Methods, kits, and systems for determining the HER2 status of a cancer, and methods of treating cancer based thereon

Abstract

This disclosure specifically includes methods, kits, and systems for determining the HER2 status of cancers such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer. In various embodiments, this disclosure relates to the use of one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation as characteristics of the HER2 status of cancer. In some embodiments, differential modifications and / or differential accessibility are detected and quantified at one or more genomic sites in a biological sample, such as cell-free DNA (cfDNA), obtained from or derived from a liquid biopsy sample from a subject with cancer. In various embodiments, the determined HER2 status can be used, for example, to select for the treatment and / or therapy of cancers such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer.

Description

[0001] Cross-references to related applications This application claims the benefit of U.S. Provisional Application No. 63 / 590,165, filed October 13, 2023; and U.S. Provisional Application No. 63 / 650,878, filed May 22, 2024; the entire contents of each application are incorporated herein by reference in their entirety. Background Technology

[0002] Human epidermal growth factor receptor 2 (HER2), also known as receptor tyrosine protein kinase erbB-2, is a well-known oncogenic driver in a variety of tumors and is an approved therapeutic target for breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer. Cancers that express too much HER2, i.e., HER2-positive cancers, tend to grow and spread more aggressively than HER2-negative cancers. Several agents targeting the HER2 protein, including trastuzumab, pertuzumab, margetuximab, trastuzumab-mtansine, trastuzumab-deruxtecan, lapatinib, neratinib, and tucatinib, have been approved by the US Food and Drug Administration (FDA) for the treatment of HER2-positive breast cancer and / or gastric / esophageal cancer (Meric-Bernstam et al., Clin Cancer Res (2019) 25(7):2033-2041 and Djaballah et al., American Society of Clinical Oncology Educational Book (2022) 42:219-232). The combination of tucatinib and trastuzumab was recently approved by the FDA for the treatment of patients with HER2-positive metastatic colorectal cancer (Casak et al., Clin Cancer Res (2023) 15:CCR-23-1041). Trastuzumab-drutecan was also recently approved by the FDA for the treatment of patients with unresectable or metastatic non-small cell lung cancer (NSCLC) tumors with activating HER2 mutations (Li et al., N Engl J Med (2022) 386:214-251).

[0003] To determine whether cancer is HER2-positive, medical professionals currently require tissue samples to be tested. Two types of tests are currently approved for HER2 diagnosis: immunohistochemistry (IHC) and in situ hybridization (ISH) (see Wolff et al., Arch Pathol Lab Med (2023) 147(9):993-1000, the entire contents of which are incorporated herein by reference). IHC is usually performed first because ISH is more expensive and takes longer to produce results.

[0004] IHC samples are scored by pathologists as 0, 1+, 2+, or 3+. If the IHC result is 0, the cancer is considered HER2-negative. These cancers do not respond to treatment with drugs targeting HER2 and can instead be treated with one or more of surgery and / or radiation, endocrine therapy (if hormone receptors such as estrogen receptors are positive), chemotherapy, and immunotherapy. If the IHC result is 1+, the cancer is considered HER2-negative. These cancers typically do not respond to treatment with drugs targeting HER2, but have recently shown responses to some HER2-targeting agents (e.g., see Nicolò et al., TherAdv Med Oncol (2023) 15:1-16). If the IHC result is 2+, the HER2 status of the tumor is indeterminate and is referred to as “indeterminate.” This means that an ISH test is needed to determine the HER2 status. If the IHC result is 3+, the cancer is HER2-positive. These cancers are typically treated with agents targeting HER2.

[0005] An IHC result of 1+ or 2+ with no detectable abnormalities based on IHC testing. ERRB2 Amputated cancers are also often referred to as HER2-low cancers. These cancers have recently been shown to respond to certain HER2-targeting agents, such as trastuzumab-drutecan, which was recently approved by the FDA for the treatment of patients with HER2-low metastatic breast cancer (see Nicolò et al., Ther Adv Med Oncol (2023) 15:1-16).

[0006] Currently approved IHC and ISH tests were originally developed to identify HER2-positive patients who could benefit from trastuzumab, and therefore are designed to identify HER2-overexpressing tumors rather than distinguishing between low HER2 expression levels and HER2 deficiency. Therefore, it remains unclear whether IHC and ISH tests are sufficient to detect low HER2 expression. Several pre-analytical and in-analytical issues may also affect the assessment of the extent of low HER2 expression, ultimately leading to high inconsistency among different pathologists when evaluating HER2 IHC tests. Different studies have reported differences in HER2 status observed in local and central laboratories. One study found that 85% of breast cancer specimens with a local HER2-0 score (n = 102) were HER2-low in a centralized reanalysis (Lambein et al., Am J Clin Pathol (2013) 140:561–566). Another published study aimed to assess the accuracy of low HER2 IHC scores, reporting that pathologists had only 26% agreement on scores of 0 and 1+, compared to 58% agreement on scores of 2+ and 3+ (Fernandez et al., JAMA Oncol (2022) 8(4):1-4). Another recently published study evaluated the agreement and inter-rater reliability of HER2 IHC scoring in 170 breast cancer biopsies by 18 breast cancer specialist pathologists from 15 institutions. The study assessed the observer methods required for subjective testing to determine the plateau in agreement and the minimum number of pathologists needed to estimate large inter-rater agreement values, as seen in real-world settings. The paper reported considerable inconsistency among intermediate categories in the four-category HER2 IHC scoring system (agreement <1% for 1+ and 3.6% for 2+) (Robbins et al., Mod Pathol (2023) 36(1):100032).

[0007] Because inaccuracies in HER2 status assessment (low or high scores) can lead to patients being incorrectly assigned to receive certain HER2-targeting agents, more accurate and objective diagnostic methods for determining HER2 status remain needed, including methods independent of IHC or ISH scores. Improved diagnostic methods will also better support future clinical trials designed to identify patient subgroups that respond to HER2-targeting agents. They will also expand our understanding of the underlying biology of cancer and help identify new treatment options. Summary of the Invention

[0008] This disclosure is based, at least in part, on the evidence that the HER2 status of a subject's cancer can be determined by detecting and quantifying the presence of histone modifications and / or DNA methylation at one or more genomic sites in cell-free DNA (cfDNA) from liquid biopsy samples (e.g., plasma samples obtained from or derived from a subject). This disclosure also covers methods for detecting chromatin accessibility and / or binding of one or more transcription factors at one or more genomic sites, rather than (or in addition to) histone modifications and / or DNA methylation. This disclosure is also based, at least in part, on the evidence that genomic sites differentially modified based on different types of histone modifications (e.g., histone methylation markers such as H3K4me3 and histone acetylation markers such as H3K27ac) and / or DNA methylation can be combined into multimodal classifiers to determine HER2 status. These novel unimodal and multimodal classifiers provide minimally invasive methods for determining HER2 status that are more accurate, objective, and comprehensive than current tissue-based methods. To date, no liquid biopsy platform has provided an operational solution for therapy-related transcriptional regulatory phenotypes such as HER2 status. The importance of HER2 intermediate expression states (e.g., low HER2) for HER2-targeting agents is becoming increasingly apparent, further highlighting this need; however, current tissue-based HER2 IHC / ISH methods struggle to objectively determine these states.

[0009] This disclosure specifically includes techniques for determining HER2 status, and techniques for detecting, monitoring, and / or treating cancers (including, for example, breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) based on HER2 status. In various embodiments, this disclosure relates to measuring histone modifications in samples obtained from or derived from a subject to detect and / or treat cancers (including, for example, breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) based on HER2 status. This disclosure specifically includes measurements of histone modifications in cell-free DNA (cfDNA) that are characteristic of cancer and, in various embodiments, can be used, for example, for detecting, monitoring, selecting treatments for, and / or treating cancers (including, for example, breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) based on HER2 status. In some embodiments, measurements of histone modifications in cfDNA can be used to detect or determine resistance to therapy or transformation of cancer (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) from one subtype to another. In various embodiments, this disclosure includes exemplary genomic sites that are differentially modified in HER2-positive and HER2-negative cancers, such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer. In various embodiments, the differentially modified genomic site in cfDNA is or includes one or more enhancers. In various embodiments, the differentially modified genomic site in cfDNA is or includes one or more promoters. In various embodiments, the differentially modified genomic site in cfDNA originates from a HER2 amplicon. In various embodiments, the differentially modified genomic site in cfDNA does not originate from a HER2 amplicon. Those skilled in the art can determine whether a genomic site originates from a HER2 amplicon. In some embodiments, “HER2 amplicon” corresponds to a region surrounding the ERBB2 transcript on chromosome 17, which typically contains the genes STARD3, GRB7, and NR1D1. In some embodiments, “HER2 amplicon” corresponds to chr17:37,776,630-38,274,514 based on human genome version hg19.

[0010] In various embodiments, the genomic site is differentially modified if it is characterized by an increase or decrease in histone modifications compared to a reference (e.g., a sample from a HER2-negative subject or a healthy subject). The increase or decrease in histone modifications can be, or include, for example, an increase or decrease in histone methylation of one or more specific methylation markers (high methylation or low methylation, respectively), or a combination thereof; an increase or decrease in pan-methylation; an increase or decrease in histone acetylation of one or more specific acetylation markers (high acetylation or low acetylation, respectively), or a combination thereof; and / or an increase or decrease in pan-acetylation (e.g., pan-H3 acetylation). In various embodiments, histone methylation can be, or include, histone methylation markers selected from H3K4me1, H3K4me2, H3K4me3, or combinations thereof. In various embodiments, histone methylation can be, or include H3K4me3. In various embodiments, histone acetylation may be or includes a histone acetylation marker selected from H3K9ac, H3K14ac, H3K18ac, H3K23ac, H3K27ac, or combinations thereof. In various embodiments, histone acetylation may be or includes H3K27ac.

[0011] In various embodiments, this disclosure relates to measuring DNA methylation in samples obtained from or derived from a subject for HER2-based detection and / or treatment of cancers (including, for example, breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer). This disclosure particularly includes measurements of DNA methylation in cell-free DNA (cfDNA), histone modification measurements that are characteristic of cancers, and in various embodiments, can be used, for example, for HER2-based detection, monitoring, selection of treatments for cancers (including, for example, breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer), and / or treatment of cancer. In some embodiments, measurements of DNA methylation in cfDNA can be used to detect or determine resistance to therapy or transformation of cancer (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) from one subtype to another. In various embodiments, this disclosure includes exemplary genomic sites that are differentially DNA-methylated in HER2-positive cancers versus HER2-negative cancers, such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer. In various embodiments, the genomic site is differentially modified if it is characterized by increased or decreased DNA methylation compared to a reference (e.g., a sample from a HER2-negative subject or a healthy subject). In various embodiments, the differentially modified genomic site in the cfDNA is or includes one or more enhancers. In various embodiments, the differentially modified genomic site in the cfDNA is or includes one or more promoters. In various embodiments, the differentially modified genomic site in the cfDNA originates from a HER2 amplicon. In various embodiments, the differentially modified genomic site in the cfDNA does not originate from a HER2 amplicon.

[0012] This invention further relates to measuring chromatin accessibility in cell-free DNA (cfDNA) in various embodiments to determine HER2 status. This disclosure particularly includes measurements of chromatin accessibility in cfDNA that are characteristic of HER2-positive cancers, which in various embodiments may be used, for example, to detect, monitor, select for treatment of HER2-positive cancers, and / or to treat HER2-positive cancers. In some embodiments, measurements of chromatin accessibility in cfDNA may be used to detect or determine resistance to therapy or conversion of cancer (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) from one subtype to another. In various embodiments, this disclosure includes genomic sites that are differentially accessible in HER2-positive and HER2-negative cancers. In various embodiments, differentially accessible genomic sites in cfDNA are or include one or more enhancers. In various embodiments, differentially accessible genomic sites in cfDNA are or include one or more promoters. In various embodiments, differentially accessible genomic sites in cfDNA originate from HER2 amplicones. In various implementation schemes, the differentially accessible genomic sites in cfDNA do not originate from the HER2 amplicon.

[0013] In various embodiments, while not wishing to be bound by any particular scientific theory, histone acetylation (e.g., H3K27ac) corresponds to and / or is related to chromatin accessibility. In various embodiments, while not wishing to be bound by any particular scientific theory, histone methylation (e.g., H3K4me3) corresponds to and / or is related to chromatin accessibility. In various embodiments, while not wishing to be bound by any particular scientific theory, DNA methylation corresponds to and / or is related to chromatin accessibility.

[0014] In various embodiments, a genomic locus is differentially accessible if it is characterized by increased or decreased chromatin accessibility compared to a reference (e.g., a sample from a HER2-negative subject or a healthy subject). Increased or decreased histone modifications may be, or include, for example, increases or decreases in accessibility as determined by various chromatin accessibility assays known in the art.

