 non-cancer samples to determine the noise floor. Fig.10 As shown in , no sample had methylation of greater than 114 tumor QEs. Therefore, the noise floor was set to 120 for additional analysis. However, one skilled in the art will appreciate that the noise floor will vary based on similar types of analysis applied to different samples or different sets of samples.

[0387] The new global blacklist is then compared with the old global blacklist. Fig.11 As shown in , most subjects had equal, if not fewer, tumor molecules when comparing normalized total tumor molecules using the old or new global blacklist. In addition, the methylation results for all 30 subjects were manually reviewed and assessed for changes in clinical interpretation. Only 2 of the 30 subjects had slightly different results, but both were within the range of statistical noise. For subjects 8272 and 6885, the results from the last acquisition changed slightly, but were still within the range of molecular noise error (see Figures 12A-12D ). Overall, these data suggest that small changes in the blacklist do not significantly affect clinical outcomes.

[0388] Then, samples from five patients whose liver cancer had been surgically removed but whose cancer stages were different (see Figures 13A-13E BTO-1 to BTO-5 in the table compare the new global blacklist with the old global blacklist. Figures 13A-13E The trends for each patient were consistent with previous analyses, as shown in Figure 2. In addition, tumor methylation decreased in each patient, except for BTO-1, a patient with stage 4 HCC who had high methylation levels and provided a sample one day after surgery.

[0389] 7.8.3 Conclusion

[0390] The data demonstrate the benefits of a global blacklist relative to a personalized blacklist (patient or sample specific blacklist). Specifically, a global blacklist avoids having to deal with unanswered biological variability and temporal variability issues. However, this analysis does not exclude the use of a personalized blacklist, but additional optimization may be required to design an effective personalized blacklist.

[0391] The data also demonstrate that adjusting the blacklist from 22 to 32 amplicons did not significantly change clinical outcomes beyond measurement noise in both clinical study datasets. Importantly, the noise floor increased from 50 to 120 normalized total tumor QEs.

[0392] 7.9 Example 7. Analysis and Clinical Validation of Pan-Cancer Assay

[0393] The Pan-Cancer Assays provided in this disclosure and described, for example, in Example 5, were subjected to analytical validation, including testing for accuracy, precision, reproducibility, sensitivity, and specificity.

[0394] 7.9.1 Introduction and summary of verification arms

[0395] Accurate, rapid, and accessible treatment monitoring for cancer patients is an unmet medical need. When the outcome of a cancer treatment regimen is uncertain, determining early on whether a treatment is working for a patient can allow them to switch to a different treatment regimen, potentially extending life, reducing unnecessary side effects from ineffective treatments, improving quality of life, and increasing the overall efficiency of the healthcare system. Early assessment of treatment efficacy, or even prediction of ultimate treatment outcome, may be particularly important for advanced cancer patients when time is of the essence or for efficiently conducting clinical studies of new cancer therapies.

[0396] ctDNA (circulating tumor DNA) obtained through liquid biopsies from cancer patients has been shown to reflect the amount of cancer present in the patient. Additionally, methylation has been shown to be a robust biomarker for cancer, and several research groups are developing methylation-based assays that use ctDNA for various cancer diagnostic applications. Using our patented QCT technology (Tsao et al., 2019), the applicants developed a novel assay called the Pan-Cancer Assay to quantify the amount of methylation in ctDNA for accurate and precise treatment monitoring applications.

[0397] A validation study was performed to confirm the analytical and clinical validity of the pan-cancer therapy monitoring assay.

[0398] Analytical validation included: Arm 1: accuracy, precision, and reproducibility on sheared tumor DNA samples; Arm 2: consistency on clinical samples with known clinical outcomes; Arm 3: limit of detection using sheared tumor DNA samples; and Arm 4: diversity across cancer types.

[0399] 7.9.2 Arm 1: Accuracy, Precision, and Reproducibility for Sheared Tumor DNA Samples

[0400] The purpose of arm 1 is to calculate the accuracy, precision and reproducibility based on sheared tumor DNA samples. Replicates of sheared tumor DNA samples with 1% and 2% tumor fractions were made from a total of 7 different tumors from 7 different cancer types. Sheared tumor DNA samples were processed in two different batches and loaded onto two different sequencers. The reaction score for each sample was calculated, which describes the amount of methylation in the sample, and the reaction scores of 1% and 2% tumor fraction samples were compared to each other.

[0401] To summarize and described in further detail below, data from Arm 1 showed: 100% accuracy (i.e., 40 correct calls out of 40 comparisons); 100% precision (i.e., 0 subjects within each batch had inconsistent results); and 100% reproducibility (i.e., 9 out of 19 comparisons were consistent between batches).

