Tumor fraction (TF) is associated with real-world progression-free survival (RWPFS) in non-small-cell lung cancer (NSCLC) patients treated with platinum-based chemotherapy (CHEMO)

Abstract

Described herein are methods and compositions related to using genomic and / or methylation-based TF from a genomic, epigenomic detection platform to predict clinical benefit to therapies, including through evaluation of real-world patients. Tumor detection, methylation-based tumor fraction, and methylation-based molecular response are prognostic and predictive of outcomes, Prolonged treatment of patients not deriving benefit risks increased toxicity, may reduce chance of success for targeted treatments, and may increase clonal hematopoietic burden, wherein methylation based detection largely eliminates noise in determining genomic molecular response (gMR), thereby demonstrating advantageous use of methylation based molecular response (mMR).

Description

Attorney Docket No. GH0240WOTUMOR FRACTION (TF) IS ASSOCIATED WITH REAL-WORLD PROGRESSION-FREE SURVIVAL (RWPFS) IN NON-SMALL-CELL LUNG CANCER (NSCLC) PATIENTS TREATED WITH PLATINUM-BASED CHEMOTHERAPY (CHEMO)CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 687,244, filed August 26, 2024, which is incorporated by reference herein in its entirety.BACKGROUND

[0002] Current response evaluation methods have accuracy limitations. Monitoring with a circulating tumor DNA (ctDNA) methylation-based tumor fraction (TF) may provide a more accurate assessment of tumor burden across solid tumors. Genomic ctDNA molecular response (gMR) predicts clinical benefit from immune checkpoint inhibitors, but associations with benefit from various therapies, such as immunecheckpoint inhibitors, chemotherapy, etc. are inconsistent. gMR may predict benefit from chemo after filtering to tissue-confirmed somatic variants, suggesting interference from clonal hematopoiesis (CH) mutations. Using methylation tumor fraction (TF) for MR assessment may mitigate the risk of CH interference without tissue.

[0003] Described herein are methods and compositions related to using methylation-based TF from an epigenomic detection platform to predict clinical benefit in cohorts of patients, including NSCLC and aNSCLC. In particular, it is demonstrated that tumor detection, methylation-based tumor fraction, and methylation-based Response are prognostic and predictive of outcomes, Prolonged treatment of patients not deriving benefit risks increased toxicity, may reduce chance of success for targeted treatments, and may increase CH burden, which could contribute to significant noise in determining gMR.

[0004] Association with outcomes validates methylation for TF and methylation-based MR (mMR). Improved prediction of outcomes on various therapies such as immune checkpoint inhibitor, chemotherapy, etc. would support MR including methylation based MR (mMR) utility in a setting where other measurements of molecular response are challenging, providing opportunity for earlier intervention and potentially improved patient outcomes.BRIEF DESCRIPTION OF THE FIGURES

[0005] Figure 1. Candidate thresholds for TF change indicating MR and nMR among the patients where either TO or T1 were quantifiable (n=459) (A) Forest plot for various tumorAttorney Docket No. GH0240WO fraction changes (B) Kaplan-Meier analysis of rwPFS by select candidate thresholds across the fulldataset. Log-rank test between groups: 100% vs any decrease, p < 0.005; 80% decrease vs any decrease, p = 0.01; 100% decrease vs 80% decrease, p = 0.04; 100% decrease vs any increase, p< 0.005; 80% decrease vs any increase, p < 0.005; any decrease vs any increase, p < 0.005.

[0006] Figure 2A-Figure 2B. Longitudinal assessment of TF predicts ICI benefit in pansolid tumor cohort (n=521; patients where rwPFS or rwOS event occurred before re-index date were excluded where indicated) Figure 2A. Kaplan-Meier analysis comparing rwPFS of MR at any timepoint (NR [95% CI NR, NR]) vs. nMR at all timepoints (3.93 months [95% CI 3.33, 4.40]) (n=513) Figure 2B. Kaplan-Meier analysis comparing rwOS of MR at any timepoint vs. nMR at all timepoints (NR [95% CI NR, NR] vs. 7.166 months [95% CI 5.73, 9.10) (n=520).

[0007] Figure 3A-Figure 3D. MR outcome prediction is complementary to RECIST in subset of patients with 6-month best RECIST response available (n=75). Kaplan-Meier plots for rwPFS Figure 3A. RECIST association with rwPFS by two groups combining CR / PR / SD vs. PD Figure 3B. MR and nMR at any timepoint association with rwPFS in RECIST subset Figure 3C. Waterfall plot of greatest percentage TF change (best molecular response) within 6-months by RECIST response Figure 3D. SD patients stratified by MR and nMR at any timepoint and association with rwPFS (n=31).

[0008] Figure 4A-Figure 4D. ctDNA change in Genomic vs Methylation MR - all regimens. Figure 4A. Histogram of ctDNA percentage change (genomic). Figure 4B. Histogram of ctDNA percentage change (methylation-based). Figure 4C. Unique patients with decreased or increased ctDNA levels (genomic). Figure 4D. Unique patients with decreased or increased ctDNA levels (methylation-based).

[0009] Figure 5. Expected pattern of TF over course of treatment.

[0010] Figure 6. Expected pattern of TF over course of treatment.

[0011] Figure 7. Baseline TF.

[0012] Figure 8. Max variant allele fraction (VAF) from genomic / epigenomic report vs Max VAF from prior genomic only test. Pattern largely consistent with expectation of the same sample rerun on a larger panel space and capturing some additional variants. Max VAF based on all 740 genes.

[0013] Figure 9. Early view of TMS across sequential timepoints. Pattern largely consistent with initial response / lack of response at early timepoints and later progression.Attorney Docket No. GH0240WO

[0014] Figure 10. MR patient inclusion criteria. 67 pts after restricting to outcomes data. Restrict MR pairs w / outcome data. Baseline is prior to start of treatment but not more than 105 days prior. On-treatment sample is more than two weeks start of therapy but not more than 105 days after treatment start. On-treatment sample must be more than 10 days prior to treatment end. Take the earliest non-targeted, non-IO line for multiple on-treatment / multiple regimen sample / treatment pairs. Take the "baseline" that is closest to the start of treatment for pts with 2 baseline samples.

[0015] Figure 11 Differences between genomic and methylation MR results.

[0016] Figure 12. Genomic MR vs Methylation MR. gMR>0, mMR=0. This plot includes all sample pairs.

[0017] Figure 13. Exemplary Genomic MR non-responder. Genomic MR is not significant at any cutpoint in this small cohort for identifying pts not responding / responding to therapy. E.g. Expect those pts without decreasing ctDNA have shorter rwTTNT (progress faster). Top is clearance. Different levels of decrease. Down to a 17% or greater decrease. Including more and more pts in the non-responder group. Comparing to responders as reference group. Makes HR values >1.

[0018] Figure 14. Exemplary Methylation MR non-responder. Methylation MR surprisingly significant even in this small cohort. Optimal range between 95%or greater decrease and 70% or greater decrease. Looks at mMR score only. Algorithm-based. Same trend, but IS significant.

[0019] Figure 15 TF based correct trend and TF based >95% decrease successfully identifies responders.

[0020] Figure 16. Exemplary on-treatment TMS. On treatment TMS cutpoint near median does separate into two groups with significantly different rwTTNT.SUMMARY OF THE INVENTION

[0021] Described herein is a method of determining a molecular response score at least partially using a computer. In some embodiments, the method includes receiving and / or determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject. In some embodiments, the method includes determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined at a first time point and the second plurality of sequence reads are determined at a later time point, further including additional pluralities of sequence reads at a plurality of later time points. In someAttorney Docket No. GH0240WO embodiments, the method includes determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy. In some embodiments, the method includes, classifying a plurality of variants in the first plurality of sequence reads and the second plurality of sequence reads as somatic or germline. In various embodiments, the method includes determining, for at least one variant of the plurality of variants classified as somatic, based on a first variant allele fraction (VAF) and a second VAF, a quantitative metric for the first VAFs and a quantitative metric for the second VAFs, determining, for the subject, a ratio of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs, determining, based on a comparison of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs and outputting, a molecular response score. In various embodiments, the quantitative metric can be a mutant allele fraction (MAF), tumor fraction (TF), circulating tumor DNA (ctDNA) levels, etc., including methylation based detection of any of the aforementioned, for comparison between and / or amongst a first plurality of sequence reads and any additional plurality of sequence reads.

[0022] In some embodiments, the method includes determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy, classifying a plurality of variants in the first plurality of sequence reads and the second plurality of sequence reads as somatic or germline, determining, for at least one variant of the plurality of variants classified as somatic, based on a first variant allele fraction (VAF) and a second VAF, a quantitative metric for the first VAFs and a quantitative metric for the second VAFs, determining, for the subject, a ratio of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs, determining, based on a comparison of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs and outputting, a molecular response score.

[0023] In some embodiments, the method includes determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy, classifying a plurality of variants in the first plurality of sequence reads and the second plurality of sequence reads as somatic or germline. In various embodiment, the method includesAttorney Docket No. GH0240WO determining, for at least one variant of the plurality of variants classified as somatic, based on a first variant allele fraction (VAF) and a second VAF, a quantitative metric for the first VAFs and a quantitative metric for the second VAFs, determining, for the subject, a ratio of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs, determining, based on a comparison of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs and outputting, a molecular response score.

[0024] Described herein is a method of determining a molecular response score at least partially using a computer. In some embodiments, the method includes determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined at a first time point and the second plurality of sequence reads are determined at a later time point, further including additional pluralities of sequence reads at a plurality of later time points. In various embodiments, the method includes receiving and / or determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy. In various embodiments, the method includes classifying a plurality of variants in the first plurality of sequence reads and the second plurality of sequence reads as somatic or germline. In various embodiments, the method includes determining a quantitative measure associated with a mutant allele fraction (MAF), tumor fraction (TF), circulating tumor DNA (ctDNA) levels, etc., including methylation based detection of any of the aforementioned for comparison between and / or amongst a first plurality of sequence reads and any additional plurality of sequence reads.

[0025] In various embodiments, the method include determining, for at least one variant of the plurality of variants classified as somatic, based on a first variant allele fraction (VAF) and a second VAF, an VAF ratio. In various embodiments, the method include determining, for at least one variant of the plurality of variants classified as somatic, based on a first variant allele fraction (VAF) and a second VAF, an VAF ratio. In various embodiments, the method includes determining, for the subject, quantitative metric for the VAF ratios. In various embodiments, the method includes determining, based on the quantitative metric for the VAF ratios, and optionally determining a confidence interval associated with quantitative metric for the VAF ratios, and outputting, a molecular response score.

[0026] Described herein is a method of determining a molecular response score at least partially using a computer. In some embodiments, the method includes determining a firstAttorney Docket No. GH0240WO plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined at a first time point and the second plurality of sequence reads are determined at a later time point, further including additional pluralities of sequence reads at a plurality of later time points. In some embodiments, the method includes receiving and / or determining a first plurality of sequence reads and a second plurality of sequence reads associated with a subject, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy. In some embodiments, the method includes classifying a plurality of variants in the first plurality of sequence reads as somatic or germline, classifying the plurality of variants in the second plurality of sequence reads as somatic or germline. In some embodiments, the method includes reclassifying at least one variant of the plurality of variants to resolve a classification discrepancy between the first plurality of sequence reads and the second plurality of sequence reads, determining, for at least one variant of the plurality of variants classified or reclassified as somatic, based on at least a portion of the first plurality of sequence reads, one or more quantitative metrics (e.g., a first variant allele fraction) determining, for at least one variant of the plurality of variants classified or reclassified as somatic, based on at least a portion of the second plurality of sequence reads, one or more quantitative metrics (e.g., a second variant allele fraction). In some embodiments, the method includes determining, based on the first quantitative metric and the second quantitative metric, a molecular response score. In some embodiments, the method includes determining, based on the first variant allele fraction and the second variant allele fraction, a molecular response score. In various embodiments, determination of the molecular response score is determined using a predefined training set. In various embodiments, the molecular response score includes comparison to a threshold, such as a population-based training threshold. In various embodiments, the quantitative metric may include a weighted mean, a measure of methylation-based measurements such as tumor fraction (TF), clonal hematopoiesis (GH), mutant allele fraction (MAF), circulating tumor DNA (ctDNA) levels, etc. as may be measured through differentially methylated regions (DMRs). In various embodiments, the molecular response score includes a confidence interval. Additional information is found in PCT Pub. No. WO2022187862A1, WO2024107599, which are each incorporated by reference herein.

[0027] Described herein is a method of characterizing a sample, comprising: obtaining or having obtained sequence information for a first plurality of sequence reads and at least a second plurality of sequence reads derived from a sample from subject; determining at leastAttorney Docket No. GH0240WO quantitative metric for the first plurality of sequence reads and at least one quantitative metric for the second plurality of sequence reads; and determining, based on the first quantitative metric and the second quantitative metric, a molecular response score, thereby characterizing the sample. In various embodiments, thefirst plurality of sequence reads are determined at a first time point and the second plurality of sequence reads are determined at a later time point. In various embodiments, method includes additional pluralities of sequence reads at a plurality of later time points. In various embodiments, the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy. In various embodiments, the method includes classifying a plurality of variants in the first plurality of sequence reads as somatic or germline; and classifying the plurality of variants in the second plurality of sequence reads as somatic or germline. In various embodiments, the method includes determining a molecular response score using a predefined training set. In various embodiments, the score includes comparison to a threshold. In various embodiments, the threshold includes a populationbased training threshold. In various embodiments, the method includes determining, for at least one variant of the plurality of variants classified as somatic, a quantitative metric for the first plurality of sequence reads and a quantitative metric for the second plurality of sequence reads; comparing the quantitative metric for the first plurality of sequence reads and the quantitative metric for the second plurality of sequence reads; and determining a molecular response score based on the comparison. In various embodiments, the method includes determining on a tumor fraction (TF) estimation for the first, the second and additional pluralities of sequence reads. In various embodiments, the method includes methylationbased detection of TF. In various embodiments, the comparison includes determining a ratio of the quantitative metric for a first TF and the quantitative metric for a second TF; and determining a molecular response score. In various embodiments, the ratio includes a hazard ratio. In various embodiments, the quantitative metric includes a mean, median, mode, weighted mean, central median. In various embodiments, methylation-based detection is measured through differentially methylated regions (DMRs).

[0028] In various embodiments of any of the aforementioned method, determination of the molecular response score is determined using a predefined training set. In various embodiments, a Cox proportional hazards models and or other model is used to determine one or more ratios between a first plurality of sequence reads derived from a first plurality of subject and one or more plurality of sequence reads derived from at least one additional plurality of subjects. In various embodiments, the first plurality of subjects includeAttorney Docket No. GH0240WO responders, and a second plurality of subjects include non-responders. In various embodiments, the hazard ratio is 2, 3, 4, 5 or more. For example, one can model a hazard ratio (HR) of 2 between tumor-fraction (TF) responders and non-responders. In various embodiments, a power calculation is determined, including 70, 80, 95, 95% or more power at various alpha (e.g., 0.05).

[0029] In various embodiments, the molecular response score includes comparison to a threshold, such as a population-based training threshold. In various embodiments, a first plurality of subjects and an at least one additional plurality of subjects includes quantitative measure measurements at one or more timepoints. In various embodiments, threshold generation includes one or more parameters, including for example, hazard ratio (HR), OS, rwPFS, TTNT, TF fraction etc. In various embodiments, the one or more parameters support determination of ctDNA change threshold distinguishing molecular response (MR) in a first plurality of subject and at least one additional plurality of subjectsDETAILED DESCRIPTION

[0030] Establishing responsiveness to therapy and detecting progression is vital to treatment decisions for patients with advanced cancer. In common solid tumor malignancies, radiological imaging is currently used, and when available, blood proteins. The most widely utilized and validated response criteria is the Response Evaluation Criteria in Solid Tumors (RECIST), which

[0031] measures changes in tumor burden based on radiological imaging. iRECIST was later introduced to address the challenges with pseudoprogression when evaluating response to immunotherapy. RECIST has limitations, including inter-reader reliability of radiological images, requiring measurable disease at baseline, constraints regarding selection of target lesions, detecting new lesions, inability to accurately measure bone metastases and body cavity-based disease such as pleural effusion and ascites, and for patients with stable disease (SD), uncertainty regarding clinical benefit . Therefore, improved methods of clinical response evaluation are needed.

[0032] Liquid biopsy assays have been studied as a method for evaluating molecular response to therapy and ctDNA levels may be easily monitored with simple blood draws throughout a patient’s journey to provide early information on response and progression. Liquid biopsy has also been explored as a tool to distinguish between pseudoprogression and true progression in patients receiving immune checkpoint inhibitors (ICI), complementing imaging-based response assessments. Furthermore, patients deemed to have SD by RECISTAttorney Docket No. GH0240WO may be further adjudicated by ctDNA results. Liquid biopsy assays have thus far focused on genomic analysis, using variant allele fraction (VAF) changes for monitoring. This approach has limitations, including reduced sensitivity at low ctDNA levels and potential interference from copy number variation and clonal hematopoiesis (26,29,33). Tumor-informed ctDNA monitoring assays have been developed to address these limitations, but this approach requires sufficient tissue, which may affect evaluability and increase turnaround time. Methylation-based ctDNA quantification has been explored as an alternative approach. This study evaluates a tissue-free, methylation-based next-generation sequencing (NGS) assay for quantifying ctDNA levels and assessing correlations between changes in tumor fraction (TF) and patient outcomes in the real-world pan-cancer RADIOHEAD study.Ligation of Adapters

[0033] Double-stranded nucleic acids e.g., DNA molecules in a sample, and single stranded nucleic acid molecules converted to double stranded molecules, can be linked to adapters at either one end or both ends. In the methods of the disclosure, adapters can be ligated to sample nucleic acids prior to the partitioning and / or conversion steps. In some embodiments, adapters may be ligated to the sample nucleic acids after the partitioning and conversion steps, but before the step of amplifying the nucleic acids which have been subjected to partitioning and conversion steps.

[0034] In some embodiments, the DNA is made ligatable, e.g., by extending the end overhangs of the DNA molecules, and adding adenosine residues to the 3’ ends of fragments and phosphorylating the 5’ end of each DNA fragment. Typically, double stranded molecules are blunt ended by treatment with a polymerase with a 5'-3 ' polymerase and a 3 '-5' exonuclease (or proof reading function), in the presence of all four standard nucleotides. Klenow large fragment and T4 polymerase are examples of suitable polymerase.