[0015] This invention further relates to measuring transcription factor binding in cell-free DNA (cfDNA) in various embodiments to determine HER2 status. This disclosure particularly includes transcription factor binding measurements in cfDNA that are characteristic of HER2-positive cancers, which in various embodiments can be used, for example, to detect, monitor, select for treatment of HER2-positive cancers, and / or to treat HER2-positive cancers. In some embodiments, transcription factor binding measurements in cfDNA can be used to detect or determine resistance to therapy or conversion of cancer (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, and lung cancer) from one subtype to another. In various embodiments, this disclosure includes genomic sites in HER2-positive and HER2-negative cancers that are differentially bound by transcription factors. In various embodiments, the genomic sites in cfDNA that are differentially bound by transcription factors are or include one or more enhancers. In various embodiments, the genomic sites in cfDNA that are differentially bound by transcription factors are or include one or more promoters. In various embodiments, the genomic sites in cfDNA that are differentially bound by transcription factors originate from HER2 amplicones. In various implementation schemes, the genomic sites in cfDNA that are differentially bound by transcription factors do not originate from the HER2 amplicon.

[0016] In various embodiments, while not wishing to be bound by any particular scientific theory, histone acetylation (e.g., H3K27ac) corresponds to and / or is associated with transcription factor binding. In various embodiments, while not wishing to be bound by any particular scientific theory, histone methylation (e.g., H3K4me3) corresponds to and / or is associated with transcription factor binding. In various embodiments, while not wishing to be bound by any particular scientific theory, DNA methylation corresponds to and / or is associated with transcription factor binding.

[0017] In various embodiments, if a genomic locus is characterized by increased or decreased transcription factor binding compared to a reference (e.g., a sample from a HER2-negative subject or a healthy subject), then the genomic locus is differentially bound by transcription factors. Increased or decreased transcription factor binding may be, for example, an increase or decrease in transcription factor binding as determined by various transcription factor binding assays known in the art.

[0018] In one aspect, this disclosure provides a method for determining the HER2 status of a subject's cancer, the method comprising: quantifying at one or more genomic sites in a biological sample obtained from or derived therefrom, optionally from a liquid biopsy sample, one or more of the following genomic sites: (i) one or more histone modifications, (ii) chromatin accessibility, (iii) binding of one or more transcription factors, and / or (iv) DNA methylation, optionally wherein one or more quantified genomic sites are not derived from HER2 amplicon.

[0019] In some embodiments, a histone modification assay is used to quantify one or more histone modifications, said assay measuring one or more of H3K9ac, H3K14ac, H3K18ac, H3K23ac, H3K27ac, H3K4me1, H3K4me2, H3K4me3, and panacetylation. In some embodiments, the histone modification assay detects H3K4me3 modification. In some embodiments, the histone modification assay detects H3K27ac modification. In some embodiments, the histone modification assay is selected from ChIP-seq (chromatin immunoprecipitation sequencing), CUT&RUN (target cleavage and nuclease release) sequencing, and CUT&Tag (target cleavage and fragmentation labeling) sequencing.

[0020] In some implementations, chromatin accessibility is quantified using chromatin accessibility assays selected from ATAC-seq (transposon accessibility chromatin sequencing assay), NOMe-seq (nucleosome occupancy and methylome sequencing), FAIRE-seq (formaldehyde-assisted separation of regulatory elements sequencing), MNase-seq (micrococcal nuclease digestion sequencing), and DNase hypersensitivity assays.

[0021] In some embodiments, a transcription factor binding assay is used to quantify the binding of one or more transcription factors, said assay detecting the binding of one or more of p300, mediator complex, cohesin complex, RNA pol II, FOXA1, ESR1, PR, MYC, EN1, FOXM1, KLF4, AP-2, RARA, or RUNX1. In some embodiments, the transcription factor binding assay is selected from ChIP-seq (chromatin immunoprecipitation sequencing), CUT&RUN (target cleavage and nuclease release) sequencing, and CUT&Tag (target cleavage and fragmentation labeling) sequencing.

[0022] In some implementations, DNA methylation is quantified using bisulfite sequencing (BS-Seq), whole-genome bisulfite sequencing (WGBS), methylated DNA immunoprecipitation sequencing (MeDIP-seq), or methyl-CpG-binding domain sequencing (MBD-seq).

[0023] In some embodiments, the method includes quantifying two or more of the following at one or more genomic sites in cell-free DNA (cfDNA) obtained from or derived from a liquid biopsy sample of a subject: (i) one or more histone modifications, (ii) chromatin accessibility, (iii) transcription factor binding, and / or (iv) DNA methylation. In some embodiments, the method includes quantifying two or more histone modifications, for example, quantifying H3K4me3 and H3K27ac modifications. In some embodiments, the method includes quantifying one or more histone modifications and DNA methylation, for example, quantifying H3K4me3 and / or H3K27ac modifications and DNA methylation. In some embodiments, the method includes quantifying H3K4me3 modification, H3K27ac modification, and DNA methylation.

[0024] In some embodiments, the biological sample is a liquid biopsy sample, such as a plasma sample, serum sample, or urine sample. In some embodiments, the method includes isolating DNA (e.g., cfDNA) from a liquid biopsy sample (e.g., a plasma sample) of 1 mL, 2 mL, 3 mL, 4 mL, or 5 mL.

[0025] In some implementations, the sample is a liquid biopsy sample containing cfDNA, and the method includes: (a) Quantify H3K4me3 modification at one or more genomic loci using an assay that includes enriching cfDNA containing one or more H3K4me3 modifications and sequencing the cfDNA enriched for H3K4me3 modifications (e.g., using cfChIP-seq assay). (b) Quantifying H3K27ac modifications at one or more genomic loci using an assay that includes enriching cfDNA containing one or more H3K27ac modifications and sequencing the H3K27ac-enriched cfDNA (e.g., using cfChIP-seq assay); and / or; (c) Quantify methylated DNA using a assay that includes enriching methylated cfDNA and sequencing the enriched cfDNA to determine the sequence count of sequences having one or more methylated nucleotides (e.g., using an MBD-seq assay).

[0026] In some implementation schemes, (a) Enriching cfDNA containing H3K4me3 modification using a method that includes incubating the sample with an agent that binds to H3K4me3 modification (e.g., an antibody); (b) Enriching cfDNA containing H3K27ac modification using methods including incubating the sample with an agent that binds to H3K27ac modification (e.g., an antibody); and / or (c) Enrich methylated cfDNA using a method that includes incubating the sample with an agent that binds methylated DNA (e.g., an antibody or a methyl-binding domain).

[0027] In some embodiments, an agent that binds to H3K4me3 modification, an agent that binds to H3K27ac modification, and / or an agent that binds to methylated DNA (e.g., via covalent or non-covalent bonds) is attached to a physical support (e.g., beads, magnetic beads, agarose beads, or magnetic epoxy beads) and then incubated with the sample.

[0028] In some implementations, if the method includes incubation with two or more of the following: (a) an agent that binds to H3K4 modification, an agent that binds to H3K27ac modification, and an agent that binds to methylated DNA, the sample is incubated with the two or more agents in the following manner: (1) sequentially, or (2) in parallel (e.g., where the sample is divided into several portions and each portion is incubated with a different agent).

[0029] In some implementations, sequencing is performed using next-generation sequencing methods.

[0030] In some implementations, the method includes attaching (e.g., linking) an adaptor to cfDNA obtained from a subject (e.g., after enriching cfDNA for cfDNA containing one or more H3K4me3 modifications, cfDNA containing one or more H3K27ac modifications, and / or methylated cfDNA).

[0031] In some implementations, the method includes amplifying multiple transformed DNA fragments after attaching an adaptor to multiple DNA fragments.

[0032] In some implementations, mapped sequence reads are mapped to a reference genome. In some implementations, non-unique mappings and / or redundant sequence reads are discarded.

[0033] In some implementations, sequence reads are mapped to a reference genome, and one or more genomic loci correspond to sequence read peaks, where sequence read peaks correspond to regions in the genome where the number of sequence reads is higher than the local background. In some implementations, when identifying genomic loci where the number of sequence reads is higher than the local background, peaks in high-noise regions are ignored. In some implementations, peaks in regions with high levels of epigenetic markers in leukocytes are removed. In some implementations, peaks in regions that may be artifacts are removed. In some implementations, peaks shorter than 50 bp are removed.

[0034] In some embodiments, quantifying H3K4me3 modifications, H3K27ac modifications, and / or DNA methylation involves summing the number of sequence reads that overlap with at least one nucleotide at one or more genomic sites. In some embodiments, before summing, the sequence reads are adjusted according to sequencing depth (e.g., normalizing the sequence read quantiles to a common reference distribution) and / or ChIP quality. In some embodiments, the sequence counts are normalized to an aggregate count of a set of regions (e.g., 10,000 regions) in a given sample that have previously been identified as having DNAse hypersensitivity in most cell types. In some embodiments, before summing, an estimate of the local background signal is subtracted from the sequence reads at each genomic site.

[0035] In some embodiments, the method further includes comparing measurements of one or more epigenetic biomarkers with reference values. In some embodiments, the reference value is a predetermined threshold, a measurement of a liquid biopsy sample, a measurement of a liquid biopsy sample obtained from a cohort of subjects, and / or a normalized value. In some embodiments, the predetermined threshold or normalized value has been previously demonstrated to distinguish between HER2-positive and HER2-negative cancers (e.g., by an AUROC greater than 0.5). In some embodiments, the reference value is a measurement of a liquid biopsy sample obtained from a cohort of subjects previously identified as having HER2-positive and HER2-negative cancer. In some embodiments, a cohort of subjects has been previously identified as having cancer (e.g., breast cancer).

[0036] In some implementations, the method includes calculating the α-sequence read density at one or more genomic loci. In some implementations, the sequence read density can be calculated by methods including the following: (a) Summing the background-adjusted sequence counts at each of one or more genomic loci and dividing by the sum of the kilobases at one or more genomic loci; or (b) For each genomic locus, divide the background adjustment fragment count by the number of kilobases of the genomic locus, and then sum over each locus.

[0037] In some implementations, one or more genomic loci include one or more genomic loci with increased levels of one or more of the following epigenetic biomarkers: (a) samples obtained from subjects with HER2-positive cancer compared to samples obtained from subjects with HER2-negative cancer; and / or (b) samples obtained from subjects with HER2-positive cancer compared to samples obtained from subjects with HER2-negative cancer.

[0038] In some implementations, the method includes calculating a HER2-positive / HER2-negative ratio score. In some implementations, the HER2-positive / HER2-negative ratio score can be calculated using methods including the following: (a) The HER2 positive sequence read density is calculated by summing the background-adjusted sequence counts at each of the one or more genomic loci where the level of the one or more epigenetic biomarkers is increased in a sample obtained from a subject with HER2-negative cancer compared to a sample obtained from a subject with HER2-positive cancer. (b) Calculating the HER2-negative sequence read density by summing background-adjusted sequence counts at each of the one or more genomic loci where the levels of the one or more epigenetic biomarkers are increased, compared to samples obtained from subjects with HER2-positive cancer; and (c) Divide the HER2 positive sequence read density by the HER2 negative sequence read density.

[0039] In some embodiments, the method includes determining a HER2-positive / HER2-negative ratio score for two or more epigenetic markers. In some embodiments, the method includes: (a) Determine the HER2 positive / HER2 negative ratio score for H3K4me3 modification; (b) Determine the H3K27ac modified HER2 positive / HER2 negative ratio score; and / or (c) Determine the HER2 positive / HER2 negative ratio score of methylated DNA.

[0040] In some embodiments, the method includes determining HER2-positive / HER2-negative ratio scores for two or more epigenetic biomarkers, wherein the HER2-positive / HER2-negative ratio scores are combined. In some embodiments, the HER2-positive / HER2-negative ratio score for each of H3K4me3, H3K27ac, and methylated DNA is determined, and each ratio score is combined. In some embodiments, fitted values ​​determined by logistic regression may be used to combine two or more ratio scores.

[0041] In some implementations, quantification of one or more histone modifications at one or more genomic sites, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation, compared to a reference, indicates that the subject has HER2-positive cancer, optionally based on HER2-3+ cancer as determined by an IHC test.

[0042] In some implementations, quantification of one or more histone modifications at one or more genomic sites, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation, compared to a reference, indicates that the subject has HER2-negative cancer, optionally based on HER2-0 cancer as determined by an IHC test.

[0043] In some implementations, the cancer is breast cancer, stomach / gastroesophageal cancer, colorectal cancer, or lung cancer. In some implementations, the cancer is breast cancer.

[0044] In some implementations, the reference value is a predetermined threshold, a measurement of a liquid biopsy sample, and / or a normalized value, optionally wherein the reference value is a measurement of a liquid biopsy sample obtained from a group of subjects who have been previously diagnosed with HER2-negative cancer, optionally with HER2-0 cancer based on IHC testing, or who do not have cancer.

[0045] In some embodiments, the method includes quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic loci in Tables 1 through 3, optionally wherein one or more genomic loci do not originate from HER2 amplicons. In some embodiments, the method includes quantifying H3K4me3 modifications at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 1, optionally wherein one or more genomic loci do not originate from HER2 amplicons.

[0046] In some embodiments, the method includes quantifying H3K27ac modifications at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 2, optionally one or more of these genomic loci not originating from HER2 amplicones. In some embodiments, the method includes quantifying DNA methylation at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 3, optionally one or more of these genomic loci not originating from HER2 amplicones.

[0047] In some implementations, the quantification of one or more epigenetic biomarkers at one or more genomic loci, compared to reference values, indicates that the subject has HER2-positive cancer, optionally based on HER2-3+ cancer as determined by IHC testing.