[0402] In both batches, all comparisons between the 1% and 2% tumor fraction samples that passed QC were judged as "increased," consistent with the identity of these samples. Response scores from both batches were as follows Fig.14 and Fig.15 as shown in .

[0403] Additional calculations were performed for the sensitivity and specificity of the Pan-Cancer assay. Sensitivity was calculated based on comparing the 1% and 2% samples, as described above, while specificity was calculated by comparing the 1% sample to itself and the 2% sample to itself within each batch. This analysis revealed sensitivity = 100% [95% CI: 91.2%, 100%] (i.e., 40 of 40 comparisons), specificity = 100% [95% CI: 95.5%, 100%] (i.e., 80 of 80 comparisons).

[0404] Overall, results from Arm 1 demonstrated that the Pan-Cancer assay is accurate, precise, and reproducible across different operators and sequencers.

[0405] 7.9.3 Arm 2: Consistency for Clinical Samples with Known Clinical Outcomes

[0406] The purpose of Arm 2 is to evaluate the accuracy of the Pan-Cancer Assay on clinical samples from a total of 20 subjects. These 20 subjects include 8 cancer subjects with known clinical outcomes and 12 healthy subjects with no known history of cancer. Response scores are measured for each subject at two time points. Decisions are made based on the response scores from the two time points and compared to the known clinical outcomes.

[0407] To summarize and further describe in detail below, data from Arm 2 showed: 100% accuracy for cancer subjects (6 out of 6 calls were consistent); and 100% accuracy for healthy subjects (11 out of 11 calls were consistent).

[0408] Of the 6 cancer subjects with detectable signals, the determinations for all 6 subjects were consistent with known clinical findings. The results for all cancer subjects are plotted in Fig.16 The known clinical results can be found in Table 4.

[0409]

[0410] Of the 12 healthy subjects, 11 had no cancer changes or had an indeterminate verdict, which was consistent with the health status of these subjects (see Fig.17 ).

[0411] In summary, the results of Arm 2 confirmed the validity of the Pan-Cancer assay to accurately assess clinical samples from a total of 20 subjects.

[0412] 7.9.4 Arm 3: Detection Limits Using Sheared Tumor DNA Samples

[0413] The purpose of Arm 3 was to evaluate the limit of detection of the Pan-Cancer assay by measuring the sensitivity under different input conditions.

[0414] For these experiments, sheared tumor DNA samples of 0%, 0.25%, 0.5%, 1% and 2% tumor fractions were prepared from a single tumor and its matched buffy coat and analyzed using the Pan-Cancer Assay. For each tumor fraction, the response scores of 16 replicates were measured and the CV was calculated using these measurements. The assay sensitivity was calculated by comparing the response scores of samples of each tumor fraction with the response scores of the 0% sample, from which the detection limit was determined.

[0415] Summarized and described in further detail below, the results showed a limit of detection = 0.25% tumor fraction.

[0416] The detection limit was defined as the lowest tumor fraction with at least 95% sensitivity. For the 0.25% tumor fraction sample, the sensitivity was 96.5% [95% CI: 93.4%, 98.4%]. Fig.18 The sensitivity for each tumor fraction is shown in .

[0417] Fig.19 The response scores for each tumor fraction are plotted in . As expected, higher tumor fractions correspond to higher response scores. Within each set of technical replicates, the CV of the response score was calculated for a given tumor fraction (see Fig.19 There is a clear dispersion between 0% and 0.25% tumor fractions, which is consistent with the 0.25% tumor fraction detection limit (see Fig.19 ). The CV was calculated using the standard deviation and mean of each tumor score.

[0418] In summary, the results of Arm 3 demonstrated that the detection limit of the Pan-Cancer assay was at least 0.25% tumor fraction.

[0419] 7.9.5 Arm 4: Diversity of cancer types

[0420] The aim of Arm 4 was to evaluate the performance of the Pan-Cancer assay on different cancer types. For each of 54 different cancer patients (covering 10 unique cancer types), a tumor gDNA sample and a matched buffy coat gDNA sample were processed.

[0421] For these experiments, bioinformatics simulations were used to evaluate the performance of the Pan-Cancer assay for different cancer types. For example, bioinformatics was used to simulate 1% and 2% tumor DNA samples by scaling down the number of methylated molecules based on the estimated tumor purity. Using the standard analysis workflow, all 2% tumor sample response scores were correctly judged as increased relative to the 1% tumor sample response scores (see Fig. 20 ; Table 5). As shown in Table 5, the sensitivity for different cancer types was 100% [95% CI: 93.4%, 100%].