[0035] The blunt ended DNA molecules can be ligated with at least partially double stranded adapter (e.g., a Y shaped or bell-shaped adapter). Alternatively, complementary nucleotides can be added to blunt ends of sample nucleic acids and adapters to facilitate ligation.Contemplated herein are both blunt end ligation and sticky end ligation. In blunt end ligation, both the sample nucleic acid molecules and the adapters have blunt ends. In sticky-end ligation, typically, the sample nucleic acid molecules bear an “A” overhang and the adapters bear a “T” overhang.

[0036] DNA ligase and adapters are added to ligate DNA molecules in the sample with an adapter on one or both ends, i.e. to form adapted DNA. As used herein, “adapter” refers toAttorney Docket No. GH0240WO short nucleic acids (e.g., less than about 500, less than about 100 or less than about 50 nucleotides in length, or be 20-30, 20-40, 30-50, 30-60, 40-60, 40-70, 50-60, 50-70, 20-500, or 30-100 bases from end to end) that are typically at least partially double-stranded and can be ligated to the end of a given sample nucleic acid molecule. In some instances, two adapters can be ligated to a single sample nucleic acid molecule, with one adapter ligated to each end of the sample nucleic acid molecule.

[0037] Adapters can include nucleic acid primer binding sites to permit amplification of a sample nucleic acid molecule flanked by adapters at both ends, and / or a sequencing primer binding site, including primer binding sites for sequencing applications, such as various next generation sequencing (NGS) applications. Adapters can include a sequence for hybridizing to a solid support, e.g., a flow cell sequence. Adapters can also include binding sites for capture probes, such as an oligonucleotide attached to a flow cell support or the like. Adapters can also include sample indexes and / or molecular barcodes. These are typically positioned relative to amplification primer and sequencing primer binding sites, such that the sample index and / or molecular barcode is included in amplicons and sequencing reads of a given nucleic acid molecule. Adapters of the same or different sequence can be linked to the respective ends of a sample nucleic acid molecule. In some embodiments, adapters of the same or different sequence are linked to the respective ends of the nucleic acid molecule except that the sample index and / or molecular barcode differs in its sequence. In some embodiments, the adapter is a Y-shaped adapter in which one end is blunt ended or tailed as described herein, for joining to a nucleic acid molecule, which is also blunt ended or tailed with one or more complementary nucleotides to those in the tail of the adapter. In another exemplary embodiment, an adapter is a bell-shaped adapter that includes a blunt or tailed end for joining to a nucleic acid molecule to be analyzed. Other exemplary adapters include T- tailed, C-tailed or hairpin shaped adapters. For example, a hairpin shaped adapter can comprise a complementary double stranded portion and a loop portion, where the double stranded portion can be attached (e.g., ligated) to a double-stranded polynucleotide. Hairpin shaped sequencing adapters can be attached to both ends of a polynucleotide fragment to generate a circular molecule, which can be sequenced multiple times.

[0038] In some embodiments, the nucleic acids further comprise adapters in which at least one cytosine is a modification resistant cytosine, optionally wherein each cytosine in the adapters is a modification resistant cytosine. In some embodiments, methods further comprise further comprising ligating adapters to the nucleic acids, wherein at least one cytosine in the adapters is a modification resistant cytosine, optionally wherein the ligating occurs beforeAttorney Docket No. GH0240WO step (c) and / or after step (a); further optionally wherein each cytosine in the adapters is a modification resistant cytosine. The adapters may comprise barcodes, e.g., according to any of the embodiments relating to barcodes described elsewhere herein. In some embodiments, the adapters can include at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 modified nucleotides, such as modified cytosine nucleotides, that are resistant to modification, e.g., conversion. In some embodiments, the modified nucleotides are resistant to modification by a deaminase. In some embodiments, the modified nucleotides comprise a conversion resistant modified cytosine, such as 5-propynylC (5pyC), 5-pyrrolo-dC (5pyrC), 5-hydroxymethylcytosine (5hmC) along with modified variants thereof, glucosylated5- hydroxymethylcytosine (5ghmC), cytosine 5-methylenesulfonate (CMS), bulky 5-position adducts, or N4-modified cytosine. In some embodiments, the conversion resistant modified cytosine is 5pyC, 5pyrC, 5ghmC, or CMS. In some embodiments, the conversion resistant modified cytosine can protect cytosine from being converted by a deaminase, such as a cytidine deaminase, which converts a cytosine to uracil. In some embodiments, each cytosine of an adapter is a conversion resistant modified cytosine, such as any one or more of the foregoing examples. For exemplary descriptions of modified nucleotides and their use in adaptors, see WO2023 / 288222 and U.S. Pat. No. 10,260,088.

[0039] The adapters used in the methods of the present disclosure may comprise one or more known nucleosides wherein the base has a known methylation status, such as 5mC nucleic acid bases. When using adapters comprise 5mC, the adapters can be ligated to the sample nucleic acid molecules prior to the conversion procedure. Analyzing the sequence data corresponding to these known 5mC nucleic acid bases allows for the efficiency of the conversion procedure to be measured, which can be used as a quality control measure for the conversion procedure. In instances where two adapters are ligated to a sample nucleic acid (one at each end), either or both of the adapters may comprise one or more nucleosides with a known methylation status. Typically the primer binding site(s), sequencing primer binding site(s), sample index(es) and / or molecular barcode(s), if present, do not comprise the nucleosides with a methylation status that change base pairing specificity as a result of the conversion procedure. Preferably adapters (e.g., Y-shaped adapters) are ligated to the sample nucleic acids prior to the conversion and partitioning steps.

[0040] In some embodiments, the disclosed methods comprise analyzing DNA in a sample. In such methods, adapters may be added to the DNA. This may be done concurrently with an amplification procedure, e.g., by providing the adapters in a 5’ portion of a primer (where PCR is used, this can be referred to as library prep-PCR or LP-PCR), before, or after anAttorney Docket No. GH0240WO amplification step. In some embodiments, adapters are added by other approaches, such as ligation. In some such methods, first adapters are added to the 3’ ends of the nucleic acids by ligation, which may include ligation to single-stranded DNA. In some such methods, first adapters are added to the 5’ ends of the nucleic acids by ligation, which may include ligation to single-stranded DNA. In some embodiments, prior to any partitioning or capturing steps, first adapters are added to the nucleic acids by ligation, which may include ligation to singlestranded DNA (e.g., to the 3’ ends thereof). In some embodiments, the capture probes can be isolated after partitioning and ligation. For example, the hypomethylated partition can be ligated with adapters and a portion of the ligated hypomethylated partition can then be used to generate the capture probes for rearrangements. The adapter can be used as a priming site for second-strand synthesis, e.g., using a universal primer and a DNA polymerase. A second adapter can then be ligated to at least the 3’ end of the second strand of the now doublestranded molecule. In some embodiments, the first adapter includes an affinity tag, such as biotin, and nucleic acid ligated to the first adapter is bound to a solid support (e.g., bead), which may comprise a binding partner for the affinity tag such as streptavidin. For further discussion of a related procedure, see Gansauge et al., Nature Protocols 8:737-748 (2013). Commercial kits for sequencing library preparation compatible with single-stranded nucleic acids are available, e.g., the Accel-NGS® Methyl-Seq DNA Library Kit from Swift Biosciences. In some embodiments, after adapter ligation, nucleic acids are amplified.

[0041] In some embodiments, the single-stranded DNA library preparation is performed in a one-step combined phosphorylation / ligation reaction, e.g., as described in Troll et al., BMC Genomics, 20: 1023 (2019), available at doi.org / 10.1186 / sl2864-019-6355-0. This method, called Single Reaction Single-stranded LibrarY (“SRSLY,”) can be performed without endpolishing. SRSLY may be useful for converting short and fragmented DNA molecules, e.g., cfDNA fragments, into sequencing libraries while retaining native lengths and ends. The SRSLY method can create sequencing libraries (e.g., Illumina sequencing libraries) from fragmented or degraded template (input) DNA. In particular embodiments, template DNA is first heat denatured and then immediately cold shocked to render the template DNA molecules single-stranded. The DNA can be maintained as single-stranded throughout the ligation reaction by the inclusion of a thermostable single-stranded binding protein (SSB). Next, the template DNA, which at this point can be single-stranded and coated with SSB, is placed in a phosphorylation / ligation dual reaction with directional dsDNA NGS adapters that contain single-stranded overhangs. Both the forward and reverse sequencing adapters can share similar structures but differ in which termini is unblocked in order to facilitate properAttorney Docket No. GH0240WO ligations. Both sequencing adapters can comprise a dsDNA portion and a single-stranded splint overhang of random nucleotides that occurs on the 3 -prime terminus of the bottom strand of the forward adapter and the 5-prime terminus of the bottom strand of the reverse adapter. In this way, the forward adapter (e.g., (P5) Illumina adapter) can delivered to the 5- prime end of template molecules and the reverse adapter (e.g., (P7) Illumina adapter) is delivered to the 3-prime end of template molecules. Thus, the native polarity of input DNA molecules can be retained.

[0042] During the dual phosphorylation / ligation reaction, T4 Polynucleotide Kinase (PNK) can be used to prepare template DNA termini for ligation by phosphorylating 5-prime termini and dephosphorylating 3-prime termini. T4 PNK works on both ssDNA and dsDNA molecules and has no activity on the phosphorylation state of proteins. Simultaneously, the random nucleotides of the splint adapter can be annealed to the single-stranded template molecule. This creates a short, localized dsDNA molecule, enabling ligation of template to adapter with a ligase such as T4 DNA ligase, which has high ligation efficiency on dsDNA templates but low efficiency on ssDNA. After the single phosphorylation / ligation reaction is complete, the library DNA can be, e.g., purified and placed directly into standard NGS indexing PCR, compatible with both traditional single or dual index primers.Molecular Tagging

[0043] In some embodiments, the nucleic acid molecules of the sample may be tagged with sample indexes, partition tags and / or molecular barcodes (referred to generally as “tags”). Tags can form part of an adapter.

[0044] Tags can be molecules, such as nucleic acids, containing information that indicates a feature of the molecule with which the tag is associated. For example, molecules can bear a sample tag or sample index (which distinguishes molecules in one sample from those in a different sample), a partition tag (which distinguishes molecules in one partition from those in a different partition) and / or a molecular tag / molecular barcode / barcode (which distinguishes different molecules from one another (in both unique and non-unique tagging scenarios). In certain embodiments, a tag can comprise one or a combination of barcodes. As used herein, the term “barcode” refers to a nucleic acid molecule having a particular nucleotide sequence, or to the nucleotide sequence, itself, depending on context. A barcode can have, for example, between 10 and 100 nucleotides. A collection of barcodes can have degenerate sequences or can have sequences having a certain Hamming distance, as desired for the specific purpose. So, for example, a molecular barcode can be comprised of oneAttorney Docket No. GH0240WO barcode or a combination of two barcodes, each attached to different ends of a molecule. Additionally or alternatively, for different partitions and / or samples, different sets of molecular barcodes, molecular tags, or molecular indexes can be used such that the barcodes serve as a molecular tag through their individual sequences and also serve to identify the partition and / or sample to which they correspond based the set of which they are a member. For example, barcodes can be used to allow the origin of the DNA (e.g., the subject, biological sample (e.g., samples collected at various time points), enriched DNA sample (e.g., enriched DNA comprising an epigenetic target region set or enriched DNA comprising a sequence-variable target region set), partition, or similar) to be identified, e.g., following pooling of a plurality of samples for parallel sequencing.

[0045] In the methods of the disclosure, partitioning results in the generation of multiple subsamples (i.e. partitions) based on the presence or absence of 5hmC nucleic acid bases in the sample nucleic acids. Tags can be used to label the nucleic acids in each partition so as to correlate the tag (or tags) with a specific partition. For example, if multiple subsamples are carried forward after the partitioning step, tags can be used to label each of the subsamples such that the corresponding sequence data deriving from each subsample can be identified. In some embodiments, a single tag can be used to label a specific partition. In some embodiments, multiple different tags can be used to label a specific partition. In embodiments employing multiple different tags to label a specific partition, the set of tags used to label one partition can be readily differentiated for the set of tags used to label other partitions. In some embodiments, the tags may have additional functions, for example the tags can be used to index sample sources or used as unique molecular identifiers (which can be used to improve the quality of sequencing data by differentiating sequencing errors from mutations, for example as in Kinde et al., Proc Nat’l Acad Sci USA 108: 9530-9535 (2011), Kou et al., PLoS ONE,11 : eO 146638 (2016)) or used as non-unique molecule identifiers, for example as described in US Pat. No. 9,598,731. Similarly, in some embodiments, the tags may have additional functions, for example the tags can be used to index sample sources or used as non-unique molecular identifiers (which can be used to improve the quality of sequencing data by differentiating sequencing errors from mutations).

[0046] Tags may be incorporated into or otherwise joined to adapters by chemical synthesis, ligation (e.g., as described above, e.g., by blunt-end ligation or sticky-end ligation), or overlap extension polymerase chain reaction (PCR), among other methods. Such adapters are ultimately joined to the target nucleic acid molecule. In other embodiments, one or more rounds of amplification cycles (e.g., PCR amplification) may be applied to introduce sampleAttorney Docket No. GH0240WO indexes to a nucleic acid using conventional nucleic acid amplification methods. The amplifications may be conducted in one or more reaction mixtures (e.g., a plurality of microwells in an array). Molecular barcodes and / or sample indexes may be introduced simultaneously, or in any sequential order. In some embodiments, molecular barcodes and / or sample indexes are introduced prior to and / or after the conversion procedure. In some embodiments, molecular barcodes and / or sample indexes are introduced prior to and / or after the partitioning step. In some embodiments, molecular barcodes and / or sample indexes are introduced prior to and / or after sequence capturing steps, if present, are performed. In some embodiments, only the molecular barcodes are introduced prior to probe capturing and the sample indexes are introduced after sequence capturing steps are performed. In some embodiments, both the molecular barcodes and the sample indexes are introduced prior to performing probe-based sequence capturing steps, if present. In some embodiments, the sample indexes are introduced after sequence capturing steps are performed, if present. In some embodiments, sample indexes are incorporated through overlap extension polymerase chain reaction (PCR).

[0047] In some embodiments, the tags may be located at one end or at both ends of the sample nucleic acids. In some embodiments, tags are predetermined or random or semirandom sequences. In some embodiments, the tag(s) may together be less than about 500, 200, 100, 50, 20, 10, 9, 8, 7, 6, 5, 4, 3, 2, or 1 nucleotide in length. Typically, tags are about 5 to 20 or 6 to 15 nucleotides in length. The tags may be linked to sample nucleic acids randomly or non-randomly.

[0048] In some embodiments, each sample is uniquely tagged with a sample index or a combination of sample indexes. In some examples, when multiple subsamples (i.e. partitions) are subsequently processed after the partitioning step, each partition can be uniquely tagged with a partition tag or a combination of partition tags. In some embodiments, each nucleic acid molecule of a sample or subsample is uniquely tagged with a molecular barcode or a combination of molecular barcodes. In other embodiments, a plurality of molecular barcodes may be used such that molecular barcodes are not necessarily unique to one another in the plurality (e.g., non-unique molecular barcodes). In these embodiments, molecular barcodes are generally attached (e.g., by ligation) to individual nucleic acid molecules such that the combination of the molecular barcode and the sequence of the sample nucleic acid that it is attached to creates a unique sequence that may be individually tracked. Detection of nonunique molecular barcodes in combination with endogenous sequence information typically allows for the assignment of a unique identity to a particular molecule. Endogenous sequenceAttorney Docket No. GH0240WO information includes the beginning (start) and / or end (stop) genomic location / position corresponding to the sequence of the original nucleic acid molecule in the sample, start and stop genomic positions corresponding to the sequence of the original nucleic acid molecule in the sample, the beginning (start) and / or end (stop) genomic location / position of the sequence read that is mapped to the reference sequence, start and stop genomic positions of the sequence read that is mapped to the reference sequence, sub-sequences of sequence reads at one or both ends, length of sequence reads, and / or length of the original nucleic acid molecule in the sample. In some embodiments, beginning region includes the first 1, first 2, the first 5, the first 10, the first 15, the first 20, the first 25, the first 30 or at least the first 30 base positions at the 5' end of the sequencing read that align to the reference sequence. In some embodiments, the end region comprises the last 1, last 2, the last 5, the last 10, the last 15, the last 20, the last 25, the last 30 or at least the last 30 base positions at the 3' end of the sequencing read that align to the reference sequence. The length, or number of base pairs, of an individual sequence read are also optionally used to assign a unique identity to a given molecule. As described herein, fragments from a single strand of nucleic acid having been assigned a unique identity, may thereby permit subsequent identification of fragments from the parent strand, and / or a complementary strand.

[0049] In certain embodiments, the number of different tags used to uniquely identify a number of molecules, z, in a class can be between any of 2*z, 3*z, 4*z, 5*z, 6*z, 7*z, 8*z, 9*z, 10*z, 11 *z, 12*z, 13*z, 14*z, 15*z, 16*z, 17*z, 18*z, 19*z, 20*z or 100*z (e.g., lower limit) and any of 100,000*z, 10,000*z, 1000*z or 100*z (e.g., upper limit). In some embodiments, molecular barcodes are introduced at an expected ratio of a set of identifiers (e.g., a combination of unique or non-unique molecular barcodes) to molecules in a sample. One example format uses from about 2 to about 1,000,000 different molecular barcode sequences, or from about 5 to about 150 different molecular barcode sequences, or from about 20 to about 50 different molecular barcode sequences, ligated to both ends of a target molecule. Alternatively, from about 25 to about 1,000,000 different molecular barcode sequences may be used. For example, 20-50 x 20-50 molecular barcode sequences (i.e., one of the 20-50 different molecular barcode sequences can be attached to each end of the target molecule) can be used. Such numbers of identifiers are typically sufficient for different molecules having the same start and stop points to have a high probability (e.g., at least 94%, 99.5%, 99.99%, or 99.999%) of receiving different combinations of identifiers.