[0048] In some implementations, the quantification of one or more epigenetic biomarkers at one or more genomic loci, compared to reference values, indicates that the subject has HER2-negative cancer, optionally based on HER2-0 cancer as determined by IHC testing.

[0049] In some implementations, the sample contains a detectable amount of ctDNA (e.g., where the estimated tumor fraction of cfDNA is >3%, e.g., determined by iChorCNA).

[0050] In some implementations, the area under the recipient operating characteristic curve (AUROC) used to determine whether a subject has HER2-positive or HER2-negative cancer is greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95).

[0051] In some implementations, HER2-positive cancer is HER2-3+ cancer based on IHC testing, and HER2-negative cancer is HER2-0 cancer based on IHC testing. In some implementations, the subject has been previously diagnosed with cancer.

[0052] In another aspect, this disclosure provides a method for treating a subject with cancer, the method comprising: administering a cancer therapy to the subject based on the HER2 status of the cancer, wherein the HER2 status of the cancer is determined using any of the methods described above for determining HER2 status. In some embodiments, the method further comprises determining the HER2 status of the cancer using any of the methods described above for determining HER2 status. In some embodiments, the cancer has been determined to be HER2-positive, and the cancer therapy is a HER2-targeting agent. In some embodiments, the cancer has been determined to be HER2-negative, and the cancer therapy is not a HER2-targeting agent.

[0053] In another aspect, this disclosure provides a method for monitoring the HER2 status of a subject's cancer and optionally treating the cancer, the method comprising: determining the HER2 status of the cancer at a first time point and a second time point using any of the methods described above for determining HER2 status. In some embodiments, a HER2-targeting agent is administered to the subject after the first time point and before the second time point. In some embodiments, the method further comprises administering cancer therapy, optionally a HER2-targeting agent, to the subject based on the HER2 status of the cancer at the second time point, optionally wherein the type, dose, and / or frequency of administration of the cancer therapy is adjusted based on the HER2 status of the cancer at the second time point.

[0054] In another aspect, this disclosure provides a method for treating a subject with cancer, the method comprising: administering a HER2-targeting agent to the subject if, based on analysis of cell-free DNA (cfDNA) from a biological sample obtained from or derived therefrom, optionally from a liquid biopsy sample, the subject has been determined to possess a validated epigenetic characteristic indicating HER2-positive cancer; and not administering the HER2-targeting agent if, if the subject has not been determined to possess a validated epigenetic characteristic indicating HER2-positive cancer, the presence of the validated epigenetic characteristic has been determined using a validated classifier, wherein the validated classifier is obtained by: (a) determining from a first cohort previously identified as having HER2-positive cancer, optionally... (i) one or more HER2-positive cell lines or (ii) one or more genomic features of histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation in subjects with HER2-3+, HER2-2+, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing; (b) identifying genomic features of histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation in (i) one or more HER2-negative cell lines or (ii) one or more HER2-0 cancer based on IHC testing from healthy subjects in a second cohort or subjects previously diagnosed with HER2-negative cancer. (c) Genomic features of transcription factor binding and / or DNA methylation; (d) Comparing the genomic features identified in step (a) and step (b) to identify genomic sites (“differential sites”) with statistically distinct levels of histone modifications, chromatin accessibility, transcription factor binding, and / or DNA methylation; and (e) Training a classifier against the levels of histone modifications, chromatin accessibility, transcription factor binding, and / or DNA methylation at differential sites to distinguish (i) samples from one or more HER2-positive cell lines or biological samples obtained from the first cohort and (ii) samples from one or more HER2-negative cell lines or biological samples obtained from the second cohort to identify genomic sites with statistically distinct levels of histone modifications, chromatin accessibility, transcription factor binding, and / or DNA methylation. Samples with characteristics of chromatin accessibility, transcription factor binding, and / or DNA methylation levels (“epigenetic features”) indicating that the sample may be obtained from a HER2-positive cell line or from a first cohort; and (e) obtaining a validated classifier by validating the classifier in step (d) against a third cohort comprising independent, blinded subjects with HER2-positive and HER2-negative cancers, and selecting a threshold such that the validated classifier predicts HER2-positive cancers, optionally based on HER2-3+, HER2-2+, or HER2-1+ cancers tested by IHC, or HER2-low cancers tested by IHC / ISH, with an area under the recipient operating characteristic curve (AUROC) greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95), wherein subjects falling into the predicted HER2-positive cancer group exhibited validated epigenetic characteristics, and subjects not falling into the HER2-positive cancer group lacked validated epigenetic characteristics.

[0055] In some implementations, the differential sites in step (c) are determined by comparing genomic features of one or more histone modifications and / or DNA methylation in (i) one or more HER2-positive cell lines and (ii) one or more HER2-negative cell lines.

[0056] In some implementations, the classifier in step (d) is targeted at... Computer simulation Training was performed on differentially expressed sites, including histone modifications, chromatin accessibility, transcription factor binding, and / or DNA methylation levels, obtained by mixing sequence data from one or more HER2-positive cell lines and sequence data from liquid biopsy samples from healthy subjects.

[0057] In some implementations, the classifier validated in step (e) is validated using liquid biopsy samples from a third queue.

[0058] In some embodiments, the classifier in step (d) is trained on two or more histone modification levels at differentially expressed sites. In some embodiments, the two or more histone modification levels include H3K4me3 and H3K27ac modification levels.

[0059] In some embodiments, the classifier in step (d) is trained on one or more histone modification levels and DNA methylation at differential sites. In some embodiments, the one or more histone modification levels include H3K4me3 and / or H3K27ac modification levels. In some embodiments, the classifier in step (d) is trained using ridge regression, elastic network regression, or lasso regression. In some embodiments, the one or more histone modification levels include H3K4me3 and H3K27ac modification levels. In some embodiments, the biological sample is a liquid biopsy sample, such as a plasma sample, serum sample, or urine sample.

[0060] In some implementation schemes, Genomic features of one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation have been identified in one or more HER2-positive cell lines; and In step (d), sequencing fragments from healthy donor plasma samples and cell lines are mixed in different proportions. Computer simulation Dilute the sample to achieve a simulated ctDNA percentage in the range of 0.5% to 50%.

[0061] In some implementations, genomic features such as one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation have been identified in one or more HER2-positive cell lines; and classifiers have been tuned using plasma data.

[0062] In some implementations, a transfer learning process is used to adjust the classifier using plasma data, the transfer learning process including: (i) Use a classifier to calculate the predicted value of the plasma sample in the form of a probability value (e.g., using the formula log2(HER2+ probability / 1 – HER2+ probability)); (ii) In a new model (e.g., a lasso logistic regression model), use the odds value as an offset term, use all the same features, but train with plasma data (e.g., leave-one-out method) to determine new weights and coefficients, and then add these weights and coefficients to the coefficients determined during cell line training to obtain an adjusted model.

[0063] In some implementations, models are trained and adjusted for cancer-specific genomic loci in plasma data, where cancer-specific genomic loci are regions in HER2-positive subjects that are enriched with H3K4me3 and H3K27ac modifications compared to HER2-negative subjects and that are associated with ctDNA at HER2 loci.

[0064] In another aspect, this disclosure provides a kit comprising reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic loci, wherein the one or more genomic loci are selected from Tables 1 to 3, and optionally, one or more genomic loci are not derived from HER2 amplicons. In some embodiments, the kit comprises reagents for quantifying H3K4me3 at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 1, optionally, one or more genomic loci are not derived from HER2 amplicons. In some embodiments, the kit comprises reagents for quantifying H3K27ac at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 2, optionally, one or more genomic loci are not derived from HER2 amplicons. In some embodiments, the kit comprises reagents for quantifying DNA methylation at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 3, optionally, one or more genomic loci are not derived from HER2 amplicons.

[0065] In some embodiments, the kit includes one or more antibodies for ChIP-seq, optionally said one or more antibodies specifically binding to H3K4me3 or H3K27ac modified histones. In some embodiments, the kit includes one or more methyl-binding domains for MBD-seq.

[0066] In some embodiments, the kit includes reagents for isolating cell-free DNA (cfDNA) from liquid biopsy samples. In some embodiments, the kit includes reagents for sequencing library preparation. In some embodiments, the kit includes reagents for sequencing. In some embodiments, the kit includes instructions for determining whether a subject has HER2-positive cancer.

[0067] In another aspect, this disclosure provides a non-transient computer-readable storage medium encoded with a computer program, wherein the program contains instructions that, when executed by one or more processors, cause the one or more processors to perform operations to execute any of the methods described above for determining the HER2 state.

[0068] In another aspect, this disclosure provides a computer system including a memory and one or more processors coupled to the memory, wherein the one or more processors are configured to perform operations to execute any of the methods described above for determining HER2.

[0069] In another aspect, this disclosure provides a system for determining the HER2 status of a subject's cancer, the system comprising a sequencer configured to generate a sequencing dataset from a sample; and a non-transient computer-readable storage medium and / or computer system of this disclosure. In some embodiments, the sequencer is configured to generate a whole-genome sequencing (WGS) dataset from a sample. In some embodiments, the system further includes a sample preparation device. In some embodiments, the sample preparation device is configured to prepare a sample for sequencing from a biological sample (optionally, a liquid biopsy sample). In some embodiments, the sample preparation device includes reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic sites in cell-free DNA (cfDNA) from a biological sample (optionally, a liquid biopsy sample). In some embodiments, one or more genomic sites are selected from Tables 1 to 3. In some embodiments, the device includes reagents for quantifying H3K4me3 at, for example, at least 5, 10, 20, 30, 40, or 50 genomic sites in Table 1. In some embodiments, the device includes reagents for quantifying H3K27ac at, for example, at least 5, 10, 20, 30, 40, or 50 genomic loci as shown in Table 2. In some embodiments, the device includes reagents for quantifying DNA methylation at, for example, at least 5, 10, 20, 30, 40, or 50 genomic loci as shown in Table 3. In some embodiments, the reagents comprise one or more antibodies for ChIP-seq, optionally said antibodies specifically binding to H3K4me3 or H3K27ac-modified histones. In some embodiments, the reagents comprise one or more methyl-binding domains for MBD-seq. In some embodiments, the device includes reagents for isolating cell-free DNA (cfDNA) from biological samples (optionally, liquid biopsy samples). In some embodiments, the device includes reagents for sequencing library preparation. In some embodiments, the sequencer includes reagents for sequencing.

[0070] In some implementations, the method is used to determine the HER2 status of a subject's (e.g., a patient's) cancer. The method may include receiving (e.g., via a processor of a computing device) one or more genomic features of the subject, including one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation. The method may further include using a HER2 classifier to classify the genomic features to determine whether the subject possesses epigenetic features indicative of HER2-positive cancer.

[0071] In some embodiments, the HER2 classifier has been trained using one or more genomic features including one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation: (i) one or more HER2-positive cell lines and one or more HER2-negative cell lines and / or (ii) one or more biological samples obtained from one or more subject cohorts previously identified as having HER2-positive cancer, optionally HER2-3+, HER2-2+, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing, and one or more biological samples obtained from one or more subject cohorts previously identified as having HER2-negative cancer, optionally HER2-0 cancer based on IHC testing. In some embodiments, one or more genomic features used to train the HER2 classifier include one or more genomic features generated by computer simulation of diluting sequence data from HER2-positive or HER2-negative cell lines with sequence data obtained from healthy donor plasma samples to achieve a simulated ctDNA percentage in the range of 0.5% to 50%.

[0072] In some embodiments, one or more genomic features target differentially expressed genomic sites found between one or more HER2-positive cell lines and one or more HER2-negative cell lines and / or from one or more subject cohorts previously identified with HER2-positive cancer, optionally HER2-3+, HER2-2+, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing, and from one or more biological samples previously identified with HER2-negative cancer, optionally HER2-0 cancer based on IHC testing, one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation levels that are statistically significantly different. In some embodiments, differentially expressed sites are determined by comparing genomic features of one or more histone modifications and / or DNA methylation in (i) one or more HER2-positive cell lines and (ii) one or more HER2-negative cell lines. In some embodiments, the HER2 classifier targets sites identified by... Computer simulation Training was performed on differentially expressed sites, including histone modifications, chromatin accessibility, transcription factor binding, and / or DNA methylation levels, obtained by mixing sequence data from one or more HER2-positive cell lines and sequence data from liquid biopsy samples from healthy subjects.

[0073] In some embodiments, the HER2 classifier is trained against two or more histone modification levels at differentially expressed sites. In some embodiments, the HER2 classifier is trained against one or more histone modification levels and DNA methylation levels at differentially expressed sites. In some embodiments, the genomic characteristics of the subjects used for classification include two or more histone modification levels. In some embodiments, the two or more histone modification levels include H3K4me3 and H3K27ac modification levels. In some embodiments, the genomic characteristics of the subjects used for classification include one or more histone modification levels and DNA methylation levels. In some embodiments, one or more histone modification levels include H3K4me3 and / or H3K27ac modification levels. In some embodiments, one or more histone modification levels include H3K4me3 and H3K27ac modification levels.

[0074] In some implementations, the HER2 classifier has been trained using data derived from plasma. In other implementations, the HER2 classifier has been trained using data derived from liquid biopsy samples.