[0422]

[0423] In summary, the data confirm that the Pan-Cancer assay exhibits high sensitivity across cancer types. For all 54 cancer patients included in the validation arm, the Pan-Cancer assay detected an increase in response score when the simulated tumor fraction changed from 1% to 2%.

[0424] 7.9.6 Conclusion

[0425] This analysis and clinical validation report describes the results from the experimental arms conducted to support the validation of the Pan-Cancer Assay. Overall, the results demonstrate that the assay has high accuracy and precision, high reproducibility, high sensitivity and specificity, and a low limit of detection. By measuring the amount of ctDNA in cancer patients via liquid biopsy, the Pan-Cancer Assay provides oncologists with additional information that can be used to improve clinical outcomes for their cancer patients.

[0426] 7.10 Example 8. Detection of the consistency between methylation and disease progression

[0427] The Pan-Cancer Assay was used to assess the concordance between methylation and disease progression. For this analysis, both methylation and CT imaging were used to assess disease progression in four patients. Methylation was monitored using the Pan-Cancer Assay. CT imaging was performed by clinicians and used to stage disease progression. Figure 21-24 The data shown in are presented as changes in methylation ("Tumor QE") over time ("Days since start of treatment"), with triangles indicating clinician-based stage of disease progression.

[0428] Fig.21 Methylation analysis and clinician-based disease progression stage of a 56-year-old male subject with pancreatic ductal adenocarcinoma stage IV at the time of initial collection are shown (see Fig.21 The triangle in Fig.21As shown in , methylation analysis was performed at the indicated time points, which coincided with the following treatment regimen: folfirinox ("folfiri"), followed by folfiri maintenance, followed by folox. Here, methylation levels are consistent with clinician-based staging. For example, a decrease in methylation corresponds to a decrease in tumor size or stable tumor growth (as indicated by the clinician), while an increase in methylation corresponds to possible disease progression (as indicated by the clinician).

[0429] Fig. 22 Methylation analysis and clinician-based disease progression stage of a 50-year-old male subject with stage IV pancreatic adenocarcinoma at the time of initial collection are shown (see Fig. 22 The triangle in Fig. 22 As shown in , methylation analysis was performed at the indicated time points, which coincided with the following treatment regimen: 5000 mg of fluorouracil + 800 mg of leucovorin + 180 mg of oxaliplatin + 300 mg of irinotecan for 46 hours. Here, the methylation levels are consistent with the clinician-based staging. For example, as the methylation levels increase before day 200 and continue to rise after day 200, the clinician-based staging indicates disease progression.

[0430] Fig.23 Shown is the methylation analysis of an 87-year-old male subject with stage IV colorectal adenocarcinoma at the time of initial collection. Fig.23 Clinician-based staging of disease progression is also shown (see Fig.23 The triangle in Fig.23 As shown in , methylation analysis was performed at the indicated time points, which coincided with the following treatment regimen: 6 cycles of fluorouracil 1100 mg + leucovorin 1100 mg + bevacizumab 500 mg. Here, the methylation levels are consistent with the clinician-based staging. For example, as the methylation levels increase before day 200 and continue to rise after day 200, the clinician-based staging indicates disease progression.

[0431] Fig.24 Shown is the methylation analysis of a 56-year-old female subject with stage IV squamous cell carcinoma of the lung at the time of initial collection. Fig.24 Clinician-based staging of disease progression is also shown (see Fig.24 The triangle in Fig.24 Methylation analysis was performed at the indicated time points, consistent with the following treatment regimens: docetaxel. Figure 21-23 Similar to the patients analyzed in , methylation levels here were consistent with clinician-based staging.

[0432] 7.11 Example 9. Consistency with variant allele fraction

[0433] A pan-cancer assay was also used to assess the concordance between methylation and variant allele fractions.

[0434] For this experiment, 40 samples were subjected to the treatment selection assay. For each sample, an aliquot of plasma was collected at the same time point as the methylation assay. The maximum variable allele fraction (VAF) was calculated from the treatment selection assay and compared to the methylation assay. Fig.25 As shown in , there is a correlation between the maximum VAF and the number of methylated molecules. Since somatic mutations can be heterogeneous throughout a tumor and may not reflect the overall abundance of the cancer, a perfect one-to-one correlation cannot be expected.