[0050] In some embodiments, the assignment of unique or non-unique molecular barcodes in reactions is performed using methods and systems described in, for example, U.S. PatentAttorney Docket No. GH0240WOApplication Nos. 20010053519, 20030152490, and 20110160078, and U.S. Patent Nos. 6,582,908, 7,537,898, 9,598,731, and 9,902,992, each of which is hereby incorporated by reference in its entirety. Alternatively, in some embodiments, different nucleic acid molecules of a sample may be identified using only endogenous sequence information (e.g., start and / or stop positions, sub-sequences of one or both ends of a sequence, and / or lengths). The addition of tags (e.g., sample indexes, partition tags and / or molecular barcodes) to nucleic acids can be done through amplification, wherein the tags are comprised in primers used for amplification.

[0051] In some embodiments, the nucleic acids are ligated to adapters comprising molecular barcodes. These molecular barcodes (optionally in combination with endogenous sequence information) can then be used when analyzing the sequencing data to group sequence reads deriving from the same parent nucleic acids (i.e. those nucleic acids prior to any amplification). The grouped sequence reads can then be analyzed, for example, to determine a consensus sequence for parent nucleic acids. The consensus sequence will include any converted bases and thus can be used to determine the methylation status of the parent nucleic acid. Similarly, the abundance of consensus sequences from a subsample at C positions in a reference can be used to determine the 5hmC status of the parent nucleic acids. For instance, when the base coverage of a specific C position in a reference sequence is higher than other C positions in a subsample which has been enriched for 5hmC, that specific C position on that parent nucleic acid can be identified as comprising a 5hmC modification at that C position.Conversion Procedure

[0052] The conversion procedures which are used in the methods of the disclosure can either convert: (i) the base pairing specificity of 5mC (e.g. Tet-assisted conversion with a substituted borane reducing agent); or (ii) the base pairing specificity of unmethylated cytosines (e.g., bisulfite conversion). Preferably the methods of the disclosure employ conversion procedures which convert the base pairing specificity of 5mC because such methods allow for increased sensitivity when using the sequencing data to also detect genetic variants.

[0053] In conversion procedures wherein the base pairing specificity of unmethylated cytosines is converted, it is difficult to identify the presence or absence of somatic mutations of cytosines in the sample nucleic acids. In contrast, when conversion procedures which convert the base pairing specificity of 5mC are used, unmethylated cytosines are retained, thus allowing OT / G>A somatic mutations to be detected with high confidence. Moreover,Attorney Docket No. GH0240WO conversion procedures which convert the base pairing specificity of 5mC (such as TAPS P and DM-Seq) are generally not as destructive as conversion procedures which convert the base pairing specificity of unmethylated cytosines (e.g., bisulfite sequence), and thus the fragmentation pattern of the sample nucleic acids is retained. This can be advantageous, e.g., in the analysis of cfDNA. Accordingly, the use of conversion procedures which convert the base pairing specificity of 5mC additionally allows for both sensitive mutation detection and the analysis of the sample nucleic acid fragmentation pattern.

[0054] There are various methods of detecting and / or identifying methylated cytosines that rely on a conversion procedure that changes the base-pairing specificity of a cytosine, based on its methylation status. These changes of base-pairing specificity can then be detected, and thus the methylation status of the cytosine inferred, by sequencing.

[0055] The methods of the present disclosure involve subjecting the nucleic acids to a conversion procedure that selectively converts the base pairing specificity of 5- methylcytosines (5mC) or unmethylated cytosines (C).

[0056] Procedures that selectively convert the base pairing specificity of 5mC refer to methods which convert the base pairing specificity of 5mC but not C. Such procedures can include methods which involve conditions which would also result in the conversion of the base pairing specific of unprotected 5hmC (e.g., TAPS), provided that, in the methods of the disclosure, any 5hmC is protected (e.g., by glucosylation) from conversion.

[0057] Procedures that selectively convert the base pairing specificity of C refer to methods which convert the base pairing specificity of C but not 5mC. Such procedures can include methods which involve conditions which would also result in the conversion of the base pairing specific of unprotected 5hmC (e.g., oxidative bisulfite sequencing), provided that, in the methods of the disclosure, any 5hmC is protected from conversion.

[0058] In some embodiments, the conversion procedure used in the methods of the disclosure is one that changes the base pairing specificity of 5mC, but does not change the base pairing specificity of unmethylated cytosines. Advantages of methods that do not convert the basepairing specificity of unmethylated cytosines include reduced loss of sequence complexity, higher sequencing efficiency and reduced alignment losses. Additionally, methods such as TAPS, TAPS P, and DM-Seq may in some cases be preferred over methods such as bisulfite sequencing because they are less destructive (especially important for low yield samples such as cfDNA) and do not require denaturation, meaning that non-conversion errors are theoretically more likely to be random. In methods that require denaturation for conversion, failure to denature a DNA molecule will result in non-conversion of all bases in the DNAAttorney Docket No. GH0240WO molecule. As biological changes in methylation are predominantly concerted to a localized regions of interest, these non-random (localized) conversion can appear as false negatives (non-methylated regions). Random non-conversion methods can maximally affect a low percent of bases within a region, and thus the specificity of methylation change detection can be maximized (reduce false positives) by placing a threshold on the percentage of bases within a region that are methylated / non-methylated. Hence, in some embodiments, a conversion procedure that does not involve denaturation is preferred.

[0059] In other embodiments, the conversion procedure used in the methods of the disclosure is one that changes the base pairing specificity of an unmethylated cytosine, but does not change the base pairing specificity of 5mC. Such methods include, for example, bisulfite sequencing and EM-seq.

[0060] In some embodiments, the conversion procedure converts the base pairing specificity 5mC. In some embodiments, the conversion procedure which converts the base pairing specificity of 5mC comprises protection of 5hmC (e.g., using P-glucosyltransferase (PGT) or 5-hydroxymethylcytosine carbamoyltransferase) combined with Tet-assisted conversion with a substituted borane reducing agent, e.g., 2-picoline borane, borane pyridine, tert-butylamine borane, ammonia borane or pyridine borane. In this method, 5hmC can be protected from conversion, for example through glucosylation using P-glucosyltransferase (PGT), forming 5- glucosylhydroxymethylcytosine (5ghmC), or through carbamoylation using 5- hydroxymethylcytosine carbamoyltransferase, forming 5cmC. A method of protecting 5hmC from conversion, for example through glucosylation using P-glucosyltransferase (PGT), forming 5-glucosylhydroxymethylcytosine (5ghmC), is described in Yu et al., Cell 2012;149: 1368-80. Alternatively, a carbamoyltransferase enzyme, such as 5- hydroxymethylcytosine carbamoyltransferase as described in Yang et al., Bio-protocol, 2023; 12(17): e4496, can be used to protect hmC (by converting hmC to 5- carbamoyloxymethylcytosine (5cmC)), then a TET protein, such as mTetl or a TET2 comprising a T1372S mutation, can be used to convert mC to caC, and then bisulfite treatment can be used to convert C and caC to U while 5cmC remains unaffected. In this method, 5hmC can be protected from conversion, for example through glucosylation using P- glucosyltransferase (PGT), forming 5-glucosylhydroxymethylcytosine (5ghmC). Treatment with a TET protein, such as mTetl or a TET2 comprising a T1372S mutation, then converts 5mC to 5caC but does not convert C, 5ghmC, or 5cmC. 5caC is then converted to DHU by treatment with pic-borane or another substituted borane reducing agent such as borane pyridine, tert-butylamine borane, or ammonia borane, also without affecting 5ghmC, 5cmC,Attorney Docket No. GH0240WO or unmethylated C. Sequencing of the converted DNA identifies positions that are read as cytosine as being either 5hmC or unmethylated C positions. Meanwhile, positions that are read as T are identified as being T or 5mC. T and 5mC can be distinguished through alignment to a reference sequence. When the corresponding position in a reference sequence is T, the nucleoside on the sample nucleic acid is identified as a T. When the corresponding position in a reference sequence is C, the nucleoside on the sample nucleic acid is identified as a 5mC. For an exemplary description of this type of conversion, see, e.g., Liu et al., Nature Biotechnology 2019; 37:424-429. 5-hydroxymethylcytosine carbamoyltransferase is described in Yang et al., Bio-protocol, 2023; 12(17): e4496. Performing such conversion methods (e.g., TAPS P conversion) on a sample as described herein thus facilitates distinguishing positions containing unmethylated C or 5hmC on the one hand from positions containing 5mC using the sequence reads obtained. The unmethylated C can then be distinguished from the 5hmC by analyzing the sequence data and using the base coverage analysis of subsamples from the partitioning step, wherein higher base coverage of cytosines in the subsample enriched for 5hmC would indicate that those cytosines were 5hmC in the sample nucleic acids corresponding to those sequence reads.

[0061] In some embodiments, the conversion procedure which converts the base pairing specificity of 5mC comprises reacting the nucleic acids with a variant methyltransferase having carboxymethyltransferase activity in the presence of carboxy-S-adenosyl-L- methionine (CxSAM) substrate, thereby labelling any unmethylated C and rendering it resistant to deaminase action. When this method is used in the context of the present disclosure, the 5hmC nucleic acid bases are also protected from deaminase action, e.g., through glucosylation such as by PGT. In some embodiments, PGT and CxMTase reactions occur simultaneously. The nucleic acids can then be contacted with a deaminase enzyme (e.g., APOBEC3A) which deaminates 5mC to uracil. Sequencing of the converted DNA identifies positions that are read as cytosine as being either 5hmC or unmethylated C positions. Meanwhile, positions that are read as T are identified as being T or 5mC. T and 5mC can be distinguished through alignment to a reference sequence. When the corresponding position in a reference sequence is T, the nucleoside on the sample nucleic acid is identified as a T. When the corresponding position in a reference sequence is C, the nucleoside on the sample nucleic acid is identified as a 5mC. In some embodiments, the variant methyltransferase having carboxymethylase activity is a recombinant M.MpelN374K, for example. In some embodiments, the deaminase enzyme is APOBEC3A. For an exemplary description of this type of conversion, known as DM-seq, see WO2021 / 236778.Attorney Docket No. GH0240WO

[0062] In some embodiments, the conversion procedure converts the base pairing specificity of unmethylated cytosines. In some embodiments, the conversion procedure which converts unmethylated cytosines comprises bisulfite conversion. Treatment with bisulfite converts unmethylated cytosine to uracil whereas 5mC and 5hmC are not converted. Thus, where bisulfite conversion is used, the converted nucleobases are inferred as comprising unmethylated cytosine. The unconverted nucleobases are inferred as comprising 5mC and / or 5hmC. Sequencing of bisulfite-treated DNA identifies positions that are read as cytosine as being 5mC or 5hmC. Meanwhile, positions that are read as T are identified as being T or unmethylated cytosine. Thus, performing bisulfite conversion as described herein thus facilitates identifying positions containing 5mC or 5hmC versus positions containing unmethylated C. The 5mC can then be distinguished from the 5hmC by analyzing the sequence data and using the base coverage analysis of subsamples from the partitioning step, wherein higher base coverage of cytosines in the subsample enriched for 5hmC would indicate that those cytosines were 5hmC in the sample nucleic acids corresponding to those sequence reads. For an exemplary description of bisulfite conversion, see, e.g., Moss et al., Nat Commun. 2018; 9: 5068.

[0063] In some embodiments, the conversion procedure converts unmethylated cytosines and comprises a non-specific, modification-sensitive double-stranded DNA deaminase, e.g., as in SEM-seq. See, e.g., Vaisvila et al. (2023) Discovery of novel DNA cytosine deaminase activities enables a nondestructive single-enzyme methylation sequencing method for base resolution high-coverage methylome mapping of cell-free and ultra-low input DNA. bioRxiv; DOI: 10.1101 / 2023.06.29.547047, available at www.biorxiv.org / content / 10.1101 / 2023.06.29.547047vl. SEM-Seq employs a non-specific, modification-sensitive double-stranded DNA deaminase (MsddA) in a nondestructive singleenzyme 5-methylctyosine sequencing (SEM-seq) method that deaminates unmodified cytosines. Accordingly, SEM-seq does not require the TET2 and T4-PGT or 5- hydroxymethylcytosine carbamoyltransferase protection and denaturing steps that are of use, e.g., in APOEC3 A-based protocols. Additionally, MsddA does not deaminate 5-formylated cytosines (5fC) or 5-carboxylated cytosines (5caC). In SEM-seq, unmodified cytosines in the DNA are deaminated to uracil and is read as “T” during sequencing. Modified cytosines (e.g., 5mC) are not converted and are read as “C” during sequencing. Cytosines that are read as thymines are identified as unmodified (e.g., unmethylated) cytosines or as thymines in the DNA. Performing SEM-seq conversion thus facilitates identifying positions containing 5mC using the sequence reads obtained. In some embodiments, the conversion procedure whichAttorney Docket No. GH0240WO converts unmethylated cytosines comprises enzymatic conversion of the first nucleobase using MsddA.

[0064] In some embodiments, the conversion procedure is an enzymatic conversion procedure which converts the base pairing specificity of modified nucleosides (e.g., DM-seq conversion comprising adding a protective group (such as a carboxymethyl group) to unmodified cytosines, and deaminating 5mC, such as using an APOBEC enzyme) or an enzymatic conversion procedure which converts the base pairing specificity of unmodified nucleosides (such as SEM-seq).

[0065] In some embodiments, the conversion procedure used in the methods of the disclosure is one that changes the base pairing specificity of a modified nucleoside (e.g., methylated cytosine), but does not change the base pairing specificity of the corresponding unmodified nucleoside (e.g., cytosine) or does not change the base pairing specificity of any un-modified nucleoside (e.g., cytosine, adenosine, guanosine and thymidine (or uracil)). Advantages of methods that do not convert the base-pairing specificity of unmodified nucleosides include reduced loss of sequence complexity, higher sequencing efficiency and reduced alignment losses. Additionally, methods such as DM-seq may in some cases be preferred over methods such as bisulfite sequencing and EM-seq because they are less destructive (especially important for low yield samples such as cfDNA) and do not require denaturation, meaning that non-conversion errors are theoretically more likely to be random. In methods that require denaturation for conversion, failure to denature a DNA molecule will result in nonconversion of all bases in the DNA molecule. As biological changes in methylation are predominantly concerted to a localized region of interest, these non-random (localized) conversion can appear as false negatives (non-methylated regions). Random non-conversion methods can maximally affect a low percent of bases within a region, and thus the specificity of methylation change detection can be maximized (reduce false positives) by placing a threshold on % of bases within a region that are methylated / non-methylated. Hence, in some cases, a conversion procedure that does not involve denaturation is preferred.

[0066] In other cases, the conversion procedure used in the methods of the disclosure is one that changes the base pairing specificity of an unmodified nucleoside (e.g., cytosine), but does not change the base pairing specificity of the corresponding modified nucleoside (e.g., methylated cytosine).

[0067] The skilled person can select a suitable method according to their needs, including which nucleoside modifications are to be detected and / or identified.Attorney Docket No. GH0240WO

[0068] In some embodiments, the conversion procedure converts modified nucleosides. In some embodiments, the conversion procedure which converts modified nucleosides comprises enzymatic conversion, such as DM-seq, for example, as described in WO2023 / 288222A1. In DM-seq, unmodified cytosines in the DNA are enzymatically protected from a subsequent deamination step wherein 5mC in 5mCpG is converted to T. The enzymatically protected unmodified (e.g., unmethylated) cytosines are not converted and are read as “C” during sequencing. Cytosines that are read as thymines (in a CpG context) are identified as methylated cytosines in the DNA.

[0069] Thus, when this type of conversion is used, the first nucleobase comprises unmodified (such as unmethylated) cytosine, and the second nucleobase comprises modified (such as methylated) cytosine. Sequencing of the converted DNA identifies positions that are read as cytosine as being unmodified C positions. Meanwhile, positions that are read as T are identified as being T or 5mC. Performing DM-seq conversion thus facilitates identifying positions containing 5mC using the sequence reads obtained.

[0070] Exemplary cytosine deaminases for use herein include APOBEC enzymes, for example, APOBEC3 A. Generally, AID / APOBEC family DNA deaminase enzymes such as APOBEC3A (A3 A) are used to deaminate (unprotected) unmodified cytosine and 5mC. For an exemplary description of APOBEC conversion, see, e.g., Schutsky et al., Nature Biotechnology 2018; 36: 1083-1090.

[0071] The enzymatic protection of unmodified cytosines in the DNA comprises addition of a protective group to the unmodified cytosines. Such protective groups can comprise an alkyl group, an alkyne group, a carboxyl group, a carboxyalkyl group, an amino group, a hydroxymethyl group, a glucosyl group, a glucosylhydroxymethyl group, an isopropyl group, or a dye. For example, DNA can be treated with a methyltransferase, such as a CpG-specific methyltransferase, which adds the protective group to unmodified cytosines. The term methyltransferase is used broadly herein to refer to enzymes capable of transferring a methyl or substituted methyl (e.g., carboxymethyl) to a substrate (e.g., a cytosine in a nucleic acid). In some embodiments, the DNA is contacted with a CpG-specific DNA methyltransferase (MTase), such as a CpG-specific carboxymethyltransferase (CxMTase), and a substituted methyl donor, such as a carboxymethyl donor (e.g., carboxymethyl-S-adenosyl-L- methionine). See, e.g., WO2021 / 236778A2. In particular embodiments, the CxMTase can facilitate the addition of a protective carboxymethyl group to an unmethylated cytosine. In some embodiments, the unmethylated cytosine is unmodified cytosine. The carboxymethyl group can prevent deamination of the cytosine during a deamination step (such as aAttorney Docket No. GH0240WO deamination step using an APOBEC enzyme, such as A3 A). Substituted methyl or carboxymethyl donors useful in the disclosed methods include but are not limited to, S- adenosyl-L-methionine (SAM) analogs, optionally wherein the SAM analog is carboxy-S- adenosyl-L-methionine (CxSAM). SAM analogs are described, for example, in WO2022 / 197593A1. The MTase may be, for example, a CpG methyltransferase from Spiroplasma sp. strain MQ1 (M.SssI), DNA-methyltransferase 1 (DNMT1), DNA- methyltransferase 3 alpha (DNMT3 A), DNA-methyltransferase 3 beta (DNMT3B), or DNA adenine methyltransferase (Dam). The CxMTase may be a CpG methyltransferase from Mycoplasma penetrans (M.Mpel). In a particular embodiment, the methyltransferase enzyme is a variant of M.Mpel, or a sequence at least 90%, at least 92%, at least 94%, at least 96%, at least 97%, at least 98%, or at least 99% identical thereto, optionally wherein the amino acid corresponding to position 374 is R or K.