[0075] In some implementations, the HER2 classifier has been tuned using plasma data. In some implementations, the HER2 classifier has been trained using one or more genomic features, including (i) one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation of one or more HER2-positive and one or more HER2-negative cell lines. In some implementations, the HER2 classifier has been tuned using a transfer learning process comprising: (i) calculating predicted values ​​for plasma samples in the form of odds values ​​using the HER2 classifier (e.g., using the formula log2(HER2+ probability / 1 – HER2+ probability)), the HER2 classifier being trained using genomic features of one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation in one or more HER2-positive and one or more HER2-negative cell lines; and (ii) training a new model (e.g., a lasso logistic regression model) using the odds values ​​as offsets, using all the same features but trained with plasma data (e.g., using leave-one-out method) to determine new weights and coefficients, and then adding these weights and coefficients to coefficients determined during cell line training to obtain the tuned model. In some implementations, new models have been trained for cancer-specific genomic loci in plasma data, where cancer-specific genomic loci are regions in HER2-positive subjects that are enriched with H3K4me3 and H3K27ac modifications compared to HER2-negative subjects and that are associated with ctDNA at HER2 loci.

[0076] In some implementations, the HER2 classifier is a validated classifier. In some implementations, the HER2 classifier is validated by selecting a threshold such that the validated classifier predicts HER2-positive cancer, optionally based on IHC testing for HER2-3+, HER2-2+, or HER2-1+ cancer, or based on IHC / ISH testing for HER2-low cancer, with an area under the recipient operating characteristic curve (AUROC) greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95). In some implementations, the HER2 classifier has been validated against independent cohorts of subjects with HER2-positive and HER2-negative cancers, where subjects falling into the predicted HER2-positive cancer group exhibit validated epigenetic characteristics, and subjects not falling into the HER2-positive cancer group lack validated epigenetic characteristics. In some implementations, the HER2 classifier has been validated using liquid biopsy sample data.

[0077] A non-transient computer-readable storage medium may encode a computer program, wherein the program may contain instructions that, when executed by one or more processors, cause the one or more processors to perform operations to perform a method for determining the HER2 status of cancer in a subject (e.g., a patient). A computer system may include memory and one or more processors coupled to the memory, wherein the one or more processors are configured to perform operations to perform a method for determining the HER2 status of cancer in a subject (e.g., a patient).

[0078] In some embodiments, a method of treating a subject with cancer includes administering a HER2-targeting agent to the subject, wherein the subject has been determined to have validated epigenetic characteristics indicative of HER2-positive cancer based on analysis of cell-free DNA (cfDNA) from a biological sample obtained from or derived from the subject, optionally from a liquid biopsy sample. In some embodiments, the presence of validated epigenetic characteristics has been determined using a classifier (e.g., a validated classifier) ​​according to a method used to determine the HER2 status of a subject's (e.g., a patient's) cancer. Attached Figure Description

[0079] Figure 1ROC curves for exemplary HER2 state classifiers generated according to Example 2 are shown. As shown, different classifiers were generated using genomic loci from Tables 1 to 3 for different modifications, namely (i) H3K4me3 modification, (ii) H3K27ac modification, (iii) DNA methylation (DNAme), or (iv) all of the above. For specific modifications, different classifiers were also generated using different subsets of genomic loci from Tables 1 to 3, namely (i) all genomic loci with an absolute log2 (fold change) ≥ 0.5, (ii) all genomic loci with an absolute log2 (fold change) ≥ 1, (iii) all genomic loci with an absolute log2 (fold change) ≥ 2, (iv) all genomic loci with an absolute log2 (fold change) ≥ 3, and (v) all genomic loci with an absolute log2 (fold change) ≥ 4.

[0080] Figure 2 Representative, non-limiting graphs are shown to demonstrate the accuracy of HER2 states (based on AUCROC) determined using a classifier generated according to Example 2.

[0081] Figure 3 ROC curves of an exemplary HER2 state classifier generated according to Example 3 are shown. As shown, different classifiers were generated using the following methods: (i) Ridge regression (all genomic loci from a subset of relevant genomic loci) (α = 0); (ii) Elastic network regression (many genomic loci from a subset of relevant genomic loci) (α = 0.25); or (iii) Lasso (minimum absolute shrinkage and selection operator) regression (a few genomic loci from a subset of relevant genomic loci) (α = 1). As shown, different classifiers were generated using the following: (a) genomic loci in Tables 1 to 3 for different modifications, namely (i) H3K4me3 modification, (ii) H3K27ac modification, (iii) DNA methylation (DNAme), or (iv) all of the above modifications; and (b) different subsets of genomic loci in Tables 1 to 3 for specific modifications, namely (i) all genomic loci with an absolute value log2 (fold change) ≥ 0.5, (ii) all genomic loci with an absolute value log2 (fold change) ≥ 1, (iii) all genomic loci with an absolute value log2 (fold change) ≥ 2, (iv) all genomic loci with an absolute value log2 (fold change) ≥ 3, and (v) all genomic loci with an absolute value log2 (fold change) ≥ 4.

[0082] Figure 4Representative, non-limiting graphs are shown to demonstrate the accuracy of HER2 states (based on AUCROC) determined using a classifier generated according to Example 3.

[0083] Figure 5 A general overview of an exemplary epigenomic platform is provided, which can be used to dynamically resolve target and pathway biology using small amounts of biological samples containing cfDNA (cell-free DNA) (e.g., approximately 1 mL of plasma). Tumor-derived cfDNA exists in circulation as chromatin fragments, faithfully maintaining tumor-associated epigenetic modifications on histones and DNA. Binding agents (e.g., antibodies) that bind to modifications associated with active enhancers (e.g., H3K27ac), active promoters (e.g., H3K4me3), and DNA methylation can be used to enrich and sequence relevant DNA fragments from small biological samples (e.g., approximately 1 mL of plasma) to provide a whole-genome epigenomic map, thereby capturing the potential transcriptional state of tumor cells (Baca et al., Nature Medicine (2023) 29:2737-2741).

[0084] Figure 6 The diagram illustrates the correlation between ERBB2 (HER2) expression in breast cancer, as determined by RNA-seq, and ERBB2 regulatory signals measured in matched plasma samples. (A) A comparison of gene regulatory signals (obtained from plasma) and RNA-seq (obtained from tissue) is shown. The gene regulatory signals are correlated with expression. (B) Quantification of corrected gene regulatory signals at ctDNA sites at ERBB2 loci in plasma obtained from subjects with breast cancer. Loci with robust regulatory signals were selected using a leave-one-out (LOO) method and further screened for ERBB2 gene specificity, as demonstrated in the literature. These quantification results were used to classify HER2 IHC 3+ / 2+ISH+ and 2+ / 1+ / 0, with a LOO AUC of 0.81, where ctDNA was detectable in the samples (as determined by ichorCNA). (C) In plasma samples with detectable ctDNA, numerous sites in promoters, enhancers, and DNA methylation were observed that showed statistically significant signal differences between HER2 IHC 3+ / 2+ISH+ and 2+ / 1+ / 0 subjects, suggesting the potential of regions outside the HER2 amplicon for HER2 status classification. (D) All promoters and enhancers found to contain differential epigenetic modifications according to HER2 status were linked to their nearest genes. Many of these genes have been documented to support their involvement in HER2 signaling (exemplary genes), suggesting that the model provided in this disclosure can reflect the underlying biology of cancer.

[0085] Figure 7 An exemplary method for developing epigenomic and genome-wide HER2+ vs. HER2- breast cancer classifiers using information from cell lines and plasma samples is illustrated. Epigenomic data from 26 breast cancer cell lines (8 HER2IHC 3+ and 18 HER2IHC 0) were used to train a lasso logistic regression model. Sites with robust regulatory signals in subjects with breast cancer were selected using the LOO protocol, and the prediction for retained samples yielded a classification AUC of 0.81 (HER2IHC 3+ / 2+ISH+ vs. 2+ / 1+ / 0). The cell line model was then adjusted using regulatory signal information from breast cancer plasma samples, achieving an AUC of 0.9 in samples with detectable ctDNA (ichorCNA). An AUC of 0.86 was also observed in all samples, indicating that some samples with a LoD below ichorCNA but still containing ctDNA were still correctly classified. The table indicates exemplary numbers of enhancers, promoters, and methylated DNA regions in the HER2 amplicon and extra-HER2 amplicon regions in classifiers trained using cell line data and classifiers adjusted using plasma samples. Figure 7 As shown, the cell line-trained classifier described in this paper can accurately classify subjects as HER2-positive or HER2-negative. Adjustments using plasma data further improve the accuracy of the classifier.

[0086] Figure 8The paper demonstrates that the breast cancer classifier presented herein achieves an AUC of 0.83 at low ctDNA percentages (e.g., ctDNA <3%). (A) shows that a LOO AUC of 0.86 was observed in all subjects with breast cancer, including those with a LoD of ctDNA below ichorCNA (HER2 IHC 3+ / 2+ISH+ vs. 2+ / 1+ / 0). (B) illustrates an exemplary protocol for characterizing the performance of the classifier presented herein in samples with low ctDNA levels (e.g., samples with a LoD of ctDNA below ichorCNA). To characterize the performance of the classifier with low ctDNA levels, breast cancer samples (N=44) with ctDNA ≥ 10% were diluted using computer simulation with plasma data from healthy controls (N=12). Each patient sample was diluted to each of two healthy control samples, resulting in eight different simulated ctDNA fractions with a final ctDNA level of 1–3%, matched to sequencing depth. (C) As a control, fragments were extracted from the same patient dataset to match the sequencing depth, and real-world samples were found to perform better than the simulated dataset (indicating that the classifier's actual performance at low ctDNA levels is actually better than the performance estimated by the computer simulation method shown in (B)). (D) A plasma-adjusted HER2 classifier was used to predict the IHC status of each sample, simulating ctDNA% of 1–3%. The mean predicted value for a given sample at a healthy donor dilution with a given ctDNA% was calculated, and a LOO AUC of 0.83 was observed. Since the simulation program appears to produce conservative classifier results, the provided AUC is expected to be the lower limit of actual performance within the specified ctDNA range (hence the "AUC ≥" note).

[0087] Figure 9The following demonstrates that HER2 classification via epigenomic liquid biopsy, using, for example, the classifier described herein, can track changes in HER2 status over time (e.g., across different treatment regimens). (A) shows the probability of a sample being HER2+ as measured by the epigenomic plasma classifier (HER2 classification probability (%)), and HER2 mRNA expression in matched tissue samples obtained from a patient whose HER2 status changed between two time points, as detected by IHC. As shown, the results from IHC, RNA-seq, and the plasma-based classifier provide consistent results, demonstrating that the classifier described herein can track changes in HER2 status over time (e.g., post-treatment). (B) shows that HER2 classification via epigenomic liquid biopsy can potentially track status changes in treatment regimens. The following shows the HER2 IHC score, the HER2+ IHC probability measured by the epigenomic plasma classifier as described herein, and measurements of HER2 mRNA expression in matched tissues from a patient (data available for more than one time point). For this patient, consistent changes were observed in plasma, IHC, and RNA-seq. Notably, a decrease in HER2 model scores determined using epigenomic liquid biopsy was observed after initiation of anti-HER2 therapy (trastuzumab-drutecan).

[0088] Figure 10 The generation of HER2 classifiers for indications other than breast cancer (particularly gastroesophageal adenocarcinoma (GEA) and ovarian cancer (OV)) is illustrated. The HER2 breast cancer classifier described herein is used to initiate the training of HER2 classifiers for other indications. (A) shows the HER2 classifier results (3+ / 2+ISH+ vs 2+ / 1+ / 0) from plasma samples obtained from subjects with GEA and OV. The breast cancer cell line classifier was adjusted in the LOO protocol using GEA plasma samples to generate a GEA HER2 status classifier. An AUC of 0.96 was observed in samples with detectable ctDNA (ichorCNA). For OV, a classification AUC of 1 was observed. These data suggest that the method presented herein can be used to determine HER2 status and / or can be easily adjusted to determine HER2 in a variety of cancers (e.g., cancers previously shown to have HER2 and / or HER2 amplicon upregulation).

[0089] Figure 11The results show a correlation between pan-cancer HER2 classification, as determined by epigenomic lipid biopsy, and HER2 IHC. A linear trend was observed between model probabilities and HER2 IHC status when HER2 IHC predicted values ​​(3+ / 2+ISH+ vs. 2+ / 1+ / 0) were pooled across different indications. Surprisingly, although the classifier used to determine HER2 status was trained on plasma samples from 3+ / 2+ISH+ subjects, a statistically significant separation of model probabilities was observed between IHC 3+ samples and samples with other IHC scores. These results suggest the potential for patient stratification for therapies requiring HER2 IHC 3+ for inclusion.