[0435] One limitation of the treatment selection assay is false negative actionable mutations. Given the correlation between VAF and methylation, the methylation results are used to provide information on whether a false negative actionable mutation identified with VAF is likely to be a false negative. For example, if no actionable mutation is found but the methylation level is low, this indicates that the tumor fraction is low at this time point and that false negative actionable mutations are likely to be present due to the small number of tumor molecules. On the other hand, if no actionable mutation is detected but the methylation level is high, this indicates that the tumor fraction is high at this time point and that it is unlikely that a large number of false negative actionable mutations are present.

[0436] In summary, the concordance between VAF and methylation suggests that methylation (ie, methylation analyzed and identified according to the methods described herein) can be used to inform VAF analysis results, which in turn can be used to supplement treatment selection determination decisions.

[0437] 7.12 Example 10. Changes in methylation profile

[0438] A pan-cancer assay was also used to assess methylation profiles over time. For these experiments, the methylation patterns of two subjects (subject 6885 and subject 5458) were analyzed over 294 days and 250 days, respectively. In summary and described in more detail below, it was observed that some subjects had relatively constant methylation profiles over time, while some patients had large changes in their methylation profiles.

[0439] For example, Fig.26 As shown in , for subject 6885, each new time point had new methylated loci. However, except for the loci present at the day 0 time point, none of these new loci contributed a significant amount of methylation at subsequent time points. For example, the locus that first appeared at day 63 did not contribute a significant amount of methylation at subsequent time points.

[0440] exist Fig. 27In the second non-limiting example shown in , for subject 5458, total methylation at day 119 increased with an interesting methylation pattern. Specifically, at day 119, slightly more than half of the methylation came from new loci (i.e., methylated loci not detected at day 0, day 35, or day 63), indicating that the tumor had evolved significantly compared to previous time points. In fact, time points after day 119 retained a large portion of the loci that first appeared at day 119, indicating that changes in tumor methylation are persistent. Based on the substantial increase and sudden change in methylation pattern composition, the subject was indicated for repeat treatment selection assays to detect any new somatic mutations in the cancer. Clinically, patterns such as those observed in subject 5458 may indicate the need to reevaluate treatment regimens.

[0441] 7.13 Conclusion of Examples 8-10

[0442] Overall, these data establish the utility of a pan-cancer assay that can be used to (1) detect concordance between methylation and disease progression; (2) detect concordance between methylation and variant allele fraction; and (3) assess methylation profiles over time.

[0443] 7.14 Example 11: Methylation Patterns in Pancreatic Ductal Adenocarcinoma Patients

[0444] The pan-cancer assay was used to evaluate the methylation profile in pancreatic ductal adenocarcinoma patients over time. The assay is used to detect a decrease in methylated molecules in circulating tumor DNA (ctDNA) compared to initial measurements. Abnormally methylated DNA is a known marker for cancer cells (PMID15542813), and changes in methylated ctDNA correspond to changes in tumor fraction. This result indicates that the tumor fraction has decreased compared to the previous measurement.

[0445] For these experiments, plasma and buffy coat were separated from whole blood collected in Streck cell-free DNA tubes. Cell-free DNA (cfDNA) was extracted from plasma, and genomic DNA (gDNA) was extracted from the buffy coat. The number of methylated molecules in cfDNA and gDNA was quantified at >500 positions in the genome known to be hypermethylated in cancer compared to non-cancerous tissue and blood using QCT molecule counting technology (PMID: 31591409). The methylation measured in gDNA was subtracted from the cfDNA methylation to remove background from the ctDNA signal. The remaining cfDNA methylated molecules at all hypermethylated positions were added to calculate the tumor methylation score. TM . Fig.28 Tumor methylation score is shown TM, expressed as the normalized sum of methylated molecules at >500 loci that are hypermethylated in circulating tumor DNA (ctDNA). Fig.28 As shown in Figure 2, tumor methylation scores were detected. TM This corresponds to a decrease in methylated ctDNA molecules. Fig.28 It was shown that the tumor fraction was reduced by about 3.1 times compared to the previous measurement.

[0446] Tumor methylation scores from the current collection TM Compared with the recently reported tumor methylation score TM Comparisons are made to determine increases, decreases, or no changes. Tumor Methylation Scores TM A change in the tumor methylation score must exceed a significance threshold to be reported as an increase or decrease. For baseline tests without any prior acquisition, tumor methylation is not scored. TM Results should be discussed with a healthcare professional and interpreted in the context of the patient's complete clinical history at multiple time points.