[0072] In one embodiment, the methyltransferase enzyme is a variant of M.Mpel having an N374R substitution or an N374K substitution. The methyltransferase can further comprise one or more amino acid substitutions selected from a) substitution of one or both residues T300 and E305 with S, A, G, Q, D, or N; b) substitution of one or more residues A323, N306, and Y299 with a positively charged amino acid selected from K, R or H; and / or c) substitution of S323 with A, G, K, R or H, which may enhance the activity of the enzyme.

[0073] Optionally, the conversion procedure further includes enzymatic protection of 5hmCs, such as by glucosylation of the 5hmCs (e.g., using PGT) or by carbamoylation of the 5hmCs (e.g., using 5-hydroxymethylcytosine carbamoyltransferase), in the DNA prior to the deamination of unprotected modified cytosines. In this method, 5hmC can be protected from conversion, for example through glucosylation using P-glucosyl transferase (PGT), forming (5-glucosylhydroxymethylcytosine) 5ghmC, or through carbamoylation using 5- hydroxymethylcytosine carbamoyltransferase, forming 5cmC. This is described, for example, in Yu et al., Cell 2012; 149: 1368-80, and in Yang et al., Bio-protocol, 2023; 12(17): e4496. Glucosylation or carbamoylation of 5hmC can reduce or eliminate deamination of 5hmC by a deaminase such as APOBEC3A. Treatment with an MTase or CxMTase then adds a protecting group to unmodified (unmethylated) cytosines in the DNA. 5mC (but not protected, unmodified cytosine and not 5ghmC or 5cmC) is then deaminated (converted to T in the case of 5mC) by treatment with a deaminase, for example, an APOBEC enzyme (such as APOBEC3 A). Sequencing of the converted DNA identifies positions that are read as cytosine as being either 5hmC or unmodified C positions. Meanwhile, positions that are read as T are identified as being T or 5mC. Performing DM-seq conversion with glucosylation ofAttorney Docket No. GH0240WO5hmC on a sample as described herein thus facilitates distinguishing positions containing unmodified C or 5hmC on the one hand from positions containing 5mC using the sequence reads obtained.

[0074] Also provided herein are methods in which alternative base conversion schemes are used. For example, unmethylated cytosines can be left intact while methylated cytosines and hydroxymethylcytosines are converted to a base read as a thymine (e.g., uracil, thymine, or dihydrouracil).

[0075] In some embodiments, methylating a cytosine in at least one first complementary strand or second complementary strand comprises contacting the cytosine with a methyltransferase such as DNMT1 or DNMT5. In such embodiments, the step of oxidizing a 5-hydroxymethylated cytosine to 5-formylcytosine (such as by contacting the 5- hydroxymethyl cytosine in a first strand and a second strand with KRuO4) can be optional.

[0076] In some embodiments, converting the modified cytosine in at least one first or second strand to a thymine or a base read as thymine comprises oxidizing a hydroxymethyl cytosine, e.g., the hydroxymethyl cytosine is oxidized to formylcytosine. In some embodiments, oxidizing the hydroxymethyl cytosine to formylcytosine comprises contacting the hydroxymethyl cytosine with a ruthenate, such as potassium ruthenate (KRuO4).

[0077] In some embodiments, the modified cytosine is converted to thymine, uracil, or dihydrouracil. In any such embodiments, amplification methods may comprise uracil- and / or dihydrouracil-tolerant amplification methods, such as PCR using a uracil- and / or dihydrouracil-tolerant DNA polymerase.

[0078] In some embodiments, the method comprises converting a formylcytosine and / or a methylcytosine to carboxylcytosine as part of converting the modified cytosine in at least one first or second strand to a thymine or a base read as thymine. For example, converting the formylcytosine and / or the methylcytosine to carboxylcytosine can comprise contacting the formylcytosine and / or the methylcytosine with a TET enzyme, such as TET1, TET2, TET3, or a TET2 comprising a T1372S mutation. In some embodiments, the method comprises reducing the carboxylcytosine as part of converting the modified cytosine in at least one first or second strand to a thymine or a base read as thymine, and / or the carboxylcytosine is reduced to dihydrouracil. In some embodiments, reducing the carboxylcytosine comprises contacting the carboxylcytosine with a borane or borohydride reducing agent.

[0079] In some embodiments, the borane or borohydride reducing agent comprises pyridine borane, 2-pi coline borane, borane, tert-butylamine borane, ammonia borane, sodium borohydride, sodium cyanoborohydride (NaBH3CN), lithium borohydride (LiBH4),Attorney Docket No. GH0240WO ethylenediamine borane, dimethylamine borane, sodium triacetoxyborohydride, morpholine borane, 4-methylmorpholine borane, trimethylamine borane, dicyclohexylamine borane, or a salt thereof. In other embodiments, the reducing agent comprises lithium aluminum hydride, sodium amalgam, amalgam, sulfur dioxide, dithionate, thiosulfate, iodide, hydrogen peroxide, hydrazine, diisobutylaluminum hydride, oxalic acid, carbon monoxide, cyanide, ascorbic acid, formic acid, dithiothreitol, beta-mercaptoethanol, or any combination thereof.Partitioning

[0080] The methods of the disclosure employ a partitioning step, wherein nucleic acids are partitioned into two or more partitions (i.e. subsamples) based on the presence or absence of 5hmC nucleic acid bases in the sample nucleic acids. The partitioning step can be performed before or after the conversion step. When the partitioning step is performed before the conversion step, one or more (e.g., both) subsamples can be carried forward to the subsequent conversion, amplification and sequencing steps. Similarly, when the partitioning step is performed after the conversion step, one or more (e.g., both) subsamples can be carried forward to the subsequent amplification and sequencing steps. When multiple subsamples are carried forward, adapters comprising partition tags can be applied to each of the subsamples such that the subsamples can be sequenced in the same sequencing reaction while still allowing the sequencing data from each subsample to be distinguished. Tagged partitions can therefore be pooled together for collective sample prep and / or sequencing. Partitioning can be performed using an agent which: (i) directly binds 5hmC; (ii) binds to a derivative of 5hmC; or (iii) binds to an isolation tag which has been conjugated to 5hmC.

[0081] In some embodiments, partitioning comprises exposing the nucleic acids to a binding agent which selectively binds 5hmC relative to 5mC. In some embodiments, the binding agent is an anti-5hmC antibody or an antigen binding fragment thereof. Exemplary antibodies include the antibody under catalog number 39069 from Active Motif.

[0082] In some embodiments, partitioning comprises exposing the nucleic acids to a binding agent which selectively binds to an isolation tag which has been conjugated to 5hmC. The isolation tag may be conjugated to 5hmC through chemical labeling such as through “click chemistry”. In some embodiments, the conjugation of the isolation tag comprises: (i) incubating the nucleic acids with a P-glucosyltransferase and UDP glucose modified with a chemoselective group, thereby covalently labelling the 5hmC with the chemoselective group; and (ii) linking a biotin moiety to the chemoselectively-modified 5hmC via a cycloaddition reaction. The partitioning can then be performed by binding the product of step (ii) to a support that binds to biotin (e.g., beads comprising streptavidin, such as magnetic beadsAttorney Docket No. GH0240WO comprising streptavidin). In some embodiments, the UDP glucose modified with chemoselective group is UDP-6-N3-Glu. In some embodiments, the biotin moiety is dibenzocyclooctyne-modified biotin. In some embodiments, the P-glucosyltransferase is T4 DNA P-glucosyltransferase.

[0083] The exemplary workflow shown in Figure 3 shows the “5hmC-SEAL” method. B- glucosyltransferase (BGT) is first applied to DNA with a UDP-6-N3-Glu substrate. This reacts selectively with 5hmC bases, resulting in a glucose moiety and N3 being transferred to the 5hmC. Standard copper-free (Cu-free) click chemistry with DBCO-biotin then is performed, in which the DBCO and N3 react, transferring biotin to the 5hmC-originating base. Streptavidin-magnetic beads can then be applied to the nucleic acid sample to isolate (‘pull-down’) the biotinylated-DNA, corresponding to originating nucleic acids containing 5hmC bases. Accordingly, the workflow of Figure 3 provides at least two subsamples of nucleic acids, wherein a first subsample is enriched for nucleic acids comprising 5hmC nucleic acid bases and wherein a second subsample is depleted of nucleic acids comprising 5hmC nucleic acid bases. Such a method is described in WO 2017 / 176630, which is incorporated herein by reference in its entirety.

[0084] In some embodiments, partitioning comprises exposing the nucleic acids to a binding agent which selectively binds to an isolation tag which has been conjugated to 5hmC. The isolation tag may be conjugated to 5hmC through chemical labeling. The isolation tag may be a glucose residue conjugated to 5hmC DNA (i.e. the glucose residue in P-glucosylated- 5hmC). In some embodiments, the conjugation of the isolation tag comprises incubating the nucleic acids with a P-glucosyltransferase and UDP -glucose, thereby covalently labelling the 5hmC with a P-glucosyl residue. The partitioning can then be performed by exposing the nucleic acids to an agent with binds glucosylated 5hmC (e.g., J-binding protein 1 (JBP1), such as biotinylated JBP1 or JBP1 bound to a support). In some embodiments, the biotinylated JBP1, and any bound nucleic acids, can then be isolated using a support (e.g., beads) comprising streptavidin. In some embodiments, the P-glucosyltransferase is T4 DNA P-glucosyltransferase. Exemplary methods are known as the JBP-l-seq method, as described in Cui et al., Genomics. 2014 p368-375, which is incorporated by reference.

[0085] After partitioning, either or both of the subsamples can be amplified and sequenced. The partition depleted in 5hmC containing nucleic acids may also be amplified and sequenced. The sequenced nucleic acids in the enriched partition can be deemed to have contained a 5hmC base at one of the cytosines present in the sequence. As 5hmC bases areAttorney Docket No. GH0240WO relatively rare in nature, by analyzing per base coverage across sequenced nucleic acids, the location of the 5hmC bases can be estimated with high confidence.Amplification

[0086] Sample nucleic acids flanked by adapters can be amplified by PCR and other amplification methods. Amplification is typically primed by primers binding to primer binding sites in adapters flanking a DNA molecule to be amplified. Amplification methods can involve cycles of denaturation, annealing and extension, resulting from thermocycling or can be isothermal as in transcription-mediated amplification. Other amplification methods include the ligase chain reaction, strand displacement amplification, nucleic acid sequence based amplification, and self-sustained sequence based replication.

[0087] In some embodiments, the present methods perform dsDNA ligations with T-tailed and C-tailed adapters when the sample nucleic acids have been subjected to A-tailing, e.g., using T4 polymerase or KI enow large fragment. This increases the efficiency of ligation and results in amplification of at least 50, 60, 70 or 80% of double stranded nucleic acids. Such methods can increase the amount or number of amplified molecules relative to control methods performed with T-tailed adapters alone by at least 10, 15 or 20%.

[0088] Amplification is performed after the conversion and partitioning steps. Amplification may be performed before or after any sequence capture step. In some embodiments, the ligating occurs before or simultaneously with amplification. In some embodiments, amplification is primed by primer binding to primer binding site(s) in the adapter(s).Capturing using capture probes

[0089] Nucleic acids in a sample can be subject to a sequence capture step, in which molecules having target sequences are captured for subsequent analysis. Capture may be performed using any suitable approach known in the art. Target capture can involve use of a bait set comprising oligonucleotide baits labeled with a capture moiety, such as biotin or the other examples noted below. The probes can have sequences selected to tile across a panel of regions, such as genes. Such bait sets are combined with a sample under conditions that allow hybridization of the target molecules with the baits. Then, captured molecules are isolated using the capture moiety. For example, a biotin capture moiety by bead-based streptavidin. Such methods are further described in, for example, U.S. patent 9,850,523, issuing December 26, 2017, which is incorporated herein by reference.Attorney Docket No. GH0240WO

[0090] Capture moieties include, without limitation, biotin, avidin, streptavidin, a nucleic acid comprising a particular nucleotide sequence, a hapten recognized by an antibody, and magnetically attractable particles. The extraction moiety can be a member of a binding pair, such as biotin / streptavidin or hapten / antibody. In some embodiments, a capture moiety that is attached to an analyte is captured by its binding pair which is attached to an isolatable moiety, such as a magnetically attractable particle or a large particle that can be sedimented through centrifugation. The capture moiety can be any type of molecule that allows affinity separation of nucleic acids bearing the capture moiety from nucleic acids lacking the capture moiety. Exemplary capture moieties are biotin which allows affinity separation by binding to streptavidin linked or linkable to a solid phase or an oligonucleotide, which allows affinity separation through binding to a complementary oligonucleotide linked or linkable to a solid phase.

[0091] In some embodiments, the methods herein comprise capturing nucleic acids comprising epigenetic and / or sequence-variable target regions. Such regions may be captured from a sample (e.g., a subsample) that has undergone attachment of adapters, conversion, partitioning, and / or amplification). Enriching for or capturing DNA comprising epigenetic and / or sequence-variable target regions may comprise contacting the DNA with a set of target- specific probes. The set of target-specific probes may have any of the features described herein for sets of target-specific probes, including but not limited to in the embodiments set forth above and the sections relating to probes below. Capturing may be performed on one or more subsamples prepared during methods disclosed herein. In some embodiments, DNA is captured from the first subsample and / or the second subsample, e.g., the first subsample and the second subsample. In some embodiments, the subsamples are differentially tagged (e.g., as described herein) and then pooled before undergoing capture.

[0092] The capturing step may be performed using conditions suitable for specific nucleic acid hybridization, which generally depend to some extent on features of the probes such as length, base composition, etc. Those skilled in the art will be familiar with appropriate conditions given general knowledge in the art regarding nucleic acid hybridization. In some embodiments, complexes of target-specific probes and DNA are formed.

[0093] In some embodiments, methods described herein comprise capturing a plurality of sets of target regions of cfDNA obtained from a subject (e.g., test subject). The target regions comprise intronic regions or VDJ regions that may comprise rearrangements, epigenetic target regions, which may show differences in methylation levels and / or fragmentation patterns depending on whether they originated from a tumor or from healthy cells, andAttorney Docket No. GH0240WO sequence-variable regions, which may show differences in sequence, other than rearrangements, depending on whether they originated from a tumor or from healthy cells. The capturing step produces a captured set of cfDNA molecules. In some embodiments, the cfDNA molecules corresponding to the sequence-variable target region set are captured at a greater capture yield in the captured set of cfDNA molecules than cfDNA molecules corresponding to the epigenetic target region set. In some embodiments, a method described herein comprises contacting cfDNA obtained from a subject (e.g., a test subject) with a set of target- specific probes, wherein the set of target-specific probes is configured to capture cfDNA corresponding to the sequence-variable target region set at a greater capture yield than cfDNA corresponding to the epigenetic target region set. For additional discussion of capturing steps, capture yields, and related aspects, see W02020 / 160414, which is incorporated herein by reference for all purposes.

[0094] It can be beneficial to capture cfDNA corresponding to the sequence-variable target region set at a greater capture yield than cfDNA corresponding to the epigenetic target region set because a greater depth of sequencing may be necessary to analyze the sequence-variable target regions with sufficient confidence or accuracy than may be necessary to analyze the epigenetic target regions. The volume of data needed to determine fragmentation patterns (e.g., to test for perturbation of transcription start sites or CTCF binding sites) or methylation status is generally less than the volume of data needed to determine the presence or absence of cancer-related sequence mutations. Capturing the target region sets at different yields can facilitate sequencing the target regions to different depths of sequencing in the same sequencing run (e.g., using a pooled mixture and / or in the same sequencing cell).

[0095] In some embodiments, amplification is performed before the capturing step. In some embodiments, amplification is performed after the capturing step. In some embodiments, amplification is performed before and after the capturing step. In some embodiments, the methods further comprise sequencing the captured cfDNA to different degrees of sequencing depth for the epigenetic and sequence-variable target region sets and for rearrangements, consistent with the discussion herein.

[0096] In some embodiments, a capturing step is performed with probes for a sequencevariable target region set and probes for an epigenetic target region set in the same vessel at the same time, e.g., the probes for the sequence-variable and epigenetic target region sets are in the same composition. This approach provides a relatively streamlined workflow. In some embodiments, the concentration of the probes for the sequence-variable target region set is greater that the concentration of the probes for the epigenetic target region set.Attorney Docket No. GH0240WO

[0097] Alternatively, a capturing step is performed with a sequence-variable target region probe set in a first vessel and with an epigenetic target region probe set in a second vessel, or a contacting step is performed with a sequence-variable target region probe set at a first time and a first vessel and an epigenetic target region probe set at a second time before or after the first time. This approach allows for preparation of separate first and second compositions comprising captured DNA corresponding to a sequence-variable target region set and captured DNA corresponding to an epigenetic target region set. The compositions can be processed separately as desired (e.g., to partition based on methylation as described herein). These can then be pooled in appropriate proportions to provide material for further processing and analysis such as sequencing.Sequencing

[0098] In general, sample nucleic acids flanked by adapters can be subject to sequencing after amplification. Sequencing methods include, for example, Sanger sequencing, high- throughput sequencing, pyrosequencing, sequencing-by-synthesis, single-molecule sequencing (also known as long-read sequencing or third generation sequencing), nanopore sequencing (a type of long-read sequencing), 5-letter sequencing or 6-letter sequencing, semiconductor sequencing, sequencing-by-ligation, sequencing-by-hybridization, Digital Gene Expression (Helicos), next generation sequencing (NGS), Single Molecule Sequencing by Synthesis (SMSS) (Helicos), enzymatic methyl sequencing (EM-Seq), Tet-assisted pyridine borane sequencing (TAPS), massively-parallel sequencing, Clonal Single Molecule Array (Solexa), shotgun sequencing, Ion Torrent, Oxford Nanopore, Roche Genia, Maxim- Gilbert sequencing, primer walking, and sequencing using PacBio, SOLiD, Ion Torrent, or Nanopore platforms. Sequencing reactions can be performed in a variety of sample processing units, which may include multiple lanes, multiple channels, multiple wells, or other means of processing multiple sample sets substantially simultaneously. Sample processing unit can also include multiple sample chambers to enable processing of multiple runs simultaneously. For example, long-read sequencing (also referred to herein as singlemolecule sequencing or third generation sequencing) methods include those that can generate longer sequencing reads, such as reads in excess of 10 kilobases, as compared to short-read sequencing methods, which generally produce reads of up to about 600 bases in length. Compared to short reads, long reads can improve de novo assembly, transcript isoform identification, and detection and / or mapping of structural variants. Furthermore, long-read sequencing of native DNA or RNA molecules reduces amplification bias and preserves baseAttorney Docket No. GH0240WO modifications, such as methylation status. Long-read sequencing technologies useful herein can include any suitable long-read sequencing methods, including, but not limited to, Pacific Biosciences (PacBio) single-molecule real-time (SMRT) sequencing, Oxford Nanopore Technologies (ONT) nanopore sequencing, and synthetic long-read sequencing approaches, such as linked reads, proximity ligation strategies, and optical mapping. Synthetic long-read approaches comprise assembly of short reads from the same DNA molecule to generate synthetic long reads, and may be used in conjunction with “true” long-read sequencing technologies, such as SMRT and nanopore sequencing methods.