[0090] Figure 12 This demonstrates that plasma-based epigenomic classifiers for HER2 status have the potential to be applied to other indications with evidence of HER2 activity (beyond breast cancer, GEA, and OV). Trastuzumab-drutecan recently expanded its labeling to include all HER2-positive (IHC 3+) solid tumors, enhancing treatment options across various solid tumor types. The demand for tissue samples will intensify with the expansion of tumor-nonspecific therapies, particularly those requiring immunohistochemistry (IHC) for biomarker identification. Comprehensive multianalyte blood assays capable of providing whole-genome expression (e.g., as provided in this disclosure) will enable the identification of biomarkers currently reliant on tissue-based analyses using minimally invasive, multi-strategy approaches. RNA-seq expression data from the TCGA Pan-Cancer Atlas were analyzed across various tumor types to identify potential disease candidates who may benefit from HER2-targeted therapies. The results are shown. In some embodiments, the methods provided in this disclosure can be used to treat… Figure 12 The following lists HER2+ cancers. The red boxes correspond to cancers that have been identified as HER2+ using the HER2 classifier described in this article.

[0091] Figure 13 An exemplary method for developing a HER2 classifier for breast cancer using cell lines and plasma samples obtained from patients is illustrated. Sites with differentially expressed epigenetic modifications (DNAme, H3K27ac, and H3K4me) in HER2+ and HER2- cell lines were identified, and a HER2 status classifier was constructed using this data. The cell line classifier was then adjusted using plasma sample data obtained from subjects with HER2+ or HER2- breast cancer. The “final classifier” refers to the plasma sample-adjusted classifier. Exemplary weights and numbers of features that can be used in the classifier are shown. As illustrated, the classifier produced using this scheme is capable of accurately classifying the HER2 status of subjects with breast cancer.

[0092] Figure 14 This demonstrates that the multianalyte HER2 classifier can robustly stratify patients based on HER2 status. The HER2+ probabilities of 3+, 1+ / 2+, and 0 (as determined by IHC) determined using the classifier constructed using the method described herein are shown. As shown, using the classifier provided herein, a HER2-0 sample exhibits a high HER2+ probability. Consistent with the HER2 classifier, FISH data show high HER2 amplification, indicating that the status determined by IHC is incorrect. Surprisingly, the HER2 classifier results are consistent with the FISH data and accurately classify the subjects, suggesting that, in some implementations, the method presented herein can produce more accurate results than the current gold standard IHC method used to determine HER2 status.

[0093] Figure 15 This paper demonstrates the potential of the HER2 classifier described herein to determine the HER2 status of colorectal cancer (CRC). (A) shows the promoter signal at the ERBB2 site in samples obtained from subjects with CRC. As shown, variable promoter signaling was observed, demonstrating the potential of the classifier described herein to detect epigenetic modifications associated with HER2 status in CRC, and the potential to develop a classifier to determine the HER2 status of CRC subjects. ACTB is shown as a positive control. As shown, variable HER2 signaling was observed even at low amounts of ctDNA (approximately 1–2%). (B) shows the results for a set of CRC samples (probability of sample being HER2+).

[0094] Figure 16 A list of cancers for which ERBB2 amplicones have been previously detected is provided. The Y-axis represents the number of previously detected amplicones. Cancers within boxes represent those for which anti-HER2 therapy has been approved. Figure 16 This demonstrates the potential of HER2-targeted therapies for treating a variety of cancers. In some implementations, the methods described herein can be used to treat cancers previously indicated by the detection of ERBB2 amplicon (e.g., Figure 16 (The cancers listed in the text).

[0095] Figure 16 This is a block diagram of an exemplary network environment for the methods and systems described herein, according to an illustrative embodiment of this disclosure.

[0096] Figure 17 This is a block diagram of an exemplary computing device and an exemplary mobile computing device used in illustrative embodiments of this disclosure. Detailed Implementation

[0097] This disclosure is based, at least in part, on the evidence that the HER2 status of a subject's cancer can be determined by detecting and quantifying the presence of histone modifications and / or DNA methylation at one or more genomic sites in cell-free DNA (cfDNA) from liquid biopsy samples (e.g., plasma samples obtained from or derived from a subject). This disclosure also covers methods for detecting chromatin accessibility and / or binding of one or more transcription factors at one or more genomic sites, rather than (or in addition to) histone modifications and / or DNA methylation. This disclosure is also based, at least in part, on the evidence that genomic sites differentially modified based on different types of histone modifications (e.g., histone methylation markers such as H3K4me3 and histone acetylation markers such as H3K27ac) and / or DNA methylation can be combined into multimodal classifiers to determine HER2 status. These novel unimodal and multimodal classifiers provide minimally invasive methods for determining HER2 status that are more accurate, objective, and comprehensive than current tissue-based methods. To date, no liquid biopsy platform has provided an operational solution for therapy-related transcriptional regulatory phenotypes such as HER2 status. The importance of HER2 intermediate expression states (e.g., low HER2) for HER2-targeting agents is becoming increasingly apparent, further highlighting this need; however, current tissue-based HER2 IHC / ISH methods struggle to objectively determine these states.

[0098] HER2 status and cancer The human epidermal growth factor receptor (HER) receptor family plays a central role in the pathogenesis of several human cancers. They regulate cell growth, survival, and differentiation via multiple signal transduction pathways and participate in cell proliferation and differentiation. The family consists of four main members: HER1, HER2, HER3, and HER4, also known as ErbB1, ErbB2, ErbB3, and ErbB4, respectively (Riese and Stern, BioEssays (1998) 20:41-48). All four HER receptors contain a cysteine-rich extracellular ligand-binding site, a transmembrane lipophilic region, and an intracellular domain with tyrosine kinase catalytic activity. The neu oncogene (also known as HER2, ErbB2, or p185) was discovered in the 1980s (Padhy et al., Cell (1982) 28(4):865-871 and Schechter et al., Nature (1984) 312(5994):513-516). The HER2 receptor is a 185 kD transmembrane glycoprotein of 1,255 amino acids located on the long arm (17q12) of human chromosome 17 (Brandt-Rauf et al., Crit Rev Oncogen (1994) 5(2-3):313-329). HER2 is expressed in many tissues, and its main role in these tissues is to promote excessive / uncontrolled cell growth and tumorigenesis.

[0099] Dimerization of the HER2 receptor leads to autophosphorylation of tyrosine residues within the receptor's cytoplasmic domain, initiating multiple signaling pathways that result in cell proliferation and tumorigenesis. HER2 amplification or overexpression occurs in approximately 15-30% of breast cancers and 10-30% of gastric / gastroesophageal cancers. These cancers are termed HER2-positive breast cancers, which tend to be more aggressive in growth and spread compared to HER2-negative breast cancers and gastric / gastroesophageal cancers. HER2 overexpression has also been found in other cancers, such as ovarian cancer, endometrial cancer, bladder cancer, lung cancer, colon cancer, and head and neck cancer.

[0100] To determine whether cancer is HER2-positive, medical professionals currently require tissue samples to be tested. Two types of tests are currently approved for HER2 diagnosis: immunohistochemistry (IHC) and in situ hybridization (ISH). IHC is usually performed first because ISH is more expensive and takes longer to provide results.

[0101] IHC samples are scored by pathologists as 0, 1+, 2+, or 3+. If the IHC result is 0, the cancer is considered HER2-negative. These cancers do not respond to treatment with drugs targeting HER2 and can instead be treated with one or more of surgery and / or radiation, endocrine therapy (if hormone receptors such as estrogen receptors are positive), chemotherapy, and immunotherapy. If the IHC result is 1+, the cancer is considered HER2-negative. These cancers typically do not respond to treatment with drugs targeting HER2, but have recently shown responses to some HER2-targeting agents (e.g., see Nicolò et al., TherAdv Med Oncol (2023) 15:1-16). If the IHC result is 2+, the HER2 status of the tumor is indeterminate and is referred to as “indeterminate.” This means that an ISH test is needed to determine the HER2 status. If the IHC result is 3+, the cancer is HER2-positive. These cancers are typically treated with agents targeting HER2.

[0102] An IHC result of 1+ or 2+ with no detectable abnormalities based on IHC testing. ERRB2 Amputated cancers are also often referred to as HER2-low cancers. These cancers have recently been shown to respond to certain HER2-targeting agents, such as trastuzumab-drutecan, which was recently approved by the FDA for the treatment of patients with HER2-low metastatic breast cancer (see Nicolò et al., Ther Adv Med Oncol (2023) 15:1-16).

[0103] HER2-targeted agents The introduction of HER2-targeted therapies has significantly impacted the prognosis of patients with HER2-positive breast cancer and gastric / esophageal cancer, and more recently, the prognosis of patients with HER2-positive colorectal and lung cancer. Many HER2-targeted therapies are under development and being tested in clinical trials for these and other HER2-positive cancers. These therapies include CAR-T, CAR-NK, and CAR-M therapies, as well as cancer vaccines (see, for example, Vila et al., Cancers (Basel) (2023) 15(7):1987, the entire contents of which are incorporated herein by reference).

[0104] Antibody Trastuzumab is a HER2-targeting antibody used to treat both early and advanced HER2-positive breast cancer. This treatment is typically administered in conjunction with chemotherapy, but can also be used alone (especially if chemotherapy alone has already been tried). When started before surgery (neoadjuvant therapy) or after surgery (adjuvant therapy) in early breast cancer, treatment usually lasts 6 months to a year. For advanced breast cancer, treatment typically continues until the treatment is no longer effective. This treatment is administered intravenously. Another form of trastuzumab, called trastuzumab plus hyaluronidase injection, is administered subcutaneously.

[0105] Pertuzumab is another HER2-targeting antibody that can be administered in combination with trastuzumab and chemotherapy to treat early-stage breast cancer before or after surgery, or to treat advanced-stage breast cancer. This treatment is administered intravenously.

[0106] Trastuzumab, pertuzumab, and hyaluronidase injection is a combination of these therapeutic agents, administered subcutaneously.

[0107] Magituximab is another HER2-targeting antibody that can be administered in conjunction with chemotherapy to treat advanced breast cancer, typically after at least two other HER2-targeting therapies have been tried. This therapy is administered intravenously.

[0108] Antibody-drug conjugates Antibody-drug conjugates (ADCs) consist of antibodies or antibody fragments linked to chemotherapeutic agents. In this case, the HER2-targeting antibody acts as a homing signal, delivering the chemotherapeutic agent directly to the cancer cells by attaching to the HER2 protein on the cancer cells.

[0109] Trastuzumab-Mettansin is an ADC that contains trastuzumab, an antibody targeting HER2, conjugated to the chemotherapy agent mettansin, which is similar to paclitaxel. It can be used alone to treat early-stage breast cancer after surgery (where chemotherapy and trastuzumab have been administered prior to surgery, and cancer cells are still present at the time of surgery), or to treat advanced breast cancer in women who have received both trastuzumab and chemotherapy. This treatment is administered intravenously.

[0110] Trastuzumab-drutecan is an adjuvant therapy (ADC) that comprises trastuzumab, a HER2-targeting antibody conjugated to the chemotherapy agent derutecan, which is similar to irinotecan. It can be used alone to treat breast cancer that cannot be surgically removed or has spread (metastasized) to other parts of the body, typically after at least one other HER2-targeting agent has been tried. This treatment is administered intravenously. Trastuzumab-drutecan can also be used to treat HER2-low-expressing breast cancer that cannot be surgically removed or has spread to other parts of the body, typically after a trial of chemotherapy or in cases of cancer recurrence within 6 months of completing adjuvant chemotherapy.

[0111] kinase inhibitors HER2 is a type of protein known as a kinase. Kinases are proteins in cells that are typically responsible for transmitting signals (such as telling cells to grow). Therapeutic agents that inhibit kinases are called kinase inhibitors and are usually small molecules. Several of these kinase inhibitors have been approved for the treatment of HER2-positive cancers.

[0112] Lapatinib is administered in tablet form and taken daily. It is used to treat advanced breast cancer. It is usually administered in combination with trastuzumab and the chemotherapy agent capecitabine.

[0113] Neratinib is administered in tablet form and taken daily. It is used to treat early-stage breast cancer in women who have received trastuzumab treatment for one year, and is typically given for one year. It can also be administered in combination with the chemotherapy agent capecitabine to treat people with metastatic disease, usually after trying at least two other HER2-targeting agents.

[0114] Tucatinib is administered in tablet form, usually twice a day. It is used to treat advanced breast cancer, provided that at least one other HER2-targeting agent has been tried. It is often administered in combination with trastuzumab and the chemotherapy agent capecitabine.

[0115] Other HER2-targeting agents and other cancers While the preceding sections focus on FDA-approved HER2-targeted therapies, many other HER2-targeted therapies are being developed and / or evaluated in clinical trials. These therapies include CAR-T, CAR-NK, and CAR-M therapies, as well as cancer vaccines (see, for example, Vila et al., Cancers (Basel) (2023) 15(7):1987, the entire contents of which are incorporated herein by reference). It should be understood that these other HER2-targeted therapies may also be used in the treatments described in this disclosure. Furthermore, while the preceding sections focus on the treatment of HER2-positive breast cancer, many of these HER2-targeted therapies may also be used to treat other HER2-positive cancers, such as gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc. For example, HER2-targeted therapies available for the treatment of HER2-positive colorectal cancer include, but are not limited to, trastuzumab, pertuzumab, tucatinib, lapatinib, and trastuzumab-drutecan. Of these HER2-targeted therapies, only the combination of tucatinib and trastuzumab has been approved by the FDA specifically for the treatment of colorectal cancer. However, because other HER2-targeted agents are approved for treating other types of HER2-positive cancers, such as breast cancer, doctors can prescribe them for colorectal cancer (off-label use). For advanced HER2-positive colorectal cancer that has already received chemotherapy, the most common HER2-targeted therapy regimens include trastuzumab-gatubacinib, lapatinib, or pertuzumab. Trastuzumab is FDA-approved for the treatment of metastatic HER2-positive gastric / esophageal cancer and is typically administered every 3 weeks along with chemotherapy. Trastuzumab-drutecan can also be used alone to treat advanced HER2-positive gastric / esophageal cancer, usually after an initial trial of trastuzumab. For gastric / esophageal cancer, trastuzumab-drutecan is typically administered every 3 weeks.