[0447] In some cases, when the sample contains insufficient amounts of DNA, methylation may not be reported. TM Limit of detection (LOD) results (depicted as cross-hatching on the graph) are not to be interpreted and will be reported as below the LOD. Performance specifications based on internal validation studies indicate that the assay can distinguish a 0.2 percentage point change in tumor score with a dispersion of 3 standard deviations. This pan-cancer assay is designed to quantify tumor methylation scores in patients with solid tumors. TM Results may differ or be invalidated if the patient has recently undergone a blood transfusion, stem cell transplant, or other surgery that could significantly affect the composition of cfDNA or buffy coat gDNA.

[0448] 8. EQUIVALENTS AND INCORPORATION BY REFERENCE

[0449] While the invention has been particularly shown and described with reference to a preferred embodiment and various alternative embodiments, it will be understood by those skilled in the relevant art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.

Claims

1. A method for quantifying DNA methylation in a sample containing a DNA sequence, the method comprising: treating the sample to encode the presence or absence of DNA methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; Adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a target sequence region of an endogenous target molecule; generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplification mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; as well as The number of methylated molecules in the sample is quantified based on the number of methylated reads from the sample and the number of reads from the collection of synthetic molecules.

2. The method of claim 1, wherein the sample is obtained from a blood draw comprising plasma and a buffy coat, wherein the plasma comprises cfDNA and the buffy coat comprises genomic DNA (gDNA) sequences.

3. The method of claim 2, further comprising extracting cfDNA from the plasma from the sample and extracting gDNA sequences from the buffy coat from the sample before processing the sample to encode the presence or absence of DNA methylation.

4. The method of claim 1, further comprising quantifying DNA methylation in a sample containing gDNA sequences from the buffy coat.

5. The method according to claim 4, comprising: treating the sample containing DNA sequences from the buffy coat to encode the presence or absence of DNA methylation in the DNA sequences, wherein the sample comprises at least ten target loci from the DNA sequences; Adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a target sequence region of an endogenous target molecule; generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplification mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; as well as determining the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules, This allows quantification of DNA methylation in gDNA sequences from the buffy coat.

6. The method according to any one of claims 1-5, further comprising adding a spike of known sequence and amount to the sample prior to the processing step.

7. The method of any one of claims 1-6, wherein processing the sample to encode the presence or absence of DNA methylation comprises bisulfite conversion or enzymatic conversion.

8. The method of any one of claims 1-7, wherein the collection of synthetic molecules is a collection of quantitative counting templates (QCTs).

9. The method of any one of claims 1-8, wherein each of the at least ten target loci is selected based on a predicted increase in DNA methylation at each locus in cancer tissue compared to non-cancerous tissue.

10. The method of any one of claims 1-9, wherein at least 100 target loci are amplified or at least 500 target loci are amplified.

11. The method of any one of claims 1-10, wherein the co-amplified mixture is sequenced at a read depth of at least 1 read per molecule.

12. The method of any one of claims 1-11, wherein the co-amplification mixture is sequenced at a read depth of 10 or more reads per molecule, 100 or more reads per molecule, or 1000 or more reads per molecule.

13. The method according to any one of claims 1-12, further comprising aggregating the number of methylated molecules of at least two of the at least ten target loci to quantify DNA methylation in the sample.

14. The method of claim 13, wherein the aggregating comprises aggregating target loci of the at least ten target loci that contain methylated molecules below a threshold amount in the buffy coat sample.

15. The method of any one of claims 1-14, wherein the sample further comprises at least one normalizing locus that is predicted to have high methylation in cfDNA in both cancerous and non-cancerous tissues.

16. The method of claim 15, wherein the co-amplification mixture comprises a collection of amplified synthetic molecules, a collection of amplified all or a subset of the at least 10 target loci, and amplified at least one normalizing locus.

17. The method according to any one of claims 1-16, further comprising standardizing the aggregated number of methylated molecules from at least two of the at least 10 target loci by the methylated molecules of the amplified at least one normalizing locus.

18. The method of any one of claims 1-17, further comprising subtracting background methylation.

19. The method according to any one of claims 1-18, further comprising filtering out selected hypermethylated target loci.

20. The method of any one of claims 1-19, further comprising subtracting background methylation and filtering out selected hypermethylated loci.

21. The method of any one of claims 18-20, wherein the subtraction of background methylation comprises subtracting the number of methylated molecules measured in the buffy coat layer quantified as described in claim 5 from the number of methylated molecules in the cfDNA.

22. The method of claim 21, wherein the subtraction of background methylation comprises subtracting the number of methylated molecules measured in the buffy coat as quantified as described in claim 5 from the number of methylated molecules in the cfDNA on a per locus basis.

23. The method of any one of claims 19-22, wherein filtering out selected hypermethylated target loci comprises filtering out target loci having hypermethylated cfDNA signals in non-cancerous tissues.