[0099] Single-molecule real-time (SMRT) sequencing facilitates direct detection of, e.g., 5- methylcytosine and 5-hydroxymethylcytosine as well as unmodified cytosine (Weirather JL, et al., “Comprehensive comparison of Pacific Biosciences and Oxford Nanopore Technologies and their applications to transcriptome analysis,” FlOOOResearch, 6: 100, 2017). Whereas next-generation sequencing methods detect augmented signals from a clonal population of amplified DNA fragments, SMRT sequencing captures a single DNA molecule, maintaining base modification during sequencing. The error rate of raw PacBio SMRT sequencing-generated data is about 13-15%, as the signal -to-noise ratio from single DNA molecules not high. To increase accuracy, this platform uses a circular DNA template by ligating hairpin adaptors to both ends of target double-stranded DNA. As the polymerase repeatedly traverses and replicates the circular molecule, the DNA template is sequenced multiple times to generate a continuous long read (CLR). The CLR can be split into multiple reads (“subreads”) by removing adapter sequences, and multiple subreads generate circular consensus sequence (“CCS”) reads with higher accuracy. The average length of a CLR is >10 kb and up to 60 kb, with length depending on the polymerase lifetime. Thus, the length and accuracy of CCS reads depends on the fragment sizes. PacBio sequencing has been utilized for genome (e.g., de novo assembly, detection of structural variants and haplotyping) and transcriptome (e.g., gene isoform reconstruction and novel gene / isoform discovery) studies.

[0100] ONT is a nanopore-based single molecule sequencing technology (Weirather JL, et al., FlOOOResearch, 6: 100, 2017). ONT directly sequences a native single-stranded DNA (ssDNA) molecule by measuring characteristic current changes as the bases are threaded through the nanopore by a molecular motor protein. ONT uses a hairpin library structure similar to the PacBio circular DNA template: the DNA template and its complement are bound by a hairpin adaptor. Therefore, the DNA template passes through the nanopore, followed by a hairpin and finally the complement. The raw read can be split into two “ID”Attorney Docket No. GH0240WO reads (“template” and “complement”) by removing the adaptor. The consensus sequence of two “ID” reads is a “2D” read with a higher accuracy.

[0101] 5 -letter and 6-letter sequencing methods include whole genome sequencing methods capable of sequencing A, C, T, and G in addition to 5mC and 5hmC to provide a 5-letter (A, C, T, G, and either 5mC or 5hmC) or 6-letter (A, C, T, G, 5mC, and 5hmC) digital readout in a single workflow. The processing of the DNA sample is entirely enzymatic and avoids the DNA degradation and genome coverage biases of bisulfite treatment. In an exemplary 5-letter sequencing method developed by Cambridge Epigenetix, the sample DNA is first fragmented via sonication and then ligated to short, synthetic DNA hairpin adaptors at both ends (Fullgrabe, et al. 2022, bioRxiv doi: https: / / doi.org / 10.1101 / 2022.07.08.499285). The construct is then split to separate the sense and antisense sample strands. For each original sample strand a complementary copy strand is synthesized by DNA polymerase extension of the 3 ’-end to generate a hairpin construct with the original sample DNA strand connected to its complementary strand, lacking epigenetic modifications, via a synthetic loop. Sequencing adapters are then ligated to the end. Modified cytosines are enzymatically protected. The unprotected Cs are then deaminated to uracil, which is subsequently read as thymine. In any such embodiments, amplification methods may comprise uracil- and / or dihydrouracil-tolerant amplification methods, such as PCR using a uracil- and / or dihydrouracil-tolerant DNA polymerase (i.e., a DNA polymerase that can read and amplify templates comprising uracil and / or dihydrouracil bases). The deaminated constructs are no longer fully complementary and have substantially reduced duplex stability, thus the hairpins can be readily opened and amplified by PCR. The constructs can be sequenced in paired-end format whereby read 1 (Pl primed) is the original stand and read 2 (P2 primed) is the copy stand. The read data is pairwise aligned so read 1 is aligned to its complementary read 2. Cognate residues from both reads are computationally resolved to produce a single genetic or epigenetic letter. Pairings of cognate bases that differ from the permissible five are the result of incomplete fidelity at some stage(s) comprising sample preparation, amplification, or erroneous base calling during sequencing. As these errors occur independently to cognate bases on each strand, substitutions result in a non-permissible pair. Non-permissible pairs are masked (marked as N) within the resolved read and the read itself is retained, leading to minimal information loss and high accuracy at read-level. The resolved read is aligned to the reference genome. Genetic variants and methylation counts are produced by read-counting at base-level.

[0102] 5hmC has been shown to have value as a marker of biological states and disease which includes early cancer detection from cell-free DNA. In adapting 5-letter to 6-letterAttorney Docket No. GH0240WO sequencing, 5mC is disambiguated from 5hmC without compromising genetic base calling within the same sample fragment. The first three steps of the workflow are identical to 5- letter sequencing described above, to generate the adapter ligated sample fragment with the synthetic copy strand. Methylation at 5mC is enzymatically copied across the CpG unit to the C on the copy strand, whilst 5hmC is enzymatically protected from such a copy. Thus, unmodified C, 5mC and 5hmC in each of the original CpG units are distinguished by unique 2-base combinations. The unmodified cytosines are then deaminated to uracil, which is subsequently read as thymine. The DNA is subjected to PCR amplification and sequencing as described earlier. The reads are pairwise aligned and resolved using a 2-base code. Each of unmodified C, 5mC, and 5hmC can be resolved as the three CpG units are distinct sequencing environments of the 2-base code.

[0103] In some embodiments, sequence coverage of the genome may be, for example, less than 5%, 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, 99%, 99.9% or 100%. In some embodiments, the sequence reactions may provide for sequence coverage of, for example, at least 5%, 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 70%, or 80% of the genome. Sequence coverage can be performed on, for example, at least 5, 10, 20, 70, 100, 200 or 500 different genes, or up to, for example, 5000, 2500, 1000, 500 or 100 different genes.

[0104] Simultaneous sequencing reactions may be performed using multiplex sequencing. In some embodiments, cell-free nucleic acids may be sequenced with at least, for example, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000, 50000, or 100,000 sequencing reactions. In other embodiments, cell-free nucleic acids may be sequenced with less than, for example, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000, 50000, or 100,000 sequencing reactions. Sequencing reactions may be performed sequentially or simultaneously. Subsequent data analysis may be performed on all or part of the sequencing reactions. In some embodiments, data analysis may be performed on at least, for example, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000, 50000, or 100,000 sequencing reactions. In other embodiments, data analysis may be performed on less than, for example, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000, 50000, or 100,000 sequencing reactions. An exemplary read depth is 1000-50000 or 1000-10000 or 1000-20000 reads per locus (base).

[0105] In general, sequencing of epigenetic target regions, e.g., to analyse a methylation profile of DNA, requires a lesser depth of sequencing than sequencing of a sequence-variableAttorney Docket No. GH0240WO target region, e.g., for analysis of mutations. Hence, lesser sequencing depths, as described herein, may in some cases be adequate for the methods described herein.Analyzing the sequence data

[0106] The sequencing data obtained by the methods of the present disclosure can be used to resolve unmethylated Cs, 5mC and 5hmC on a single molecule level.

[0107] The conversion procedures used in the methods of the present disclosure allow for 5mC to be distinguished from unmethylated Cs. For example, methods using a conversion procedure that selectively converts the base pairing specificity of 5mC means that 5mC in the sample nucleic acids will be read as T in sequencing. As noted elsewhere, 5hmC can be protected from conversion, e.g., through prior glucosylation. Aligning the sequence reads to a reference sequence (e.g., a reference genome) and identifying C>T alterations allows for the identification of 5mCs in the sample nucleic acids. Unmethylated Cs and protected 5hmCs are read as Cs in the sequencing data. The partitioning step allows 5hmCs and unmethylated Cs to be distinguished. Nucleic acids comprising a 5hmC will be partitioned into the subsample enriched for nucleic acids comprising 5hmC. Typically nucleic acids will contain at most one 5hmC nucleic acid base due to their scarcity in nature. The position of the 5hmC nucleic acid base in the nucleic acid can be identified using base coverage analysis. Base coverage analysis can involve aligning sequence reads (e.g., individual sequence reads, or consensus sequence reads for parent nucleic acids, as described elsewhere) from a subsample to a reference sequence. Analysis of the frequency (e.g., the proportion) of sequence reads which align to a specific C position in a reference sequence can identify the C position which comprised a 5hmC nucleic acid base in the sample nucleic acids. Specifically, the C position that comprised a 5hmC nucleic acid base in the sample nucleic acids would be expected to have a higher base coverage in the subsample enriched for nucleic acids comprising 5hmC nucleic acid bases compared to those C positions that did not comprise a 5hmC nucleic acid base in the sample nucleic acids.

[0108] Methods using a conversion procedure that selectively converts the base pairing specificity of unmethylated C (e.g., bisulfite sequencing) means that unmethylated C in the sample nucleic acids will be read as T in sequencing. Aligning the sequence reads to a reference sequence (e.g., a reference genome) and identifying OT alterations allows for the identification of unmethylated Cs in the sample nucleic acids. 5mCs and 5hmCs are read as Cs in the sequencing data. The partitioning step allows 5hmCs and 5mCs to be distinguished. Nucleic acids comprising a 5hmC will be partitioned into the subsample enriched for nucleicAttorney Docket No. GH0240WO acids comprising 5hmC. Typically nucleic acids will contain at most one 5hmC nucleic acid base due to their scarcity in nature. The position of the 5hmC nucleic acid base in the nucleic acid can be identified using base coverage analysis, as described above.

[0109] As noted above, identifying nucleic acid bases that have undergone conversion generally involves comparing the sequence data obtained from the nucleic acids that has been subjected to the conversion procedure to a reference sequence (e.g., a reference genome). Typically, the method involves (i) comparing the sequence data with (A) one or more predetermined reference sequence, such as reference sequences corresponding to one or more epigenetic target regions where particular significance is attached to the methylation profile, e.g., in diagnosing, prognosing or characterizing a cancer; or (B) sequence data obtained by sequencing a subsample of the nucleic acid that was not subjected to the conversion procedure, for example a subsample that was separated before subjecting a separate subsample to the conversion procedure; and (ii) identifying point differences between the converted nucleic acid sequences and the reference sequence(s) (A) or non-converted nucleic acid sequences (B) as nucleosides (in the initial sample) having a methylation status that permits a change in base pairing specificity on exposure to the conversion procedure.

[0110] The identification of the methylation status of the sample nucleic acids has a variety of utilities. For example, methylation status can be used to characterize disease states, including for example, identifying the presence or absence of cancer, identification of cancer type, and / or identifying the tissue of origin of cfDNA molecules.

[0111] Analyzing the sequence data may also include the analysis of non-methylation features, such as fragmentation patterns (e.g., in the case of cfDNA analysis) or genetic variants (such as SNVs, indels and / or CNVs). When analyzing fragmentation patterns and / or genetic variants, conversion procedure that selectively converts the base pairing specificity of 5mC are preferred. These conversion procedures are generally less destructive and thus maintain the fragmentation pattern of the sample nucleic acids. Moreover, they do not convert the base pairing specificity of unmethylated C, thus allowing for more sensitive mutation detection.

[0112] Fragmentation patterns of DNA molecules in cfDNA samples carry information about the chromatin organization of the cells or tissues from which the cfDNA fragments originate. In particular, DNA fragments released to the bloodstream is often fragmented or cleaved around nucleosomes and / or other DNA bound proteins in the cells or tissues of origin.Further, nucleosome positioning and the location of DNA binding proteins is highly tissue specific and thus is used herein to amplify signal coming from the cells or tissues from whichAttorney Docket No. GH0240WO the cfDNA fragments originate (e.g., tumor cells as well as cells in the tumor microenvironment and cells involved in the immune response). Accordingly, in some embodiments, analyzing the sequencing data may comprise analyzing the methylation profile and the fragmentation pattern of cfDNA. Such analysis can be used to identify the tissue of origin of the cfDNA and / or diagnose or prognose cancer. In some embodiments, analyzing the sequencing data may comprise analyzing the methylation profile and the presence or absence of genetic variants in cfDNA. Such analysis can be used to identify the tissue of origin of the cfDNA and / or diagnose or prognose cancer.

[0113] In some embodiments, analyzing the sequencing data may comprise analyzing: (i) the methylation profile; (ii) the fragmentation pattern; and (iii) the presence or absence of genetic variants in cfDNA. Such analysis can be used to identify the tissue of origin of the cfDNA and / or diagnose or prognose cancer.Modification sensitive sequencing

[0114] In some embodiments, the nucleic acids are linked at both ends to Y-shaped adapters including primer binding sites and tags. The molecules are amplified

[0115] The methods disclosed herein may use modification sensitive sequencing to detect the modification status of one or more nucleotides. This may include nucleotides present in the original sample and / or at least one type of dNTP comprising a modified base (such as mCTP) used in the end repair reaction. In some embodiments, a DNA sample comprising a plurality of DNA molecules is subjected to modification-sensitive sequencing to obtain sequencing data derived from the DNA sample, wherein the modification-sensitive sequencing comprises subjecting the plurality of DNA molecules to a procedure that affects a first nucleobase of the plurality of DNA molecules differently from a second nucleobase of the plurality of DNA molecules, wherein the first nucleobase is a modified or unmodified nucleobase, the second nucleobase is a modified or unmodified nucleobase different from the first nucleobase, and the first nucleobase and the second nucleobase have the same base pairing specificity, thereby producing a plurality of converted DNA molecules comprising one or more inappropriately converted bases and / or one or more inappropriately unconverted bases at one or more locations. Such embodiments may also comprise a step of end-repair prior to the modification-sensitive sequencing.

[0116] Modification sensitive sequencing involves a sequencing workflow which is capable of distinguishing at least two modification states of a nucleotide bases. These two states may be: (i) whether a base is modified or not (e.g. 5mC and / or 5hmC vs unmethylated cytosine);Attorney Docket No. GH0240WO or (ii) the type of modification which a base exhibits (e.g. 5mC vs 5hmC). Modification sensitive sequencing does not necessarily require that a specific type of modification is identified as present or absent at a specific position, just whether one or more modification types (e.g. 5mC and 5hmC) is present or absent. For instance, in some embodiments, modification sensitive sequencing includes sequencing comprising a bisulfite conversion step which can distinguish 5mC and 5hmC from unmethylated C, but it cannot distinguish between 5mC and 5hmC. In some embodiments, the modification-sensitive sequencing comprises subjecting a DNA sample (such as a DNA sample from a subject) to a procedure that affects a first nucleobase of the DNA differently from a second nucleobase of the DNA, wherein the first nucleobase is a modified or unmodified nucleobase, the second nucleobase is a modified or unmodified nucleobase different from the first nucleobase, and the first nucleobase and the second nucleobase have the same base pairing specificity.

[0117] The type of modification sensitive sequencing used will depend on the type of modified base(s) used in the end repair, such that the type of modification sensitive sequencing will be able to detect at least the presence or absence of at least that modified base.

[0118] As outlined below, there are various methods of detecting and / or identifying modified nucleosides that rely on a conversion procedure that changes the base-pairing specificity of a nucleoside, based on the modification status of the nucleosides. These changes of basepairing specificity can then be detected, and thus the modification status of the nucleoside inferred, by sequencing. Together, the conversion procedure and the sequencing itself constitutes one form of modification aware sequencing, as referred to herein.

[0119] In some cases, a conversion procedure used in the methods of the disclosure is one that changes the base pairing specificity of a modified nucleoside (e.g. methylated cytosine), but does not change the base pairing specificity of the corresponding unmodified nucleoside (e.g. cytosine) or does not change the base pairing specificity of any un-modified nucleoside (e.g. cytosine, adenosine, guanosine and thymidine (or uracil)). Advantages of methods that do not convert the base-pairing specificity of unmodified nucleosides include reduced loss of sequence complexity, higher sequencing efficiency and reduced alignment losses. Additionally, methods such as TAPS may in some cases be preferred over methods such as bisulfite sequencing and EM-seq because they are less destructive (especially important for low yield samples such as cfDNA or FFPE samples) and do not require denaturation, meaning that non-conversion errors are theoretically more likely to be random. In methods that require denaturation for conversion, failure to denature a DNA molecule will result inAttorney Docket No. GH0240WO non-conversion of all bases in the DNA molecule. As biological changes in methylation are predominantly concerted to a localized regions of interest, these non-random (localized) nonconversion events can appear as false negatives (non-methylated regions). Random nonconversion methods can maximally affect a low percent of bases within a region, and thus the specificity of methylation change detection can be maximized (reduce false positives) by placing a threshold on percentage of bases within a region that are methylated / non- methylated. Hence, in some cases, a conversion procedure that does not involve denaturation is preferred. In other cases, a conversion procedure used in the methods of the disclosure is one that changes the base pairing specificity of an unmodified nucleoside (e.g. cytosine), but does not change the base pairing specificity of the corresponding modified nucleoside (e.g. methylated cytosine such as 5hmC and / or 5mC). Such methods include, for example, bisulfite sequencing. The skilled person can select a suitable method according to their needs, including which nucleoside modifications are to be detected and / or identified and which type of modified base is used in the end repair reaction.