[0116] Subjects and samples The samples used in the methods, kits, and systems analyses provided herein can be any biological sample, including any processed sample containing circulating tumor DNA (ctDNA) derived from a biological sample. In various embodiments, the samples used in the methods, kits, and systems analyses provided herein can be samples obtained from mammalian subjects. In various embodiments, the samples used in the methods, kits, and systems analyses provided herein can be samples obtained from human subjects.

[0117] In various contexts, a human subject is someone who has been diagnosed with or is seeking a diagnosis of HER2-positive cancer (e.g., HER2-positive breast cancer, stomach / esophageal cancer, colorectal cancer, lung cancer), someone who has been diagnosed with or is seeking a diagnosis of being at risk of developing HER2-positive cancer, and / or someone who has been diagnosed with or is seeking a diagnosis of being at direct risk of developing HER2-positive cancer, etc. In various contexts, a human subject is someone identified as requiring HER2 status screening. In some cases, a human subject is someone identified by a medical practitioner as requiring HER2 status screening.

[0118] The subject may not have previously received cancer treatment, such as that described in this disclosure. In other embodiments, the subject has previously received cancer treatment, such as that described in this disclosure.

[0119] In various implementations, subjects possess one or more biomarkers and / or risk factors for cancer (e.g., HER2-positive cancer, such as HER2-positive breast cancer, gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc.). In some implementations, the requirement for HER2 status screening in human subjects is determined based on an initial cancer diagnosis (e.g., diagnosis of breast cancer, gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc.). In all cases, a human subject is defined as a subject who has not been diagnosed with cancer, has no risk of developing cancer, has no direct risk of developing cancer, has not been diagnosed with cancer, and / or has not sought a cancer diagnosis. Genetic factors may also increase the risk of HER2-positive cancer, as demonstrated by individuals with a family history of HER2-positive cancer.

[0120] In some embodiments, HER2-targeting agents may be used for lung cancer patients identified as HER2-positive based on the methods described herein. In some embodiments, the lung cancer is unresectable or metastatic non-small cell lung cancer (NSCLC) with an activated HER2 mutation in the tumor. In some embodiments, the HER2 mutation is an S310F, S310Y, R678Q, D769H, or I767M mutation, and the subject is administered an antibody or antibody-drug conjugate of this disclosure, such as trastuzumab-drutecan (see Gaibar et al., J Oncol (2020) 2020:6375956). In some embodiments, the HER2 mutation is an L755S or D769Y mutation, and the subject is administered a kinase inhibitor of this disclosure, such as neratinib (see Gaibar et al., J Oncol (2020) 2020:6375956).

[0121] In various embodiments, samples from a subject (e.g., a person) can be obtained from a liquid biopsy. In some embodiments, the sample and / or reference is obtained from serum, plasma, or urine. In some embodiments, the sample is serum. In some embodiments, the sample contains circulating tumor DNA (ctDNA). In some embodiments, the sample is derived from about 1 mL of blood obtained from the subject. In some embodiments, the sample is derived from about 0.5–5 mL of blood obtained from the subject, for example, about 0.5 to about 2 mL, about 0.5 to 1.75 mL, about 0.5 to 1.5 mL, about 0.75 to 1.25 mL, about 0.9 to 1.1 mL, about 1 mL, about 2 mL, about 3 mL, about 4 mL, or about 5 mL of blood.

[0122] In various implementations, the sample is a cell-free DNA (cfDNA) sample. cfDNA typically exists in human biological fluids (e.g., plasma, serum, or urine) as short double-stranded fragments. cfDNA concentrations are usually low, but can increase significantly under certain conditions, including but not limited to pregnancy, autoimmune diseases, myocardial infarction, and cancer. Circulating tumor DNA (ctDNA) is a component of cell-free DNA specifically derived from cancer cells. ctDNA may or may not be present in human biological fluids and may bind to leukocytes and erythrocytes. Various tests for detecting tumor-derived ctDNA are based on detecting genetic or epigenetic modifications of cancer characteristics (e.g., features of the associated cancer). Genetic or epigenetic characteristics of cancer include, but are not limited to, oncogenetic or cancer-related mutations in tumor suppressor genes, activated oncogenes, chromosomal abnormalities, histone modifications (e.g., histone methylation and / or histone acetylation), chromatin accessibility, binding to one or more transcription factors, and / or DNA methylation.

[0123] In various implementations, ctDNA accounts for less than 30%, less than 20%, or less than 10% of cfDNA in the liquid biopsy sample obtained from the subject, for example, less than 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, or less than 1% of cfDNA in the sample. In some implementations, the percentage of ctDNA in the liquid biopsy sample is assessed using ichorCNA, which estimates the percentage of ctDNA in the sample in a probabilistic manner (see Adalsteinsson et al., Nat Commun (2017) 8(1):1324, the entire contents of which are incorporated herein by reference).

[0124] cfDNA and ctDNA can provide real-time or near-real-time indicators of the state of the source tissue. The half-life of cfDNA and ctDNA in blood is approximately 2 hours, therefore samples collected at a given time reflect the state of the source tissue relatively promptly.

[0125] In some embodiments, the method includes isolating DNA (e.g., cfDNA) from a liquid biopsy sample. Various methods for isolating nucleic acids from a sample are known in the art (e.g., isolating cfDNA from blood or plasma). Nucleic acids can be isolated using, but not limited to, standard DNA purification techniques via direct gene capture (e.g., by clarifying the sample to remove assay inhibitors, and, if the target nucleic acid is present, capturing the target nucleic acid from the clarified sample with a capture agent to generate a capture complex, and separating the capture complex to recover the target nucleic acid).

[0126] Reagents and protocols for obtaining and analyzing cfDNA and ctDNA (such as circulating cfDNA and ctDNA in blood or other tissues) are commercially available, as described in the examples, and are well known in the art (see, for example, Anker et al., Cancer and Metastasis Rev (1999) 18:65-73; Wua et al., Clin Chim Acta (2002) 321:77-87; Fiegel et al., Cancer Res (2005) 15:1141-1145; Pathak et al., Clin Chem (2006) 52:1833-1842; Schwarzenbach et al., Clin Cancer Res (2009) 15:1032-1038; Schwarzenbach et al., Nat Rev Cancer (2011) 11:426-437, the contents of each of which are incorporated herein by reference in their entirety).

[0127] In various implementations, samples can be repeatedly collected from individuals over a period of time (e.g., daily, weekly, monthly, annually, semi-annually, etc.). In various implementations, such samples can be used to validate early detection results and / or identify changes in biological patterns, such as due to disease progression, resistance to therapy, treatment, remission, etc. For example, according to this disclosure, subject samples can be collected and monitored monthly, every two months, or in combinations of one, two, or three-month intervals. In various implementations, samples can be collected from or at certain clinically determined stages (such as resistance to therapy, before radiological progression, after radiological progression, and / or tissue biopsy) for monitoring over a period of time. Furthermore, HER2 status obtained at different time points can be conveniently compared with each other and with the status of normal controls during the monitoring period, thus providing the subject's own values ​​as an internal or personal control for long-term monitoring.

[0128] The samples include materials prepared by processes including, but not limited to, the following steps: concentration, dilution, pH adjustment, removal of high-abundance peptides (e.g., albumin, gamma globulin, and transferrin), addition of preservatives, addition of calibrators, addition of protease inhibitors, addition of denaturants, desalting, concentration, and / or extraction of sample nucleic acids, and / or amplification of sample nucleic acids (e.g., by PCR or other nucleic acid amplification techniques). The samples also include materials prepared by techniques for isolating, for example, nucleosomes or transcription factors and / or nucleic acids associated with nucleosomes or transcription factors.

[0129] Proteins unsuitable for the relevant purpose or background (e.g., high-abundance, non-informative, or undetectable proteins) can be removed from samples using high-affinity reagents, high-molecular-weight filters, ultracentrifugation, and / or electrodialysis. High-affinity reagents include antibodies or other reagents (e.g., aptamers) that selectively bind to high-abundance proteins. Sample preparation may also include ion-exchange chromatography, metal ion affinity chromatography, gel filtration, hydrophobic chromatography, chromatographic focusing, adsorption chromatography, isoelectric focusing, and related techniques. Molecular-weight filters include membranes that separate molecules based on size and molecular weight. Such filters can be further employed with reverse osmosis, nanofiltration, ultrafiltration, and microfiltration. Ultracentrifugation involves centrifuging the sample at approximately 15,000–60,000 rpm while monitoring particle settling (or non-settling) using an optical system. Electrodialysis is a procedure that uses electroporation or semipermeable membranes, involving the transfer of ions from one solution to another under the influence of a potential gradient. Because the membranes used in electrodialysis can selectively transport positively or negatively charged ions, repel ions with opposite charges, or allow substances to migrate through a semipermeable membrane based on size and charge, electrodialysis can be used for the concentration, removal, or separation of electrolytes.

[0130] The separation and purification methods disclosed herein may include any procedures known in the art, such as capillary electrophoresis (e.g., in a capillary or on a chip) or chromatography (e.g., in a capillary, column, or on a chip). Electrophoresis is a method that can be used to separate ionic molecules under the influence of an electric field. Electrophoresis can be performed in microchannels on a gel, capillary, or chip. Examples of gels used for electrophoresis include starch, acrylamide, polyethylene oxide, agarose, or combinations thereof. Gels can be modified by cross-linking, adding detergents or denaturing agents, immobilizing enzymes or antibodies (affinity electrophoresis) or substrates (zymography), and introducing pH gradients. Examples of capillaries used for electrophoresis include capillaries with electrospray interfaces.

[0131] Capillary electrophoresis (CE) is preferably used to separate complex hydrophilic molecules and highly charged solutes. CE technology can also be applied to microfluidic chips. Depending on the type of capillary and buffer used, CE can be further subdivided into separation techniques such as capillary zone electrophoresis (CZE), capillary isoelectric focusing (CIEF), capillary isovelocity electrophoresis (CITP), and capillary electrochromatography (CEC). One embodiment of coupling CE technology with electrospray ionization involves using a volatile solution, for example, an aqueous mixture containing volatile acids and / or bases and organic matter such as alcohols or acetonitrile.

[0132] Capillary isotachophoresis (CITP) is a technique in which analytes move through a capillary at a constant velocity but are still separated by their respective mobilities. Capillary zone electrophoresis (CZE), also known as free solution electrophoresis (FSCE), is based on the differences in electrophoretic mobilities of analytes, which depend on the analyte charge and the frictional resistance encountered during migration; this frictional resistance is typically proportional to the size of the analyte. Capillary isoelectric focusing (CIEF) allows weakly ionizable amphoteric molecules to be separated by electrophoresis over a pH gradient. CEC is a hybrid technique combining traditional high-performance liquid chromatography (HPLC) and capillary electrophoresis (CE).

[0133] The separation and purification techniques used in this disclosure may include any chromatographic procedure known in the art. Chromatography may be based on the differential adsorption and elution of certain analytes, or on the partition of the analyte between the mobile and stationary phases. Different examples of chromatography include, but are not limited to, liquid chromatography (LC), gas chromatography (GC), high-performance liquid chromatography (HPLC), etc.

[0134] In some implementations, whole blood is collected from the subject, and the plasma layer is separated by centrifugation. cfDNA can then be extracted from the plasma using methods known in the art.

[0135] Histone modification, chromatin accessibility and transcription factor binding Histone methylation is thought to increase or decrease the expression of related coding sequences, depending on which histone residues are methylated. Histone methylation is a necessary modification that leads to monomethylation (me1), dimethylation (me2), and trimethylation (me3) of several amino acids, directly affecting heterochromatin formation, gene imprinting, X chromosome inactivation, and gene transcription regulation. Histone methyltransferases promote monomethylation, dimethylation, or trimethylation of histones, while histone demethylases promote demethylation. Generally, lysine (Lys or K), arginine (Arg or R), and the rare histidine (His or H) are the most common histone methyl receptors. Histone methylation occurs only at specific lysine and arginine sites in histones H3 and H4. In histone H3, lysines 4, 9, 26, 27, 36, 56, and 79, and arginines 2, 8, and 17 can be methylated. In contrast, histone H4 has fewer methylation sites, with only lysines 5, 12, and 20, and arginine 3, capable of methylation. Histone methylation is generally associated with the activation or repression of transcription in downstream genes. Methylation of histones H3K4, R8, R17, K26, K36, K79, H4R3, and K12 can activate gene transcription. However, methylation of histones H3K9, K27, K56, H4K5, and K20 can repress gene transcription. For example, H3K4 methylation typically activates gene expression, while H3K27 methylation typically represses it.