24. The method according to any one of claims 19-23, wherein the filtering out selected hypermethylated target loci comprises filtering out target loci whose total number of methylated molecules in the buffy coat is higher than a threshold value.

25. The method of claim 24, wherein the threshold is sample specific.

26. The method of claim 24, wherein the threshold is subject-specific.

27. The method of claim 24, wherein the threshold comprises a predetermined mean tumor methylation quantitative equivalent (QE), a predetermined maximum tumor methylation QE, or a combination thereof.

28. The method of claim 27, wherein the threshold comprises a predetermined mean tumor methylation QE.

29. The method according to any one of claims 19-28, wherein the filtering out selected hypermethylated target loci is performed before determining the number of methylated molecules in the sample.

30. The method of any one of claims 18-29, wherein the subtracting of background methylation is performed prior to determining the number of methylated molecules in the sample.

31. The method according to any one of claims 18-28, wherein said filtering out selected hypermethylated loci and said subtracting background methylation are performed before quantifying the number of methylated molecules in said sample.

32. A method for determining a DNA methylation profile in a subject, the method comprising: treating a sample isolated from a subject to encode the presence or absence of DNA methylation in a DNA sequence; Adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a target sequence region of an endogenous target molecule; generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of the methylated sequence reads in the sample and the number of reads from the set of synthetic molecules; as well as The methylation profile in the subject is determined based on the number of methylated molecules in the sample.

33. A method of determining a DNA methylation profile in a subject over time, the method comprising: At the first point in time: i) treating a sample isolated from the subject to encode the presence or absence of DNA methylation in a DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; ii) adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a target sequence region of an endogenous target molecule; iii) generating a co-amplification mixture comprising the set of amplified synthetic molecules and the amplified set of said at least ten target loci from said DNA sequence; iv) sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; v) determining the number of sequence reads that are methylated sequence reads; vi) quantifying the number of methylated molecules in the sample based on the number of methylated sequence reads in the sample and the number of reads from the set of synthetic molecules; Repeat steps i) to vi) at a second time point; as well as The DNA methylation profile in the subject is determined based on the number of methylated molecules in the sample at the first time point and the number of methylated molecules in the sample at the second time point.

34. A method according to claim 32, wherein the number of methylated molecules from each target locus in the sample is quantified based on the number of methylated sequence reads from each target locus in the sample and the number of reads from the set of synthetic molecules.

35. The method of claim 33, wherein determining the methylation profile in the subject at the first time point and the second time point identifies a change in the methylation profile between the first time point and the second time point.

36. The method according to any one of claims 33-35, further comprising: Repeat steps i) to vi) at a third time point; as well as A methylation profile in the subject is determined based on the number of methylated molecules for each target locus in the sample at the third time point.

37. The method of claim 36, wherein determining the methylation profile in the subject at the first time point, the second time point, and the third time point identifies a change in the methylation profile between the first and second time points, between the second and the third time points, between the first and the third time points, or a combination thereof.

38. The method of any one of claims 35-37, wherein a change in the methylation profile is indicative of a change in a tumor in the subject.

39. The method of claim 38, wherein the change in the tumor comprises a change in tumor size, a change in the abundance of somatic mutations associated with the tumor, the presence of new somatic mutations associated with the tumor, a chromosomal abnormality associated with the tumor, or resistance of the tumor to treatment.

40. The method of any one of claims 35-39, wherein changes in the methylation profile are incorporated into clinical recommendations for the subject.

41. The method of any one of claims 35-40, further comprising assigning a measure of increase, decrease, or no change to the change in the methylation profile based on a comparison to a significance threshold.

42. The method of claim 41, wherein the significance threshold is predetermined or dynamically calculated.

43. The method of any one of claims 32-41, wherein the sample is obtained from a blood draw comprising plasma and buffy coat, wherein the plasma comprises cfDNA and the buffy coat comprises genomic DNA (gDNA) sequences.

44. The method of claim 43, further comprising extracting cfDNA from the plasma from the sample and extracting gDNA sequences from the buffy coat from the sample prior to processing the sample to encode the presence or absence of DNA methylation.

45. The method of claim 43 or 44, further comprising quantifying DNA methylation in a sample containing gDNA sequences from the buffy coat.

46. ​​The method of claim 45, comprising: treating the sample containing DNA sequences from the buffy coat to encode the presence or absence of DNA methylation in the DNA sequences, wherein the sample comprises at least ten target loci from the DNA sequences; Adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule; generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplification mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules; as well as aggregating the number of methylated molecules of at least two target loci to quantify DNA methylation in the sample containing DNA sequences from the buffy coat, DNA methylation in the buffy coat is thereby quantified.