[0120] In some embodiments, the conversion procedure converts modified nucleosides. In some embodiments, the conversion procedure which converts modified nucleosides comprises Tet-assisted conversion with a substituted borane reducing agent, optionally wherein the substituted borane reducing agent is 2-picoline borane, borane pyridine, tertbutylamine borane, ammonia borane or pyridine borane. In Tet-assisted pic-borane conversion with a substituted borane reducing agent conversion, a TET protein is used to convert 5mC and 5hmC to 5caC, without affecting unmodified C. 5caC, and 5fC if present, are then converted to dihydrouracil (DHU) by treatment with 2-picoline borane (pic-borane) or another substituted borane reducing agent such as borane pyridine, tert-butylamine borane, or ammonia borane, also without affecting unmodified C. See, e.g., Liu et al., Nature Biotechnology 2019; 37:424-429 (e.g., at Supplementary Fig. 1 and Supplementary Note 7). Thus, when this type of conversion is used, the first nucleobase comprises one or more of 5mC, 5fC, 5caC, or 5hmC, and the second nucleobase comprises unmodified cytosine. DHU is read as a T in sequencing. Sequencing of the converted DNA identifies positions that are read as cytosine as being unmodified C positions. Meanwhile, positions that are read as T are identified as being T, 5mC, 5fC, 5caC, or 5hmC. Performing TAP conversion, such as on a DNA sample as described herein, thus facilitates identifying positions containing unmodified C using the sequence reads obtained.

[0121] Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein the at least one type of dNTP comprises a 5mC or 5hmC, and regions synthesizedAttorney Docket No. GH0240WO during the end repair reaction can be identified as those regions comprising 5mC or 5hmC (via T being called at positions which are C in the reference) at non-CpG positions. This procedure encompasses Tet-assisted pyridine borane sequencing (TAPS), described in further detail in Liu et al. 2019, supra. In this method Tet enzyme is used to progressively oxidize 5mC and 5hmC to 5fC or 5caC, then pyridine borane deaminates 5fC, 5CaC to DHU, amplified as T.

[0122] Alternatively, protection of 5hmC (e.g., using PGT or 5-hydroxymethylcytosine carbamoyltransferase) can be combined with Tet-assisted conversion with a substituted borane reducing agent, e.g. as described above. In this method (TAPS-P), 5hmC can be protected from conversion, for example through glucosylation using P-glucosyl transferase (PGT), forming (forming 5-glucosylhydroxymethylcytosine) 5ghmC, or through carbamoylation using 5-hydroxymethylcytosine carbamoyltransferase, forming 5cmC. This is described in Yu et al., Cell 2012; 149: 1368-80. Treatment with a TET protein such as mTetl then converts 5mC to 5caC but does not convert C, 5ghmC, or 5cmC. 5caC is then converted to DHU by treatment with pic-borane or another substituted borane reducing agent such as borane pyridine, tert-butylamine borane, or ammonia borane, also without affecting ghmC, 5cmC, or unmodified C. Thus, when Tet-assisted conversion with a substituted borane reducing agent is used, the first nucleobase comprises mC, and the second nucleobase comprises one or more of unmodified cytosine or hmC, such as unmodified cytosine and optionally hmC, fC, and / or caC. Sequencing of the converted DNA identifies positions that are read as cytosine as being either 5hmC or unmodified C positions. Meanwhile, positions that are read as T are identified as being T, 5fC, 5caC, or 5mC. Performing TAPSP conversion on a sample as described herein thus facilitates distinguishing positions containing unmodified C or 5hmC on the one hand from positions containing 5mC using the sequence reads obtained. Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein the at least one type of dNTP comprises a 5mC, and regions synthesized during the end repair reaction can be identified as those regions comprising 5mC (via T being called at positions which are C in the reference) at non-CpG positions. For an exemplary description of this type of conversion, see, e.g., Liu et al., Nature Biotechnology 2019; 37:424-429. 5-hydroxymethylcytosine carbamoyltransferase is described in Yang et al., Bio-protocol, 2023; 12(17): e4496.

[0123] In some embodiments, the conversion procedure converts modified nucleosides. In some embodiments, the conversion procedure which converts modified nucleosides comprises chemi cal -assisted conversion with a substituted borane reducing agent, optionallyAttorney Docket No. GH0240WO wherein the substituted borane reducing agent is 2-picoline borane, borane pyridine, tertbutylamine borane, borane pyridine or ammonia borane. In chemi cal -assisted conversion with a substituted borane reducing agent, an oxidizing agent such as potassium perruthenate (KRuO4) (also suitable for use in ox-BS conversion) is used to specifically oxidize 5hmC to 5fC. Treatment with pic-borane or another substituted borane reducing agent such as borane pyridine, tert-butylamine borane, or ammonia borane converts 5fC and 5caC to DHU but does not affect 5mC or unmodified C. Thus, when this type of conversion is used, the first nucleobase comprises one or more of hmC, fC, and caC, and the second nucleobase comprises one or more of unmodified cytosine or mC, such as unmodified cytosine and optionally mC. Sequencing of the converted DNA identifies positions that are read as cytosine as being either 5mC or unmodified C positions. Meanwhile, positions that are read as T are identified as being T, 5fC, 5caC, or 5hmC. Performing this type of conversion as described herein thus facilitates distinguishing positions containing unmodified C or 5mC on the one hand from positions containing 5hmC using the sequence reads obtained. Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein at least one type of dNTP comprises a 5hmC, and regions synthesized during the end repair reaction can be identified as those regions comprising 5hmC (via T being called at positions which are C in the reference) at non-CpG positions. For an exemplary description of this type of conversion, see, e.g., Liu et al., Nature Biotechnology 2019; 37:424-429.

[0124] Exemplary conversion procedures that change the base-pairing specificity of modified cytosines have been described. However, the methods described herein could in principle use any modified nucleoside and suitable conversion procedure (i.e. single-base epigenetic conversion assay) that changes the base-pairing specificity of the modified nucleoside and thereby allows the modified base to be distinguished from the corresponding unmodified nucleoside and / or other types of modification when sequenced. For example, any conversion procedure could be used allowing any one of N6-methyladenine (6mA), N6- hydroxymethyladenine (6hmA), or N6-formyladenine (6fA) to be distinguished from unmodified adenosine.

[0125] In some embodiments, the conversion procedure converts unmodified nucleosides. In some embodiments, the conversion procedure which converts unmodified nucleosides comprises bisulfite conversion. Treatment with bisulfite converts unmodified cytosine and certain modified cytosine nucleotides (e.g. 5-formyl cytosine (5fC) or 5-carboxylcytosine (5caC)) to uracil whereas other modified cytosines (e.g., 5mC and 5hmC) are not converted. Thus, where bisulfite conversion is used, the first nucleobase comprises one or more ofAttorney Docket No. GH0240WO unmodified cytosine, 5fC, 5caC, or other cytosine forms affected by bisulfite, and the second nucleobase may comprise one or more of 5mC and 5hmC, such as 5mC and optionally 5hmC. Sequencing of bisulfite-treated DNA identifies positions that are read as cytosine as being 5mC or 5hmC positions. Meanwhile, positions that are read as T are identified as being T or a bisulfite-susceptible form of C, such as unmodified cytosine, 5fC, or 5caC. Thus, performing bisulfite conversion, such as on a DNA sample as described herein facilitates identifying positions containing 5mC or 5hmC. Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein at least one type of dNTP comprises a 5mC and / or a 5hmC, and regions synthesized during the end repair reaction can be identified as those regions comprising 5mC or a 5hmC (via C being called at these positions) at non-CpG positions. For an exemplary description of bisulfite conversion, see, e.g., Moss et al., Nat Commun. 2018; 9: 5068.

[0126] In some embodiments, the procedure which converts unmodified nucleosides comprises oxidative bisulfite (Ox-BS) conversion. This procedure first converts 5hmC to 5fC, which is bisulfite susceptible, followed by bisulfite conversion. Thus, when oxidative bisulfite conversion is used, the first nucleobase comprises one or more of unmodified cytosine, 5fC, 5caC, 5hmC, or other cytosine forms affected by bisulfite, and the second nucleobase comprises 5mC. Sequencing of Ox-BS converted DNA identifies positions that are read as cytosine as being 5mC positions. Meanwhile, positions that are read as T are identified as being T or a bisulfite-susceptible form of C, such as unmodified cytosine, 5fC, or 5hmC. Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein at least one type of dNTP comprises a 5mC, and regions synthesized during the end repair reaction can be identified as those regions comprising 5mC (via C being called at these positions) at non-CpG positions. Performing Ox-BS conversion thus facilitates identifying positions containing mC. For an exemplary description of oxidative bisulfite conversion, see, e.g., Booth et al., Science 2012; 336: 934-937.

[0127] In some embodiments, the procedure which converts unmodified nucleosides comprises Tet-assisted bisulfite (TAB) conversion. In TAB conversion, 5hmC is protected from conversion and 5mC is oxidized in advance of bisulfite treatment, so that positions originally occupied by 5mC are converted to U while positions originally occupied by 5hmC remain as a protected form of cytosine. For example, as described in Yu et al., Cell 2012; 149: 1368-80, P-glucosyl transferase can be used to protect 5hmC (forming 5- glucosylhydroxymethylcytosine (5ghmC)), then a TET protein such as mTetl can be used toAttorney Docket No. GH0240WO convert 5mC to 5caC, and then bisulfite treatment can be used to convert C and 5caC to U while 5ghmC remains unaffected.

[0128] Alternatively, a carbamoyltransferase enzyme, such as 5-hydroxymethylcytosine carbamoyltransferase as described in Yang et al., Bio-protocol, 2023; 12(17): e4496, can be used to protect hmC (by converting hmC to 5-carbamoyloxymethylcytosine (5cmC)), then a TET protein such as mTetl can be used to convert mC to caC, and then bisulfite treatment can be used to convert C and caC to U while 5cmC remains unaffected. Thus, when TAB conversion is used, the first nucleobase comprises one or more of unmodified cytosine, 5fC, 5caC, 5mC, or other cytosine forms affected by bisulfite, and the second nucleobase comprises 5hmC. Sequencing of TAB-converted DNA identifies positions that are read as cytosine as being 5hmC positions. Meanwhile, positions that are read as T are identified as being T, or a bisulfite-susceptible form of C, such as unmodified cytosine, 5mC, 5fC, or 5caC. Performing TAB conversion on a first subsample as described herein thus facilitates identifying positions containing 5hmC. Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein at least one type of dNTP comprises a 5hmC, and regions synthesized during the end repair reaction can be identified as those regions comprising 5hmC (via C being called at these positions) at non-CpG positions.

[0129] In some embodiments, the conversion procedure which converts unmodified nucleosides comprises APOBEC-coupled epigenetic (ACE) conversion. In ACE conversion, an AID / APOBEC family DNA deaminase enzyme such as APOBEC3 A (A3 A) is used to deaminate unmodified cytosine and 5mC without deaminating 5hmC, 5fC, or 5caC. Thus, when ACE conversion is used, the first nucleobase comprises unmodified C and / or mC (e.g., unmodified C and optionally mC), and the second nucleobase comprises hmC. Sequencing of ACE-converted DNA identifies positions that are read as cytosine as being 5hmC, 5fC, or 5caC positions. Meanwhile, positions that are read as T are identified as being T, unmodified C, or 5mC. Performing ACE conversion as described herein thus facilitates distinguishing positions containing 5hmC from positions containing 5mC or unmodified C using the sequence reads obtained from the first subsample. Hence, in these embodiments, the end repair reaction can be performed with dNTPs, wherein at least one type of dNTP comprises a 5hmC, and regions synthesized during the end repair reaction can be identified as those regions comprising 5hmC (via C being called at these positions) at non-CpG positions. For an exemplary description of ACE conversion, see, e.g., Schutsky et al., Nature Biotechnology 2018; 36: 1083-1090.Attorney Docket No. GH0240WO

[0130] In some embodiments, the procedure that affects a first nucleobase in the DNA differently from a second nucleobase in the DNA of the first subsample comprises enzymatic conversion of the first nucleobase, e.g., as in EM-Seq. See, e.g., Vaisvila R, et al. (2019) EM- seq: Detection of DNA methylation at single base resolution from picograms of DNA. bioRxiv; DOI: 10.1101 / 2019.12.20.884692, available at www.biorxiv.org / content / 10.1101 / 2019.12.20.884692vl. For example, TET2 and T4-PGT or 5-hydroxymethylcytosine carbamoyltransferase (described in Yang et al., Bio-protocol, 2023; 12(17): e4496) can be used to convert 5mC and 5hmC into substrates that cannot be deaminated by a deaminase (e.g., APOBEC3A), and then a deaminase (e.g., APOBEC3A) can be used to deaminate unmodified cytosines converting them to uracils.

[0131] In some embodiments, the procedure that affects a first nucleobase in the DNA differently from a second nucleobase in the DNA comprises enzymatic conversion of the first nucleobase using a non-specific, modification-sensitive double-stranded DNA deaminase, e.g., as in SEM-seq. See, e.g., Vaisvila et al. (2023) Discovery of novel DNA cytosine deaminase activities enables a nondestructive single-enzyme methylation sequencing method for base resolution high-coverage methylome mapping of cell-free and ultra-low input DNA. bioRxiv; DOI: 10.1101 / 2023.06.29.547047, available at https: / / www.biorxiv.org / content / 10.1101 / 2023.06.29.547047vl. SEM-Seq employs a nonspecific, modification-sensitive double-stranded DNA deaminase (MsddA) in a nondestructive single-enzyme 5-methylctyosine sequencing (SEM-seq) method that deaminates unmodified cytosines. Accordingly, SEM-seq does not require the TET2 and T4- PGT or 5-hydroxymethylcytosine carbamoyltransferase protection and denaturing steps that are of use, e.g., in APOEC3A-based protocols. Additionally, MsddA does not deaminate 5- formylated cytosines (5fC) or 5-carboxylated cytosines (5caC). In SEM-seq, unmodified cytosines in the DNA are deaminated to uracil and is read as “T” during sequencing. Modified cytosines (e.g., 5mC) are not converted and are read as “C” during sequencing. Cytosines that are read as thymines are identified as unmodified (e.g., unmethylated) cytosines or as thymines in the DNA. Performing SEM-seq conversion thus facilitates identifying positions containing 5mC using the sequence reads obtained. In some embodiments, the procedure that affects a first nucleobase in the DNA differently from a second nucleobase in the DNA comprises enzymatic conversion of the first nucleobase using MsddA.

[0132] In some embodiments, the conversion procedure converts modified nucleosides. In some embodiments, the conversion procedure which converts modified nucleosidesAttorney Docket No. GH0240WO comprises enzymatic conversion, such as DM-seq, for example, as described in WO2023 / 288222A1. In DM-seq, unmodified cytosines in the DNA are enzymatically protected from a subsequent deamination step wherein 5mC in 5mCpG is converted to T. The enzymatically protected unmodified (e.g., unmethylated) cytosines are not converted and are read as “C” during sequencing. Cytosines that are read as thymines (in a CpG context) are identified as methylated cytosines in the DNA.

[0133] Thus, when this type of conversion is used, the first nucleobase comprises unmodified (such as unmethylated) cytosine, and the second nucleobase comprises modified (such as methylated) cytosine. Sequencing of the converted DNA identifies positions that are read as cytosine as being unmodified C positions. Meanwhile, positions that are read as T are identified as being T or 5mC. Performing DM-seq conversion thus facilitates identifying positions containing 5mC using the sequence reads obtained.

[0134] Exemplary cytosine deaminases for use herein include APOBEC enzymes, for example, APOBEC3 A. Generally, AID / APOBEC family DNA deaminase enzymes such as APOBEC3A (A3 A) are used to deaminate (unprotected) unmodified cytosine and 5mC. For an exemplary description of APOBEC conversion, see, e.g., Schutsky et al., Nature Biotechnology 2018; 36: 1083-1090.

[0135] The enzymatic protection of unmodified cytosines in the DNA comprises addition of a protective group to the unmodified cytosines. Such protective groups can comprise an alkyl group, an alkyne group, a carboxyl group, a carboxyalkyl group, an amino group, a hydroxymethyl group, a glucosyl group, a glucosylhydroxymethyl group, an isopropyl group, or a dye. For example, DNA can be treated with a methyltransferase, such as a CpG-specific methyltransferase, which adds the protective group to unmodified cytosines. The term methyltransferase is used broadly herein to refer to enzymes capable of transferring a methyl or substituted methyl (e.g., carboxymethyl) to a substrate (e.g., a cytosine in a nucleic acid). In some embodiments, the DNA is contacted with a CpG-specific DNA methyltransferase (MTase), such as a CpG-specific carboxymethyltransferase (CxMTase), and a substituted methyl donor, such as a carboxymethyl donor (e.g., carboxymethyl-S-adenosyl-L- methionine). See, e.g., WO2021 / 236778A2. In particular embodiments, the CxMTase can facilitate the addition of a protective carboxymethyl group to an unmethylated cytosine. In some embodiments, the unmethylated cytosine is unmodified cytosine. The carboxymethyl group can prevent deamination of the cytosine during a deamination step (such as a deamination step using an APOBEC enzyme, such as A3 A). Substituted methyl or carboxymethyl donors useful in the disclosed methods include but are not limited to, S-Attorney Docket No. GH0240WO adenosyl-L-methionine (SAM) analogs, optionally wherein the SAM analog is carboxy-S- adenosyl-L-methionine (CxSAM). SAM analogs are described, for example, in WO2022 / 197593A1. The MTase may be, for example, a CpG methyltransferase from Spiroplasma sp. strain MQ1 (M.SssI), DNA-methyltransferase 1 (DNMT1), DNA- methyltransferase 3 alpha (DNMT3 A), DNA-methyltransferase 3 beta (DNMT3B), or DNA adenine methyltransferase (Dam). The CxMTase may be a CpG methyltransferase from Mycoplasma penetrans (M.Mpel). In a particular embodiment, the methyltransferase enzyme is a variant of M.Mpel, or a sequence at least 90%, at least 92%, at least 94%, at least 96%, at least 97%, at least 98%, or at least 99% identical thereto, optionally wherein the amino acid corresponding to position 374 is R or K.

[0136] In one embodiment, the methyltransferase enzyme is a variant of M.Mpel having an N374R substitution or an N374K substitution. The methyltransferase can further comprise one or more amino acid substitutions selected from a) substitution of one or both residues T300 and E305 with S, A, G, Q, D, or N; b) substitution of one or more residues A323, N306, and Y299 with a positively charged amino acid selected from K, R or H; and / or c) substitution of S323 with A, G, K, R or H, which may enhance the activity of the enzyme.