[0136] Histone acetylation primarily occurs on lysine residues and is generally thought to increase the expression of related coding sequences. While not wanting to be bound by any theory, it is believed that acetylation of lysine residues neutralizes the positive charge of lysine, thus distancing histones from negatively charged DNA. The released structure facilitates the involvement of transcriptional mechanisms such as transcription factors and RNA polymerase II. Histone acetylation and deacetylation are typically catalyzed by histone acetyltransferases (HAT) and HDAC, respectively. Acetyl-CoA is both the source and a cofactor of acetylation. In regulatory regions, HAT can acetylate histones and recruit HAT-containing complexes to activate transcription. For example, H3K9ac and H3K27ac levels may be associated with promoter and enhancer activity. Furthermore, H3K27ac not only enhances the kinetics of transcriptional activation but also accelerates the transition of RNA polymerase II from the initiation to the elongation state.

[0137] Differential modifications at genomic sites (e.g., differences in histone methylation and / or histone acetylation) can refer to, be determined by, or detect as differences or changes in the modification status of one or more genomic sites between a first sample, disease, illness, or state and a second or reference sample, disease, illness, or state. Those skilled in the art will understand that a reference is typically generated by measuring using the same, similar, or comparable method as the non-reference measurement being compared.

[0138] Chromatin accessibility refers to the degree of physical contact between nuclear macromolecules and DNA, and depends in part on the occupancy and modification state of nucleosomes. Modified histones can regulate chromatin accessibility through various mechanisms, such as altering transcription factor (TF) binding through steric hindrance and modulating nucleosome affinity for active chromatin remodelers. The topological organization of nucleosomes in the genome is non-uniform: histones can be densely packed in both facultative and constitutive heterochromatin, and can be consumed at regulatory sites, including enhancers, insulators, and transcriptomes. Active regulatory elements of the genome are generally accessible.

[0139] Differences in accessibility to genomic loci can refer to, or be determined by, or detect by, a comparative difference or change in the modification status of one or more genomic loci between a first sample, disease, condition, or state and a second or reference sample, disease, condition, or state. Those skilled in the art will understand that a reference is typically generated by measuring using the same, similar, or comparable method as the non-reference measurement being compared.

[0140] Reference values ​​can be predetermined values ​​or a set of values, or values ​​or a set of values ​​derived from one or a set of samples. A reference can be one or a set of samples. Reference values ​​can be predetermined thresholds (values ​​that vary depending on the circumstances (e.g., based on patient subgroups, age, weight, or other variables)) or ratios. Reference ratios can be ratios related to modifications and / or accessibility at multiple sites within, across, or between individual samples and / or references. In various embodiments, references can have or represent a normal, non-disease state. In some embodiments, such as for disease staging or for assessing treatment efficacy, references can have or represent a disease state, such as cancer, cancer stage, or cancer subtype, such as HER2-positive or HER2-negative cancer. In some embodiments, references can represent HER2-3+, HER2-2+, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing. In some embodiments, references can be based on IHC testing to represent HER2-0 cancer. In some embodiments, the reference may correspond to a subject with breast cancer and / or a breast cancer subtype, such as HER2-positive or HER2-negative cancer. In some embodiments, the reference may correspond to a subject with gastric cancer / gastroesophageal cancer and / or a gastric cancer subtype / gastroesophageal cancer subtype, such as HER2-positive or HER2-negative gastroesophageal cancer. In some embodiments, the reference may correspond to a subject with colorectal cancer and / or a colorectal cancer subtype, such as HER2-positive or HER2-negative colorectal cancer.

[0141] In some embodiments, the reference is a predetermined threshold. In some embodiments, the predetermined threshold has been previously demonstrated to distinguish between HER2-positive and HER2-negative cancers (e.g., by an AUROC greater than 0.5). In some embodiments, the reference is a measurement from a liquid biopsy sample. In some embodiments, the reference is a measurement from a liquid biopsy sample obtained from a cohort of subjects. In some embodiments, the reference is a normalized sample. In some embodiments, the reference is a measurement obtained from a liquid biopsy sample obtained from a cohort of subjects previously diagnosed with HER2-positive or HER2-negative cancer (including, for example, HER2-positive or HER2-negative breast cancer).

[0142] In some cases, the reference is a non-contemporary sample from the same source, such as a previous sample from the same source, for example, a sample from the same subject. In some cases, the reference for the modification status of one or more genomic loci (e.g., one or more differentially modified genomic loci) can be the modification status of a sample (e.g., a sample from a subject) or multiple samples of one or more genomic loci (e.g., one or more differentially modified genomic loci) known to represent a specific state (e.g., HER2-positive or HER2-negative cancer). In some cases, the reference for the accessibility status of one or more genomic loci (e.g., one or more differentially accessible genomic loci) can be the accessibility status of a sample (e.g., a sample from a subject) or multiple samples of one or more genomic loci (e.g., one or more differentially accessible genomic loci) known to represent a specific state (e.g., HER2-positive or HER2-negative cancer).

[0143] In some exemplary but non-limiting embodiments of this disclosure, difference modification or difference accessibility may refer to a difference (e.g., the difference between a sample and a reference) where the absolute log2 (fold change) is greater than or equal to 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0 or higher, or any range (including extreme values) between the two, such as that measured according to the assays provided herein. In Tables 1 through 3, the log2 (fold change) values ​​are based on the ratio of HER2 positive to HER2 negative reads. Right now A positive log2 (fold change) value indicates that the sequencing read at a specific genomic locus is associated with a HER2-positive status, while a negative log2 (fold change) value indicates that the sequencing read at a specific genomic locus is associated with a HER2-negative status.

[0144] Enhancers are genomic sites that can undergo differential modification or exhibit differential accessibility in different diseases, conditions, and other states. Enhancers are cis-acting DNA regulatory regions that are thought to bind to trans-acting proteins, thereby influencing the expression patterns of related genes. Chromium immunoprecipitation sequencing (ChIP-seq) of histone modifications (e.g., acetylation) has identified millions of enhancers in the mammalian genome. The number of active enhancers in any given cell type is estimated to be in the tens of thousands. Certain transcription factors (TFs), sometimes referred to as “master” transcription factors, are associated with active enhancers and have significant effects on gene expression and cellular function. Some of these transcription factors preferentially associate with enhancers that regulate genes required to establish cellular identity and function, including enhancer domains known as “superenhancers.” Furthermore, master TFs can participate in interconnected self-regulating circuits or “cliques” that are self-enhancing, exhibit significant cell selectivity, and play a role in maintaining cell state and / or cell survival.

[0145] Techniques for detecting and quantifying histone modifications and transcription factor binding Various molecular biology techniques are well known in the art and / or disclosed in this application for the detection and quantification of histone modifications and / or transcription factor binding. In some embodiments, the methods, kits, and systems disclosed herein relate to the detection and quantification of histone modifications and / or transcription factor binding in samples (e.g., in liquid biopsy samples containing cfDNA, such as plasma samples containing cfDNA). Chromatin immunoprecipitation (ChIP) is a technique in molecular biology that can be used to detect and quantify histone modifications and transcription factor binding in samples. CUT&RUN or CUT&Tag are other newer techniques that can also be used to detect and quantify histone modifications and transcription factor binding sites. ChIP-chip, ChIP-exo, ChIP Re-ChIP, and ChIPmentation are other alternative techniques that can be used.

[0146] ChIP can involve multiple steps, including one or more of fixation, sonication, immunoprecipitation, and analysis of the immunoprecipitated DNA. ChIP has become a widely used tissue-based technique for determining the in vivo locations of various transcription factors and histone binding sites. Because proteins are captured at their DNA-binding sites, ChIP facilitates the detection of DNA-protein interactions occurring in living cells. More importantly, ChIP can be coupled with many commonly used molecular biology techniques such as PCR and real-time PCR, single-strand conformation polymorphism PCR, Southern blot analysis, Western blot analysis, cloning, and microarrays. This resulting versatility enhances the potential of this technique.

[0147] ChIP of tissue samples typically involves cross-linking chromatin-binding proteins with formaldehyde, followed by sonication or nuclease treatment to obtain small DNA fragments. Immunoprecipitation can then be performed using specific antibodies against the DNA-binding proteins of interest. The DNA can then be released from the proteins and analyzed using various methods. ChIP is also used to study RNA-protein interactions. X-ChIP uses sonication to break down fixed chromatin, while N-ChIP uses native chromatin, which may not be fixed, and performs nuclease digestion.

[0148] The first step in this technology can be cross-linking DNA and proteins. Formaldehyde is one of the most commonly used cross-linking agents. One advantage of using formaldehyde is the ease of reversibility of the cross-linking and its ability to form bonds spanning approximately 2 angstroms. This means that formaldehyde can bind molecules tightly together. Typically, formaldehyde can be added to the culture medium in cell culture flasks or cell culture plates. It enters the cells through the cell membrane and cross-links proteins with chromatin. Formaldehyde is also used for fixing tumor tissue. Other cross-linking agents that have been used include chemicals such as methylene blue and acridine orange, cisplatin, dimethylarsenic acid, potassium chromate, and ultraviolet (UV) light and lasers.

[0149] The harvested chromatin can be subjected to one or more sonication cycles. This typically breaks down DNA into 100-500 bp fragments to precisely pinpoint the location of the DNA sequence of interest. An alternative to sonication is nuclease digestion of the chromatin, such as in the N-ChIP method. Chromatin purification can be achieved using cesium chloride (CsCl) gradient centrifugation.

[0150] Chromatin can be enriched by targeting histone modifications using agents that bind to specific histone modifications (e.g., immunoprecipitation using one or more antibodies that bind to target epitopes).

[0151] For example, antibodies used in ChIP can selectively bind to specific transcription factors or one or more specific histone modifications, such as one or more specific histone acetylation or histone methylation modifications. In some embodiments, the antibody used to bind to the target epitope can be a "pan-" antibody (e.g., a pan-acetylation antibody, a pan-methylation antibody, an antibody that binds to a set of histone modifications associated with increased transcriptional activation, and / or an antibody that binds to a set of histone modifications associated with increased transcriptional repression). An antibody targeting the protein of interest is bound to a protein-DNA complex, and the complex is then precipitated. Commonly used immunoprecipitants for separating antigen-antibody complexes from lysates include salmon sperm DNA-protein A-Sepharose®, protein G, magnetic beads, and other engineered immunoprecipitation systems known to those skilled in the art.

[0152] Immunoprecipitated DNA can be eluted. Once the DNA of interest is isolated, various detection and quantification methods can be used to study the isolated gene fragments. Commonly used methods include PCR, real-time PCR, groove blot hybridization, microarray technology, and deep sequencing or next-generation sequencing. ChIP-seq combines chromatin immunoprecipitation (ChIP) with massively parallel DNA sequencing to identify binding sites of DNA-related proteins. ChIP-seq can be used to map DNA-binding proteins across the entire genome, such as transcription factor binding sites and histone modification maps.

[0153] Cell-free chromatin immunoprecipitation sequencing (cfChIP-seq) involves applying ChIP-seq to samples containing cell-free DNA, such as liquid biopsy samples containing cfDNA, like plasma samples containing cfDNA (see, for example, Sadeh et al., Nat Biotechnol (2021) 39: 586–598 and Jang et al., Life Sci Alliance (2023) 6(12):e202302003, the entire contents of each of which are incorporated herein by reference). In some embodiments, cfChIP-seq uses antibodies or antibody fragments that bind to specific histone modifications (e.g., H3K4me3 and / or H3K27ac) and / or transcription factors, which are covalently or non-covalently coupled to beads (e.g., magnetic beads such as Dynabeads® beads) and incubated with a volume, such as about 1 mL, of thawed plasma obtained from the subject. For example, exemplary antibodies binding to H3K4me3 include PA5-27029 (available from Thermo Fisher Scientific, Waltham, MA) and C15410003 (available from Diagenode, Denville, NJ), and exemplary antibodies binding to H3K27ac include ab21623 or ab4729 (both available from Abcam, Cambridge, UK) and C15210016 (available from Diagenode, Denville, NJ).

[0154] In some embodiments, antibodies or antibody fragments may be covalently coupled to beads, such as epoxy resin beads. In some embodiments, antibodies or antibody fragments may be non-covalently coupled to magnetic beads, such as protein A or protein G beads, such as Dynabeads® protein A or Dynabeads® protein G beads. After washing, a cfDNA library is typically prepared from the captured cfDNA. Library construction can be performed on beads or after releasing the captured cfDNA by digestion of bound histones (e.g., using proteinase K). The cfDNA library is then sequenced to produce reads of the captured cfDNA sequence, for example, by next-generation sequencing (NGS) as known in the art. The reads are then analyzed, for example, by alignment and counting using standard bioinformatics techniques as known in the art. The cfChIP-seq bioinformatics workflow may include, for example, aligning the sequence reads to a reference genome using BWA or Bowtie2. The aligned sequences can be used to compare with the reference sequence to identify and quantify peaks. In some embodiments, sequencing data can be used to quantify histone modifications at a given genomic site. For example, in some embodiments, histone modifications (e.g., those with at least one nucleotide overlapping with a genomic site) can be quantified by counting the number of sequence reads falling within the genomic site. In some embodiments, non-unique and / or redundant sequence reads are discarded before quantifying histone modifications. In some embodiments, sequence reads falling into high-noise regions of the genome are ignored during histone modification quantification.