47. The method of any one of claims 32-46, further comprising adding a spike of known sequence and amount to the sample prior to the processing step.

48. The method of any one of claims 32-47, wherein the collection of synthetic molecules is a collection of quantitative counting templates (QCTs).

49. The method of any one of claims 32-48, wherein each of the at least ten target loci is selected based on a predicted increase in DNA methylation at the locus in cancer tissue compared to non-cancerous tissue.

50. The method of any one of claims 32-49, wherein at least 100 target loci are amplified or at least 500 target loci are amplified.

51. The method of any one of claims 32-50, wherein the sample further comprises at least one normalizing locus predicted to have high methylation of cfDNA in both cancerous and non-cancerous tissues.

52. The method of any one of claims 32-51, wherein the co-amplification mixture comprises a collection of amplified synthetic molecules, a collection of all or a subset of the at least 10 target loci amplified, and at least one amplified normalizing locus.

53. The method of any one of claims 32-52, further comprising aggregating the number of methylated molecules of at least two of the at least ten target loci to quantify DNA methylation in the sample.

54. The method of claim 53, wherein the aggregating comprises aggregating target loci of the at least ten target loci that contain less than a threshold amount of methylated molecules in the buffy coat sample.

55. The method of claim 53 or 54, further comprising normalizing the aggregated number of methylated molecules from at least two of the at least 10 target loci by the methylated molecules of the at least one normalizing locus.

56. The method of any one of claims 32-55, further comprising subtracting background methylation.

57. The method of any one of claims 32-56, further comprising filtering hypermethylated loci.

58. The method of any one of claims 33-57, further comprising subtracting background methylation and filtering out selected hypermethylated loci.

59. The method of any one of claims 56-58, wherein the subtraction of background methylation comprises subtracting the number of methylated molecules measured in the buffy coat layer quantified as described in claim 46 from the number of methylated molecules in the cfDNA.

60. The method of any one of claims 56-59, wherein the subtraction of background methylation comprises subtracting the number of methylated molecules measured in the buffy coat as quantified as described in claim 46 from the number of methylated molecules in the cfDNA on a per locus basis.

61. The method of any one of claims 57-60, wherein filtering out selected hypermethylated loci, subtracting background methylation, or both is performed prior to quantifying the number of methylated molecules in the sample.

62. The method according to any one of claims 56-61, wherein the filtering out selected hypermethylated target loci comprises filtering out target loci whose total number of methylated molecules in the buffy coat is higher than a threshold value.

63. The method of claim 62, wherein the threshold is sample specific.

64. The method of claim 62, wherein the threshold is subject-specific.

65. The method of claim 62, wherein the threshold comprises a predetermined mean tumor methylation QE, a predetermined maximum tumor methylation QE, or a combination thereof.

66. The method of claim 65, wherein the threshold comprises a predetermined mean tumor methylation QE.

67. The method of any one of claims 32-66, further comprising subjecting the subject to a treatment selection assay after detecting a change in the DNA methylation profile.

68. The method of claim 67, wherein the treatment selection assay comprises genomic analysis to detect novel somatic mutations, the abundance of somatic mutations, or both.

69. A method for determining a confidence indication of the presence or absence of a somatic mutation in a sample containing a DNA sequence, the method comprising: determining the presence or absence of a somatic mutation in the sample; Quantifying DNA methylation in the sample, the quantification comprising the following steps: treating the sample to encode the presence or absence of v methylation in the DNA sequence, wherein the sample comprises at least ten target loci from the DNA sequence; Adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a sequence region of an endogenous target molecule; generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplified mixture at a read depth of at least one read sequence per molecule in the sample to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules; and Returns the following confidence level indication: When the number of the methylated molecules in the sample is higher than a predetermined or dynamically calculated threshold, returning a true call result for the presence or absence of the somatic mutation; or When the number of the methylated molecules in the sample is at or below the predetermined or dynamically calculated threshold, returning a no-call result for the presence or absence of the somatic mutation, This truly determines the confidence of the determination of the presence or absence of the somatic mutation.

70. The method of claim 69, wherein said determining the presence or absence of a somatic mutation in the sample comprises determining a variant allele fraction of the somatic mutation.

71. The method of claim 69 or 70, wherein the sample is obtained from a blood draw comprising plasma and buffy coat, wherein the plasma comprises cfDNA and the buffy coat comprises genomic DNA (gDNA) sequences.

72. The method of claim 71 , further comprising extracting cfDNA from plasma from the sample and extracting gDNA sequences from the buffy coat from the sample prior to processing the sample to encode the presence or absence of DNA methylation.