[0137] Optionally, the conversion procedure further includes enzymatic protection of 5hmCs, such as by glucosylation of the 5hmCs (e.g., using PGT) or by carbamoylation of the 5hmCs (e.g., using 5-hydroxymethylcytosine carbamoyltransferase), in the DNA prior to the deamination of unprotected modified cytosines. In this method, 5hmC can be protected from conversion, for example through glucosylation using P-glucosyl transferase (PGT), forming (5-glucosylhydroxymethylcytosine) 5ghmC, or through carbamoylation using 5- hydroxymethylcytosine carbamoyltransferase, forming 5cmC. This is described, for example, in Yu et al., Cell 2012; 149: 1368-80, and in Yang et al., Bio-protocol, 2023; 12(17): e4496. Glucosylation or carbamoylation of 5hmC can reduce or eliminate deamination of 5hmC by a deaminase such as APOBEC3A. Treatment with an MTase or CxMTase then adds a protecting group to unmodified (unmethylated) cytosines in the DNA. 5mC (but not protected, unmodified cytosine and not 5ghmC or 5cmC) is then deaminated (converted to T in the case of 5mC) by treatment with a deaminase, for example, an APOB EC enzyme (such as APOBEC3 A). Sequencing of the converted DNA identifies positions that are read as cytosine as being either 5hmC or unmodified C positions. Meanwhile, positions that are read as T are identified as being T or 5mC. Performing DM-seq conversion with glucosylation of 5hmC on a sample as described herein thus facilitates distinguishing positions containingAttorney Docket No. GH0240WO unmodified C or 5hmC on the one hand from positions containing 5mC using the sequence reads obtained.

[0138] Also provided herein are methods in which alternative base conversion schemes are used. For example, unmethylated cytosines can be left intact while methylated cytosines and hydroxymethylcytosines are converted to a base read as a thymine (e.g., uracil, thymine, or dihydrouracil).

[0139] In some embodiments, methylating a cytosine in at least one first complementary strand or second complementary strand comprises contacting the cytosine with a methyltransferase such as DNMT1 or DNMT5. In such embodiments, the step of oxidizing a 5-hydroxymethylated cytosine to 5-formylcytosine (such as by contacting the 5- hydroxymethyl cytosine in a first strand and a second strand with KRuO4) can be optional.

[0140] In some embodiments, converting the modified cytosine in at least one first or second strand to a thymine or a base read as thymine comprises oxidizing a hydroxymethyl cytosine, e.g., the hydroxymethyl cytosine is oxidized to formylcytosine. In some embodiments, oxidizing the hydroxymethyl cytosine to formylcytosine comprises contacting the hydroxymethyl cytosine with a ruthenate, such as potassium ruthenate (KRuO4).

[0141] In some embodiments, the modified cytosine is converted to thymine, uracil, or dihydrouracil. In any such embodiments, amplification methods may comprise uracil- and / or dihydrouracil-tolerant amplification methods, such as PCR using a uracil- and / or dihydrouracil-tolerant DNA polymerase.

[0142] In some embodiments, the method comprises converting a formylcytosine and / or a methylcytosine to carboxylcytosine as part of converting the modified cytosine in at least one first or second strand to a thymine or a base read as thymine. For example, converting the formylcytosine and / or the methylcytosine to carboxylcytosine can comprise contacting the formylcytosine and / or the methylcytosine with a TET enzyme, such as TET1, TET2, or TET3. In some embodiments, the method comprises reducing the carboxyl cytosine as part of converting the modified cytosine in at least one first or second strand to a thymine or a base read as thymine, and / or the carboxylcytosine is reduced to dihydrouracil. In some embodiments, reducing the carboxylcytosine comprises contacting the carboxylcytosine with a borane or borohydride reducing agent.

[0143] In some embodiments, the borane or borohydride reducing agent comprises pyridine borane, 2-picoline borane, borane, tert-butylamine borane, ammonia borane, sodium borohydride, sodium cyanoborohydride (NaBH3CN), lithium borohydride (LiBH4), ethylenediamine borane, dimethylamine borane, sodium triacetoxyborohydride, morpholineAttorney Docket No. GH0240WO borane, 4-methylmorpholine borane, trimethylamine borane, dicyclohexylamine borane, or a salt thereof. In other embodiments, the reducing agent comprises lithium aluminum hydride, sodium amalgam, amalgam, sulfur dioxide, dithionate, thiosulfate, iodide, hydrogen peroxide, hydrazine, diisobutylaluminum hydride, oxalic acid, carbon monoxide, cyanide, ascorbic acid, formic acid, dithiothreitol, beta-mercaptoethanol, or any combination thereof. Various TET enzymes may be used in the disclosed methods as appropriate, as described elsewhere herein. Modification sensitive sequencing also includes sequencing methods which do not rely on a conversion step, wherein the base pairing specificity of a base is changed dependent on its modification status. For instance, single molecule techniques such as nanopore based sequencing and single molecule real time sequencing can be used to directly detect modified bases.

[0144] For example, some sequencing reactions involve use of an enzyme to control passage of a nucleic acid through a nanopore, and in such cases reaction data can include both kinetics and other behavior of the enzyme and fluctuations in current through the nanopore. For example, ratchet proteins, helicases, or motor proteins can be used to push or pull a nucleic acid molecule through a hole in a biological or synthetic membrane. The kinetics of these proteins can vary depending on the sequence context of a nucleic acid on which they are acting. For example, they may slow down or pause at a modified base, and this behavior, captured as a part of the reaction data, is indicative of the presence of the modified base even where the modified base is not within the sensing portion of the nanopore. One example of a nanopore sequencing system is that commercialized by Oxford Nanopore Technologies (ONT). (See e.g., (Weirather et al., FlOOOResearch, 6: 100, 2017.) ONT sequencing directly sequences a native single-stranded DNA (ssDNA) molecule by measuring characteristic current changes as the bases are threaded through the nanopore by a molecular motor protein. ONT sequencing uses a hairpin library structure similar to the PacBio circular DNA template: the DNA template and its complement are bound by a hairpin adaptor. Therefore, the DNA template passes through the nanopore, followed by a hairpin and finally the complement. The raw read can be split into two “ID” reads (“template” and “complement”) by removing the adaptor. The consensus sequence of two “ID” reads is a “2D” read with a higher accuracy.

[0145] Nanopore sequencing can be used to detect base modifications including 5mC, 5hmC, 6mA, BrdU, FdU, IdU, and EdU (see e.g., Gouil & Keniry Essays in Biochemistry (2019) 63 639-648; Kutyavin, Biochemistry (2008), 47, 51, 13666-1367; Muller et al., Nature Methods (2019), volume 16, pages 429-436; Hennion et al., Genome Biology (2020), volume 21, Article number: 125). Accordingly, in some embodiments, the modification sensitiveAttorney Docket No. GH0240WO sequencing comprises nanopore sequencing. In such embodiments, the end repair may be performed using dNTPs, which comprise 4mC, 5mC, 5hmC, 6mA, BrdU, FdU, IdU, and / or EdU.

[0146] Another modification sensitive single molecule sequencing technique is single molecule real time sequencing (SMRT) that has been commercialized by Pacific Biosciences. SMRT sequencing relies on sequencing-by-synthesis, where the sequence of a circular DNA template is determined from the succession of fluorescence pulses, each resulting from the addition of one labelled nucleotide by a polymerase fixed to the bottom of a well. Base modifications do not affect the base-called sequence, but they affect the kinetics of the polymerase. By considering the inter-pulse duration (IPD), base modifications can be inferred from the comparison of a modified template to an in silico model or an unmodified template. Such methods can therefore use the pulse width of a signal from sequencing bases, the interpulse duration (IPD) of bases, and the identity of the bases in order to detect a modification in a base or in a neighboring base. (See e.g., Weirather et al., FlOOOResearch, 6: 100, 2017.) Single molecule real time sequencing can be used to detect base modifications such as 4mC, 5mC, 5hmC, 6mA, and 8oxoG (Gouil & Keniry Essays in Biochemistry (2019) 63 639-648). Accordingly, in some embodiments, the modification sensitive sequencing comprises single molecule real time sequencing. In such embodiments, the end repair may be performed using dNTPs, which comprise 4mC, 5mC, 5hmC, 6mA, and / or 8oxoG.Partitioning

[0147] In some instances, a heterogeneous nucleic acid sample is partitioned into two or more partitions (sub-samples). In some embodiments, each partition is differentially tagged. Tagged partitions can then be pooled together for collective sample prep and / or sequencing. The partitioning-tagging-pooling steps can occur more than once, with each round of partitioning occurring based on a different characteristics, and tagged using differential tags that are distinguished from other partitions and partitioning means.

[0148] Examples of characteristics that can be used for partitioning include sequence length, methylation level, nucleosome binding, sequence mismatch, immunoprecipitation, and / or proteins that bind to DNA. Resulting partitions can include one or more of the following nucleic acid forms: single-stranded DNA (ssDNA), double-stranded DNA (dsDNA), shorter DNA fragments and longer DNA fragments. In some embodiments, partitioning based on a cytosine modification (e.g., cytosine methylation) or methylation generally is performed and is optionally combined with at least one additional partitioning step, which may be based onAttorney Docket No. GH0240WO any of the foregoing characteristics or forms of DNA. In some embodiments, a heterogeneous population of nucleic acids is partitioned into nucleic acids with one or more base modifications and without the one or more base modifications. Examples of base modifications are described elsewhere herein. Alternatively or additionally, a heterogeneous population of nucleic acids can be partitioned into nucleic acid molecules associated with nucleosomes and nucleic acid molecules devoid of nucleosomes. Alternatively or additionally, a heterogeneous population of nucleic acids may be partitioned into singlestranded DNA (ssDNA) and double-stranded DNA (dsDNA). Alternatively, or additionally, a heterogeneous population of nucleic acids may be partitioned based on nucleic acid length (e.g., molecules of up to 160 bp and molecules having a length of greater than 160 bp).

[0149] In some cases, different procedures are applied to different partitions to determine different characteristics of the initial sample. The DNA of at least one partition is subjected to an end repair and modification sensitive sequencing procedure according to the methods of the disclosure described herein. In some embodiments at least one partition is not subjected to the end repair and modification sensitive sequencing procedure according to the methods of the disclosure described herein. In cases where the modification sensitive sequencing procedure comprises a conversion procedure, corresponding sequences from the converted and non-converted partitions can be compared to identify single nucleotides that have undergone conversion and therefore identify corresponding modified nucleosides in the initial sample.

[0150] In some embodiments, partition tagging comprises tagging molecules in each partition with a partition tag. After re-combining partitions (e.g., to reduce the number of sequencing runs needed and avoid unnecessary cost) and sequencing molecules, the partition tags identify the source partition. In another embodiment, different partitions are tagged with different sets of molecular tags, e.g., comprised of a pair of barcodes. In this way, each molecular barcode indicates the source partition as well as being useful to distinguish molecules within a partition. For example, a first set of 35 barcodes can be used to tag molecules in a first partition, while a second set of 35 barcodes can be used tag molecules in a second partition.

[0151] In some embodiments, after partitioning and tagging with partition tags, the molecules may be pooled for sequencing in a single run. In some embodiments, a sample tag is added to the molecules, e.g., in a step subsequent to addition of partition tags and pooling. Sample tags can facilitate pooling material generated from multiple samples for sequencing in a single sequencing run.Attorney Docket No. GH0240WO

[0152] Alternatively, in some embodiments, partition tags may be correlated to the sample as well as the partition. As a simple example, a first tag can indicate a first partition of a first sample; a second tag can indicate a second partition of the first sample; a third tag can indicate a first partition of a second sample; and a fourth tag can indicate a second partition of the second sample. While tags may be attached to molecules already partitioned based on one or more characteristics, the final tagged molecules in the library may no longer possess that characteristic. For example, while single stranded DNA molecules may be partitioned and tagged, the final tagged molecules in the library are likely to be double stranded. Similarly, while DNA may be subject to partition based on different levels of methylation, in the final library, tagged molecules derived from these molecules are likely to be unmethylated. Accordingly, the tag attached to a molecule in the library typically indicates the characteristic of the “parent molecule” from which the ultimate tagged molecule is derived, not necessarily to characteristic of the tagged molecule, itself.

[0153] As an example, barcodes 1, 2, 3, 4, etc. are used to tag and label molecules in the first partition; barcodes A, B, C, D, etc. are used to tag and label molecules in the second partition; and barcodes a, b, c, d, etc. are used to tag and label molecules in the third partition. Differentially tagged partitions can be pooled prior to sequencing. Differentially tagged partitions can be separately sequenced or sequenced together concurrently, e.g., in the same flow cell of an Illumina sequencer.

[0154] After sequencing, analysis of reads can be performed on a partition-by-partition level, as well as a whole DNA population level. Tags are used to sort reads from different partitions. Analysis can include in silico analysis to determine genetic and epigenetic variation (one or more of methylation, chromatin structure, etc.) using sequence information, genomic coordinates length, coverage, and / or copy number. In some embodiments, higher coverage can correlate with higher nucleosome occupancy in genomic region while lower coverage can correlate with lower nucleosome occupancy or a nucleosome depleted region (NDR).

[0155] Disclosed methods herein comprise analyzing DNA in a sample. In some embodiments described herein, the disclosed methods comprise partitioning DNA. In such methods, different forms of DNA (e.g., hypermethylated and hypomethylated DNA) can be physically partitioned based on one or more characteristics of the DNA. This approach can be used to determine, for example, whether certain sequences are hypermethylated or hypomethylated. In some embodiments, a first subsample or aliquot of a sample is subjected to steps for making capture probes as described elsewhere herein and a second subsample orAttorney Docket No. GH0240WO aliquot of a sample is subjected to partitioning. In some embodiments, a sample or subsample or aliquot thereof is subjected to partitioning and differential tagging, followed by a capture step using capture probes for rearranged sequences and optionally additional capture probes, e.g., for sequence-variable and / or epigenetic target regions.

[0156] Methylation profiling can involve determining methylation patterns across different regions of the genome. For example, after partitioning molecules based on extent of methylation (e.g., relative number of methylated nucleobases per molecule) and sequencing, the sequences of molecules in the different partitions can be mapped to a reference genome. This can show regions of the genome that, compared with other regions, are more highly methylated or are less highly methylated. In this way, genomic regions, in contrast to individual molecules, may differ in their extent of methylation.

[0157] Partitioning nucleic acid molecules in a sample can increase a rare signal, e.g., by enriching rare nucleic acid molecules that are more prevalent in one partition of the sample. For example, a genetic variation present in hypermethylated DNA but less (or not) present in hypomethylated DNA can be more easily detected by partitioning a sample into hypermethylated and hypomethylated nucleic acid molecules. By analyzing multiple partitions of a sample, a multi-dimensional analysis of a single molecule can be performed and hence, greater sensitivity can be achieved. Partitioning may include physically partitioning nucleic acid molecules into partitions or subsamples based on the presence or absence of one or more methylated nucleobases. A sample may be partitioned into partitions or subsamples based on a characteristic that is indicative of differential gene expression or a disease state. A sample may be partitioned based on a characteristic, or combination thereof that provides a difference in signal between a normal and diseased state during analysis of nucleic acids, e.g., cell free DNA (cfDNA), non-cfDNA, tumor DNA, circulating tumor DNA (ctDNA) and cell free nucleic acids (cfNA).

[0158] In some embodiments, hypermethylation and / or hypomethylation variable epigenetic target regions are analyzed to determine whether they show differential methylation characteristic of tumor cells or cells of a type that does not normally contribute to the DNA sample being analyzed (such as cfDNA), and / or particular immune cell types.

[0159] In some instances, heterogeneous DNA in a sample is partitioned into two or more partitions (e.g., at least 3, 4, 5, 6 or 7 partitions). In some embodiments, each partition is differentially tagged. Tagged partitions can then be pooled together for collective sample prep and / or sequencing. The partitioning-tagging-pooling steps can occur more than once, with each round of partitioning occurring based on a different characteristic (examplesAttorney Docket No. GH0240WO provided herein), and tagged using differential tags that are distinguished from other partitions and partitioning means. In other instances, the differentially tagged partitions are separately sequenced.

[0160] The agents used to partition populations of nucleic acids within a sample can be affinity agents, such as antibodies with the desired specificity, natural binding partners or variants thereof (Bock et al., Nat Biotech 28: 1106-1114 (2010); Song et al., Nat Biotech 29: 68-72 (2011)), or artificial peptides selected e.g., by phage display to have specificity to a given target. In some embodiments, the agent used in the partitioning is an agent that recognizes a modified nucleobase. In some embodiments, the modified nucleobase recognized by the agent is a modified cytosine, such as a methylcytosine (e.g., 5- methylcytosine). In some embodiments, the modified nucleobase recognized by the agent is a product of a procedure that affects the first nucleobase in the DNA differently from the second nucleobase in the DNA of the sample. In some embodiments, the modified nucleobase may be a “converted nucleobase,” meaning that its base pairing specificity was changed by a procedure. For example, certain procedures convert unmethylated or unmodified cytosine to dihydrouracil, or more generally, at least one modified or unmodified form of cytosine undergoes deamination, resulting in uracil (considered a modified nucleobase in the context of DNA) or a further modified form of uracil. Examples of partitioning agents include antibodies, such as antibodies that recognize a modified nucleobase, which may be a modified cytosine, such as a methylcytosine (e.g., 5- methylcytosine). In some embodiments, the partitioning agent is an antibody that recognizes a modified cytosine other than 5-methylcytosine, such as 5-carboxylcytosine (5caC). Alternative partitioning agents include methyl binding domain (MBDs) and methyl binding proteins (MBPs) as described herein, including proteins such as MeCP2.

[0161] Additional, non-limiting examples of partitioning agents are histone binding proteins which can separate nucleic acids bound to histones from free or unbound nucleic acids. Examples of histone binding proteins that can be used in the methods disclosed herein include RBBP4, RbAp48 and SANT domain peptides. In some embodiments, partitioning can comprise both binary partitioning and partitioning based on degree / level of modifications. For example, methylated fragments can be partitioned by methylated DNA immunoprecipitation (MeDIP), or all methylated fragments can be partitioned from unmethylated fragments using methyl binding domain proteins (e.g., MethylMinder Methylated DNA Enrichment Kit (ThermoFisher Scientific). Subsequently, additional partitioning may involve eluting fragments having different levels of methylation byAttorney Docket No. GH0240WO adjusting the salt concentration in a solution with the methyl binding domain and bound fragments. As salt concentration increases, fragments having greater methylation levels are eluted.