73. The method of claim 71 or 72, further comprising quantifying DNA methylation in a sample containing DNA sequences from the buffy coat.

74. The method of claim 73, comprising: treating the sample containing DNA sequences from the buffy coat to encode the presence or absence of DNA methylation in the DNA sequences, wherein the sample comprises at least ten target loci from the DNA sequences; Adding to the sample a collection of synthetic molecules, the collection of molecules comprising: a target-associated region having a nucleotide sequence that matches a target sequence region of an endogenous target molecule comprising at least one of the target loci, and a variant region having a nucleotide sequence that does not match a target sequence region of an endogenous target molecule; generating a co-amplification mixture comprising a collection of amplified synthetic molecules and a collection of amplified at least ten target loci from the DNA sequence; sequencing the co-amplification mixture to generate sequence reads; determining the number of sequence reads that are methylated sequence reads; as well as quantifying the number of methylated molecules in the sample based on the number of methylated reads from the sample and the number of reads from the set of synthetic molecules, This allows quantification of DNA methylation in gDNA sequences from the buffy coat.

75. The method of any one of claims 69-74, further comprising adding a spike of known sequence and amount to the sample prior to the processing step.

76. The method of any one of claims 69-75, wherein the collection of synthetic molecules is a collection of quantitative counting templates (QCTs).

77. The method of any one of claims 69-76, wherein each of the at least ten target loci is selected based on a predicted increase in DNA methylation at each locus in cancer tissue compared to non-cancerous tissue.

78. The method of any one of claims 69-77, wherein at least 100 target loci are amplified or at least 500 target loci are amplified.

79. The method of any one of claims 69-78, wherein the sample further comprises at least one normalizing locus predicted to have high methylation of cfDNA in both cancerous and non-cancerous tissues.

80. The method of any one of claims 69-79, wherein the co-amplification mixture comprises a collection of amplified synthetic molecules, a collection of all or a subset of the at least 10 target loci that are amplified, and at least one amplified normalizing locus.

81. The method of any one of claims 69-80, further comprising aggregating the number of methylated molecules of at least two of the at least ten target loci to quantify DNA methylation in the sample.

82. The method of claim 81, wherein the aggregating comprises aggregating target loci of the at least ten target loci that contain less than a threshold amount of methylated molecules in the buffy coat sample.

83. The method of claim 81 or 82, further comprising normalizing the aggregated number of methylated molecules from at least two of the at least 10 target loci by the methylated molecules of the at least one normalizing locus.

84. The method of any one of claims 69-83, further comprising subtracting background methylation.

85. The method of any one of claims 69-84, further comprising filtering out selected hypermethylated loci.

86. The method of any one of claims 69-85, further comprising subtracting background methylation and filtering out selected hypermethylated loci.

87. The method of any one of claims 84-86, wherein subtracting background methylation comprises subtracting the number of methylated molecules measured in the buffy coat layer quantified as described in claim 74 from the number of methylated molecules in the cfDNA.

88. The method of any one of claims 84-87, wherein subtracting background methylation comprises subtracting the number of methylated molecules measured in the buffy coat as quantified as described in claim 74 from the number of methylated molecules in the cfDNA on a per locus basis.

89. The method of any one of claims 84-88, wherein filtering out selected hypermethylated loci, subtracting background methylation, or both is performed prior to quantifying the number of methylated molecules in the sample.

90. The method of any one of claims 84-89, wherein filtering out selected hypermethylated target loci comprises filtering out target loci for which the total number of methylated molecules in the buffy coat is above a threshold.

91. The method of claim 90, wherein the threshold is sample-specific.

92. The method of claim 90, wherein the threshold is subject-specific.

93. The method of claim 90, wherein the threshold comprises a predetermined mean tumor methylation QE, a predetermined maximum tumor methylation QE, or a combination thereof.

94. The method of claim 93, wherein the threshold comprises a predetermined mean tumor methylation QE.

95. The method of any one of claims 69-94, further comprising, upon returning a true call and if the abundance of the somatic mutations indicates the presence of cancer, performing a treatment selection assay or repeating a treatment selection assay on the subject.

96. The method of claim 95, wherein the treatment selection assay comprises genomic analysis to detect novel somatic mutations, the abundance of somatic mutations, or both.

97. The method of any one of claims 69-96, further comprising, upon returning a no determination, repeating the method of any one of claims 69-96 on a different sample taken from the subject.

98. The method of any one of claims 69-97, further comprising incorporating the confidence indication into a clinical recommendation for the subject.

Images