[0162] Analyzing DNA may comprise detecting or quantifying DNA of interest. Analyzing DNA can comprise detecting genetic variants and / or epigenetic features (e.g., DNA methylation and / or DNA fragmentation). In some embodiments, methylation levels can be determined using partitioning, modification-sensitive conversion such as bisulfite conversion, direct detection during sequencing, methylation-sensitive restriction enzyme digestion, methylation-dependent restriction enzyme digestion, or any other suitable approach. For example, different forms of DNA (e.g., hypermethylated and hypomethylated DNA) can be physically partitioned based on one or more characteristics of the DNA. For example, a methylated DNA binding protein (e.g., an MBD such as MBD2, MBD4, or MeCP2) or an antibody specific for 5-methylcytosine (as in MeDIP) can be used to partition the DNA. This approach can be used to determine, for example, whether certain sequences are hypermethylated or hypomethylated. In some embodiments, DNA fragmentation pattern can be determined based on endpoints and / or centerpoints of DNA molecules, such as cfDNA molecules.

[0163] In some instances, the final partitions are enriched in nucleic acids having different extents of modifications (overrepresentative or underrepresentative of modifications). Overrepresentation and underrepresentation can be defined by the number of modifications born by a nucleic acid relative to the median number of modifications per strand in a population. For example, if the median number of 5-methylcytosine residues in nucleic acid in a sample is 2, a nucleic acid including more than two 5-methylcytosine residues is overrepresented in this modification and a nucleic acid with 1 or zero 5-methylcytosine residues is underrepresented. The effect of the affinity separation is to enrich for nucleic acids overrepresented in a modification in a bound phase and for nucleic acids underrepresented in a modification in an unbound phase (i.e. in solution). The nucleic acids in the bound phase can be eluted before subsequent processing.

[0164] When using MeDIP or MethylMiner®Methylated DNA Enrichment Kit (ThermoFisher Scientific) various levels of methylation can be partitioned using sequential elutions. For example, a hypomethylated partition (no methylation) can be separated from a methylated partition by contacting the nucleic acid population with the MBD from the kit, which is attached to magnetic beads. The beads are used to separate out the methylated nucleic acids from the non- methylated nucleic acids. Subsequently, one or more elution stepsAttorney Docket No. GH0240WO are performed sequentially to elute nucleic acids having different levels of methylation. For example, a first set of methylated nucleic acids can be eluted at a salt concentration of 160 mM or higher, e.g., at least 150 mM, at least 200 mM, 300 mM, 400 mM, 500 mM, 600 mM, 700 mM, 800 mM, 900 mM, 1000 mM, or 2000 mM. After such methylated nucleic acids are eluted, magnetic separation is once again used to separate higher level of methylated nucleic acids from those with lower level of methylation. The elution and magnetic separation steps can be repeated to create various partitions such as a hypomethylated partition (enriched in nucleic acids comprising no methylation), a methylated partition (enriched in nucleic acids comprising low levels of methylation), and a hyper methylated partition (enriched in nucleic acids comprising high levels of methylation).

[0165] In some methods, nucleic acids bound to an agent used for affinity separation based partitioning are subjected to a wash step. The wash step washes off nucleic acids weakly bound to the affinity agent. Such nucleic acids can be enriched in nucleic acids having the modification to an extent close to the mean or median (i.e., intermediate between nucleic acids remaining bound to the solid phase and nucleic acids not binding to the solid phase on initial contacting of the sample with the agent).

[0166] The affinity separation results in at least two, and sometimes three or more partitions of nucleic acids with different extents of a modification. While the partitions are still separate, the nucleic acids of at least one partition, and usually two or three (or more) partitions are linked to nucleic acid tags, usually provided as components of adapters, with the nucleic acids in different partitions receiving different tags that distinguish members of one partition from another. The tags linked to nucleic acid molecules of the same partition can be the same or different from one another. But if different from one another, the tags may have part of their code in common so as to identify the molecules to which they are attached as being of a particular partition.

[0167] For further details regarding portioning nucleic acid samples based on characteristics such as methylation, see WO2018 / 119452, WO2024 / 159053, WO2021 / 236778, each of which are incorporated herein by reference.

[0168] In some embodiments, the nucleic acid molecules can be partitioned into different partitions based on the nucleic acid molecules that are bound to a specific protein or a fragment thereof and those that are not bound to that specific protein or fragment thereof.

[0169] Nucleic acid molecules can be partitioned based on DNA-protein binding. Protein- DNA complexes can be partitioned based on a specific property of a protein. Examples of such properties include various epitopes, modifications (e.g., histone methylation orAttorney Docket No. GH0240WO acetylation) or enzymatic activity. Examples of proteins which may bind to DNA and serve as a basis for fractionation may include, but are not limited to, protein A and protein G. Any suitable method can be used to partition the nucleic acid molecules based on protein bound regions. Examples of methods used to partition nucleic acid molecules based on protein bound regions include, but are not limited to, SDS-PAGE, chromatin-immuno-precipitation (ChIP), heparin chromatography, and asymmetrical field flow fractionation (AF4).

[0170] In some embodiments, the partitioning comprises contacting the DNA with a methylation sensitive restriction enzyme (MSRE) and / or a methylation dependent restriction enzyme (MDRE). Following the treatment of the DNA with a MSRE or a MDRE, the DNA may be partitioned based on size to generate hypermethylated (longest DNA molecules following MSRE treatment and shortest DNA fragments following MDRE treatment), intermediate (intermediate length DNA molecules following MSRE or MDRE treatment), and hypomethylated (shortest DNA molecules following MSRE treatment and longest DNA fragments following MDRE treatment) subsamples. In some embodiments, the partitioning is performed by contacting the nucleic acids with a methyl binding domain (“MBD”) of a methyl binding protein (“MBP”). In some such embodiments, the nucleic acids are contacted with an entire MBP. In some embodiments, an MBD binds to 5-methylcytosine (5mC), and an MBP comprises an MBD and is referred to interchangeably herein as a methyl binding protein or a methyl binding domain protein. In some embodiments, MBD is coupled to paramagnetic beads, such as Dynabeads® M-280 Streptavidin via a biotin linker. Partitioning into fractions with different extents of methylation can be performed by eluting fractions by increasing the NaCl concentration.

[0171] In some embodiments, bound DNA is eluted by contacting the antibody or MBD with a protease, such as proteinase K. This may be performed instead of or in addition to elution steps using NaCl as discussed above.

[0172] Examples of agents that recognize a modified nucleobase contemplated herein include, but are not limited to:(a) MeCP2 is a protein that preferentially binds to 5-methyl-cytosine over unmodified cytosine. b) RPL26, PRP8 and the DNA mismatch repair protein MHS6 preferentially bind to 5- hydroxymethyl-cytosine over unmodified cytosine.(c) FOXK1, FOXK2, FOXP1, FOXP4 and FOXI3 preferably bind to 5-formyl- cytosine over unmodified cytosine (lurlaro et al., Genome Biol. 14: R119 (2013)).Attorney Docket No. GH0240WO(d) Antibodies specific to one or more methylated or modified nucleobases or conversion products thereof, such as 5mC, 5caC, or DHU.

[0173] In general, elution is a function of the number of modifications, such as the number of methylated sites per molecule, with molecules having more methylation eluting under increased salt concentrations. To elute the DNA into distinct populations based on the extent of methylation, one can use a series of elution buffers of increasing NaCl concentration. Salt concentration can range from about 100 nm to about 2500 mM NaCl. In one embodiment, the process results in three (3) partitions. Molecules are contacted with a solution at a first salt concentration and comprising a molecule comprising an agent that recognizes a modified nucleobase, which molecule can be attached to a capture moiety, such as streptavidin. At the first salt concentration a population of molecules will bind to the agent and a population will remain unbound. The unbound population can be separated as a “hypomethylated” population. For example, a first partition enriched in hypomethylated form of DNA is that which remains unbound at a low salt concentration, e.g., 100 mM or 160 mM. A second partition enriched in intermediate methylated DNA is eluted using an intermediate salt concentration, e.g., between 100 mM and 2000 mM concentration. This is also separated from the sample. A third partition enriched in hypermethylated form of DNA is eluted using a high salt concentration, e.g., at least about 2000 mM.

[0174] In some embodiments, a monoclonal antibody raised against 5-methylcytidine (5mC) is used to purify methylated DNA. DNA is denatured, e.g., at 95°C in order to yield singlestranded DNA fragments. Protein G coupled to standard or magnetic beads as well as washes following incubation with the anti-5mC antibody are used to immunoprecipitate DNA bound to the antibody. Such DNA may then be eluted. Partitions may comprise unprecipitated DNA and one or more partitions eluted from the beads. In some embodiments, the partitions of DNA are desalted and concentrated in preparation for enzymatic steps of library preparation.

[0175] Sequences that comprise aberrantly high copy numbers may tend to be hypermethylated. Accordingly, in some embodiments, the DNA contacted with capture probes specific for members of an epigenetic target region set comprising a plurality of target regions that are both type-specific differentially methylated regions and copy number variants comprises at least a portion of a hypermethylated partition. The DNA from or comprising at least a portion of the hypermethylated partition may or may not be combined with DNA from or comprising at least a portion of one or more other partitions, such as an intermediate partition or a hypomethylated partition.Attorney Docket No. GH0240WOEXAMPLESExample 1 - Molecular Response Discrimination

[0176] The optimal ctDNA change threshold distinguishing molecular response (MR) from non-molecular response (nMR) was determined using a predefined training set. A power analysiswas conducted using the lifelines library’s sample size calculation function for Cox proportional hazards models. The Inventors modeled a hazard ratio (HR) of 2 between tumor-fraction (TF) responders and non-responders as a conservative estimate, informed by earlier studies.

[0177] Under 80% and 90% power targets (at a conventional alpha of 0.05), total sample sizes of 178 and 238 participants, respectively, were predicted to be sufficient. In a secondary scenario, using a higher HR of 3 and a fixed group size of 50 per arm (responder / non- responder), the Inventors observed an approximate 88% power. For the molecular response threshold determination, participants were assigned to training and validation subsets in a 60:40 ratio, employing a data-splitting approach that stratified individuals by the occurrence of a rwPFS event, ensuring balanced representation of progressions across the two subsets. Final training and test participant sizes were 382 and 245, respectively. Thus, both the training and test cohorts would provide adequate power to detect HRs of 2 or greater according to the predetermined analyses

[0178] in sample size determination. Both training and test groups were balanced for clinical characteristics.Example 2- Molecular Response Threshold Setting

[0179] For threshold determination, a subset of patients with successful baseline (TO) and early on-treatment samples (Tl), where at least one of those was above the limit of quantification (LOQ) was used. This resulted in an analyzable subset of 459 patients, comprising 273 from the training set and 186 from the test set. These participants were drawn from the same original cohorts, with exclusions applied solely based on the sample success and LOQ criterion for this specific analysis. The optimal threshold was determined by a balance of selecting the lowest hazard ratio (HR), shortest rwPFS for nMR group as well as the largest number of patients identified with MR across a range of thresholds from 0-100%. Threshold definitions for TF changes were finalized using only the training set, and those predefined thresholds were then applied to the test set. This approach enabled data-driven threshold selection in the training partition while preventing overfitting in the test partition and full cohort analysis.Attorney Docket No. GH0240WOExample 3 - Molecular Response in Immunotherapy Setting

[0180] Plasma samples were analyzed from patients (n=67) with aNSCLC, undergoing treatment with chemo in 1st line or later setting, who had serial genomic and also a tissue- free genomic and epigenomic ctDNA NGS test results. 27 patients had a baseline sample (Tl) within 105 days prior to treatment start and an on-treatment timepoint (T2) 2-15 weeks after treatment start and >10 days prior to treatment end. gMR was based on the %change in mean variant allele fraction (VAF) T2 / T1. gMR was based on the %change in mean variant allele fraction (VAF) T2 / T1. Baseline maximum VAF (maxVAF) from genomic / epigenomic and genomic results, limited to the 74 genes in common, were compared using Spearman’s rank correlation. Relationships between real-world progression-free survival (rwPFS) and gMR, TF %change from Tl to T2 (methylation-based MR; mMR), or TF at T2 were explored using Cox proportional hazards (CPH) analysis. Multiple cutpoints were evaluated to label patients as molecular responders or non-responders for outcomes analysis. Gender, age, and therapy line were included as covariates for all analyses.Example 4 - Molecular Response in Chemotherapeutic Setting

[0181] In NSCLC Pt-Chemo cohorts. Majority of therapy is still chemotherapy-based. In advanced NSCLC, prediction of outcomes on chemo is an unmet need.Tl maxVAF values were concordant between prior genomic data and genomic / egpigenomic data from 74 genes in common (r=0.81, p<0.001), confirming similarity in the genomic-based estimate of ctDNA level across the two assays despite the use of different aliquots of the same timepoint sampling. CPH revealed that gMR defined by any ctDNA % change threshold was not associated with rwPFS (p=(0.182-0.862)). However, a significant association with rwPFS was observed for TF at T2 (HR 6.31; 95% CI (1.58, 25.22); p=0.009) and for a >95% decrease in TF from Tl to T2 (HR 0.14; 95% CI (0.03, 0.80); p=0.027).

[0182] While these results appear to indicate that gMR does not always predict benefit from chemo in patients with aNSCLC, TF > 0.05% at T2 or mMR reflecting > 95% decrease in TF each were significantly associated with improved rwPFS, even in a small cohort. These results demonstrate feasibility of methylation-based TF to predict benefit from chemo in patients with aNSCLC without tissue and lay groundwork for future studies to address validity of serial TF monitoring in this patient population.

Claims

1. Attorney Docket No. GH0240WOTHE CLAIMS1. A method of determining a molecular response score, comprising: at least partially using a determining a first plurality of sequence reads and at least a second plurality of sequence reads associated with a subject; determining at least quantitative metric for the first plurality of sequence reads and at least one quantitative metric for the second plurality of sequence reads; and determining, based on the first quantitative metric and the second quantitative metric, a molecular response score.

2. The method of any preceding claim, wherein the first plurality of sequence reads are determined at a first time point and the second plurality of sequence reads are determined at a later time point.

3. The method of any preceding claim, wherein the method comprises additional pluralities of sequence reads at a plurality of later time points.

4. The method of any preceding claim, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy.

5. The method of any preceding claim, wherein the method comprises: classifying a plurality of variants in the first plurality of sequence reads as somatic or germline; and classifying the plurality of variants in the second plurality of sequence reads as somatic or germline.

6. The method of any preceding claim, wherein the method comprises determining a molecular response score using a predefined training set.

7. The method of any preceding claim, wherein the score includes comparison to a threshold.

8. The method of any preceding claim, wherein the threshold comprises a population-based training threshold.Attorney Docket No. GH0240WO9. The method of any preceding claim, wherein the method comprises determining, for at least one variant of the plurality of variants classified as somatic, based on a first variant allele fraction (VAF) and a second VAF, a quantitative metric for the first VAFs and a quantitative metric for the second VAFs.

10. The method of any preceding claim, wherein the method comprises: determining a ratio of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs; and determining a molecular response score based on a comparison of the quantitative metric for the first VAFs and the quantitative metric for the second VAFs.

11. The method of any preceding claim, wherein the method comprises: determining, for at least one variant of the plurality of variants classified as somatic, based on a first tumor fraction (TF) and a second TF, a quantitative metric for the first TF and a quantitative metric for the second TF.

12. The method of any preceding claim, wherein the method comprises methylation-based detection of TF.

13. The method of any preceding claim, wherein the method comprises: determining a ratio of the quantitative metric for the first TF and the quantitative metric for the second TF; and determining a molecular response score based on a comparison of the quantitative metric for the first TF and the quantitative metric for the second TF.

14. The method of any preceding claim, wherein the ratio comprises a hazard ratio.

15. A method of characterizing a sample, comprising: obtaining or having obtained sequence information for a first plurality of sequence reads and at least a second plurality of sequence reads derived from a sample from subject; determining at least quantitative metric for the first plurality of sequence reads and at least one quantitative metric for the second plurality of sequence reads; andAttorney Docket No. GH0240WO determining, based on the first quantitative metric and the second quantitative metric, a molecular response score, thereby characterizing the sample.

16. The method of any preceding claim, wherein the first plurality of sequence reads are determined at a first time point and the second plurality of sequence reads are determined at a later time point.

17. The method of any preceding claim, wherein the method comprises additional pluralities of sequence reads at a plurality of later time points.

18. The method of any preceding claim, wherein the first plurality of sequence reads are determined before administering a therapy and the second plurality of sequence reads are determined after administering the therapy.

19. The method of any preceding claim, wherein the method comprises: classifying a plurality of variants in the first plurality of sequence reads as somatic or germline; and classifying the plurality of variants in the second plurality of sequence reads as somatic or germline.

20. The method of any preceding claim, wherein the method comprises determining a molecular response score using a predefined training set.

21. The method of any preceding claim, wherein the score includes comparison to a threshold,22. The method of any preceding claim, wherein the threshold comprises a population-based training threshold.

23. The method of any preceding claim, wherein the method comprises: determining, for at least one variant of the plurality of variants classified as somatic, a quantitative metric for the first plurality of sequence reads and a quantitative metric for the second plurality of sequence reads;Attorney Docket No. GH0240WO comparing the quantitative metric for the first plurality of sequence reads and the quantitative metric for the second plurality of sequence reads; and determining a molecular response score based on the comparison.

24. The method of any preceding claim, wherein the method comprises determining on a tumor fraction (TF) estimation for the first, the second and additional pluralities of sequence reads.

25. The method of any preceding claim, wherein the method comprises methylation-based detection of TF.

26. The method of any preceding claim, wherein the comparison comprises determining a ratio of the quantitative metric for a first TF and the quantitative metric for a second TF; and determining a molecular response score.

27. The method of any preceding claim, wherein the ratio comprises a hazard ratio.

28. The method of any preceding claim, wherein the quantitative metric comprises a mean, median, mode, weighted mean, central median.

29. The method of any preceding claim, wherein methylation-based detection is measured through differentially methylated regions (DMRs).

30. A system for performing the method of any preceding claim.

31. A computer readable medium for performing the method of any preceding claim.

Images