[0155] In some implementations, sequence reads are adjusted according to sequencing depth before counting. Adjustment based on sequencing depth may include, for example, normalizing the sequence read quantiles to a common reference distribution. In some implementations, sequence reads are adjusted according to ChIP quality before counting. In some implementations, sequence reads are normalized relative to aggregated counts of a set of regions (e.g., 1,000, 2,000, 3,000, 4,000, 5,000, 6,000, 7,000, 8,000, 9,000, 10,000, or more regions) that have previously been identified as having DNase hypersensitivity in most cell types. In some implementations, an estimate of the local background signal is subtracted from the sequence read count at each genomic locus.

[0156] CUT&Tag involves antibody-based target protein binding, such as transcription factors of interest or histone modifications, where chromatin cleavage and library preparation occur directly after antibody incubation (see Kaya-Okur et al., Nat Comm (2019) 10:1930). CUT&Tag assays utilize a Tn5 transposase fused to protein A, guiding the enzyme to an antibody bound to a target on chromatin. The Tn5 transposase is pre-loaded with a sequencing adaptor (generating an assembled pA-Tn5 adaptor transposon) for antibody-targeted fragmentation labeling. In a typical CUT&Tag assay, the sample is incubated with antibodies immobilized on magnetic beads coated with concanavalin A for easy subsequent washing. Cells can be incubated first with a primary antibody specific to the target protein of interest, followed by a secondary antibody. The sample can then be incubated with an assembled transposon consisting of protein A fused to a Tn5 transposase conjugated to an NGS adaptor. After incubation, unbound transposons can be washed away under strict conditions. Tn5 is a Mg 2+ Dependent enzyme, therefore Mg can be added. 2+ This activates the reaction, causing chromatin to be cleaved near the protein binding site, and simultaneously adding the NGS adaptor DNA sequence. Chromatin cleavage and library preparation can be completed in one step.

[0157] CUT&RUN is an epigenomic analysis strategy in which a micrococcal nuclease targets and cleaves an antibody to release a specific protein-DNA complex into a supernatant for paired-end DNA sequencing (see Skene and Henikoff, Elife (2017) 6:1-35, Skene et al., Nat Protoc (2018) 13:1006-1019). Because only the target fragment enters the solution and the vast majority of the DNA remains in solution, the background level of CUT&RUN is very low. In an exemplary CUT&RUN assay, the sample is incubated with an antibody or antibody fragment that binds to the target protein (e.g., a transcription factor of interest or a histone modification). The sample is then incubated with protein A-MNase, after which CaCl2 may be added to initiate the calcium-dependent nuclease activity of MNase, thereby cleaving the DNA surrounding the target protein. The protein A-MNase reaction can be quenched by adding chelating agents (EDTA and EGTA). The cleaved DNA fragment is then released, extracted, and used to construct sequencing libraries.

[0158] Those skilled in the art will understand that the DNA sequencing techniques applicable to the methods described herein include a sequencing step. Applicable DNA sequencing techniques include, for example, next-generation sequencing (NGS) methods. Additional steps required for preparing DNA for sequencing via a suitable sequencing method may be incorporated into the methods described herein. For example, in some embodiments, the methods described herein include attaching (e.g., ligating) a DNA adaptor to cfDNA. In some embodiments, the DNA adaptor may be attached before, during, or after histone modification enrichment. In some embodiments, the method includes amplifying cfDNA after attaching the DNA adaptor.

[0159] Techniques for detecting and quantifying chromatin accessibility Various molecular biology techniques are well known in the art and / or disclosed in this application for detecting and quantifying chromatin accessibility. In some embodiments, the methods, kits, and systems of this disclosure relate to detecting and quantifying chromatin accessibility in samples, for example, in liquid biopsy samples containing cfDNA (such as plasma samples containing cfDNA). ATAC-seq (transposon accessibility chromatin sequencing assay), NOMe-seq (nucleosome occupancy and methylome sequencing), FAIRE-seq (formaldehyde-assisted separation of regulatory elements sequencing), MNase-seq (micrococcal nuclease digestion sequencing), and DNase hypersensitivity assay are exemplary techniques in molecular biology that can be used to detect and quantify chromatin accessibility in samples. Sono-Seq is another alternative method that can be used (see Auerbach et al., Proc Natl Acad USA (2009) 106(35):14926-14931).

[0160] DNase hypersensitivity assays utilize the nonspecific DNA endonuclease deoxyribonuclease I (DNase I), which selectively digests accessible DNA regions. DNase I hypersensitive sites (DHS) identified by DNase-seq include open chromatin regulatory regions. A typical DNase hypersensitivity assay involves a first step in which the cell nucleus is isolated from the cell using a lysis buffer, and the nucleus is digested using DNase I. DNA fragment size is measured using gel electrophoresis to determine optimal digestion. After polishing to form blunt ends, biotinylated linkers can be ligated to the ends of the digested DNA, which can then be isolated. DNA with biotinylated linkers can be digested with the restriction endonuclease MmeI and captured by streptavidin-coated Dynabeads®, resulting in short tags to which second sequencing linkers can be ligated. The second linkers can be ligated and amplified to produce a library for sequencing. DNase-seq bioinformatics workflows may include, for example, aligning sequence reads to a reference genome using BWA or Bowtie2. The aligned sequence can be used to compare with a reference sequence to identify and quantify peaks.

[0161] MNase-seq uses micrococcal nuclease (MNase) to determine chromatin accessibility. This enzyme preferentially digests DNA that does not contain nucleosomes and is not bound to proteins. A typical MNase-seq assay may include a first step in which the cell nucleus is isolated from native or cross-linked chromatin, digested with MNase, and titrated. The in vivo formaldehyde cross-linking step is designed to capture the interaction between proteins and DNA. This cross-linking allows the bound protein to protect its associated DNA from MNase digestion. After cross-linking, the sample is digested with MNase, which can be specifically activated by adding Ca2+ to the buffer. Digestion can be prevented by a chelation reaction, in which case the sample is treated with RNase, the cross-linking is reversed, and the protein is digested from the chromatin. DNA can then be isolated via phenol-chloroform extraction. Uncut DNA is purified, and mononucleosome bands are separated and excised by gel electrophoresis. The isolated DNA can be amplified by adding adaptors to generate a library for sequencing. MNase-seq primarily sequences DNA regions bound to histones or other proteins. Therefore, it indirectly determines which DNA regions are accessible by directly identifying which regions bind to nucleosomes or proteins.

[0162] FAIRE-seq is a method for isolating nucleosome-deleted regions (NDRs) of DNA from chromatin. A typical FAIRE-seq assay may include a first step in which cells are fixed with formaldehyde, causing histones to crosslink with interacting DNA. The crosslinked chromatin is then sheared by sonication, yielding protein-free DNA and protein-crosslinked DNA fragments. Protein-free DNA can be separated using phenol-chloroform extraction: protein-crosslinked DNA remains in the organic phase, while protein-free DNA remains in the aqueous phase. Highly crosslinked DNA remains in the organic phase, while uncrosslinked DNA is pulled into the aqueous phase. The uncrosslinked DNA in the aqueous phase can then be amplified and sequenced. Reads enriched in the sequencing pool tend to have less nucleosome and transcription factor binding, thus inferring that they originate from accessible regions.

[0163] NOMe-seq is a method that uses M. CviPI methyltransferase to recognize nucleosome deletion regions (NDRs) in DNA. This methyltransferase methylates cytosine in GpC dinucleotides that are not protected by nucleosomes or other proteins. (The last sentence appears to be incomplete and possibly refers to a different topic.) m pG is different from GpC in the human genome. m GpCs are not naturally present in most cell types. m Levels can be compared to background signals and used for detection and quantification of NDR. A typical NOMe-seq protocol may include a step in which the sample is treated with M. cviPI and S-adenosylhomocysteine ​​(SAM) to methylate accessible GpC sites. The M. cviPI-treated DNA can be sonicated for sequencing of the DNA fragments. The DNA is then treated with bisulfite, using sodium bisulfite to convert unmethylated cytosine to uracil, while methylated cytosine remains unaffected. An adaptor is used to generate a library, which is then sequenced. Accessible chromatin is expected to have high levels of GpC. m and low levels of C m Therefore, NOMe-seq uses two independent methylation analyses to identify NDRs, which, as independent (but opposite) measurements, provide matching chromatin markers for each regulatory element.

[0164] ATAC-seq utilizes a highly active Tn5 transposase that preferentially cleaves accessible chromatin regions while simultaneously inserting adaptors into the fragmented regions (Buenrostro et al., Nat Methods (2013) 10(12):1213-1218, the full text of which is incorporated herein by reference). A typical ATAC-seq assay may include a first step in which the sample is incubated with a Tn5 transposase. The DNA can then be isolated and purified. The DNA fragmented and labeled by the Tn5 transposase can be purified, then amplified to generate a library and sequenced for analysis.

[0165] Techniques for detecting and quantifying DNA methylation Various molecular biology techniques are well known in the art and / or disclosed in this application for the detection and quantification of DNA methylation. In some embodiments, the methods, kits, and systems of this disclosure relate to the detection and quantification of chromatin accessibility in a sample, for example, in a liquid biopsy sample containing cfDNA (such as a plasma sample containing cfDNA). Bisulfite sequencing (BS-Seq), whole-genome bisulfite sequencing (WGBS), methylated DNA immunoprecipitation sequencing (MeDIP-seq), or methyl-CpG-binding domain sequencing (MBD-seq) are exemplary techniques in molecular biology that can be used to detect and quantify chromatin accessibility in a sample. Degenerate representative bisulfite sequencing (RRBS) is another alternative method that can be used (see Meissner et al., Nucleic Acids Res (2005) 33(18):5868-5877). Illumina Infinium arrays can also be used to detect and quantify DNA methylation.

[0166] DNA methylation generally refers to the methylation of the 5' position of cytosine (mC) by DNA methyltransferases (DNMTs). This is an important epigenetic modification in humans and many other species. In mammals, most DNA methylation occurs against a CpG dinucleotide background. DNA methylation is considered a repressive chromatin modification. Abnormal methylation can lead to a variety of diseases, including cancer (Robertson, Nat Rev Genet (2005) 6:597–610 and Bergman and Cedar, Nat Struct Mol Biol (2013) 20:274–281).

[0167] Bisulfite sequencing (BS-Seq) or whole-genome bisulfite sequencing (WGBS) is a well-established protocol for detecting methylated cytosine in genomic DNA. In this method, genomic DNA is treated with sodium bisulfite and then sequenced, providing single-base resolution of methylated cytosine in the genome. After bisulfite treatment, unmethylated cytosine is deamination to uracil, which is then converted to thymidine after sequencing. Simultaneously, methylated cytosine resists deamination and is read as cytosine. The location of methylated cytosine can then be determined by comparing the treated and untreated sequences.

[0168] In some embodiments, methylated DNA can be sequenced using methods that include enriching cfDNA containing methylated DNA. For example, enrichment can be achieved using agents that selectively bind to methylated DNA (e.g., antibodies in MeDIP-seq or methyl-CpG binding domains (MBD) in MBD-seq). In some embodiments, an agent that binds to methylated DNA (e.g., via covalent or non-covalent bonding) is attached to a physical support (e.g., beads, magnetic beads, agarose beads, or magnetic epoxy beads), wherein attachment can be performed before, during, or after incubation with the sample.

[0169] MeDIP-seq was first reported by Weber et al., Nat Genet (2005) 37:853–862. In a typical MeDIP-seq protocol, methylated DNA fragments are enriched using antibodies or antibody fragments that bind to 5-methylcytidine (5mC), and these fragments are then sequenced and analyzed. If 5mC-specific antibodies or antibody fragments are used, methylated DNA is isolated from genomic DNA via immunoprecipitation. Anti-5mC antibodies are incubated with fragmented genomic DNA and precipitated, followed by DNA purification and sequencing.

[0170] Methyl-CpG-binding domain sequencing (MBD-seq) is similar to MeDIP-seq, except that it uses methyl-binding domain (MBD) proteins instead of antibodies or antibody fragments to bind methylated DNA. In a typical MBD-seq protocol, genomic DNA is first sonicated and then incubated with a labeled MBD protein that binds to methylated cytosine. The protein-DNA complex is then precipitated using beads conjugated with antibodies specific to the MBD protein tag, followed by DNA purification and sequencing.

[0171] In some implementations, DNA methylation at a given genomic site can be quantified by sequencing the methylated DNA. For example, in some implementations, DNA methylation at a genomic site can be quantified by counting the number of sequence reads that overlap with the genomic site (e.g., containing at least one nucleotide that overlaps with the genomic site).

[0172] Those skilled in the art will understand that the DNA sequencing techniques applicable to the methods described herein include a sequencing step. Applicable DNA sequencing techniques include, for example, next-generation sequencing (NGS) methods. Additional steps required for preparing DNA for sequencing via a suitable sequencing method may be incorporated into the methods described herein. For example, in some embodiments, the methods described herein include attaching (e.g., ligating) a DNA adaptor to cfDNA. In some embodiments, the DNA adaptor may be attached before, during, or after histone modification enrichment.

[0173] Classifier In some embodiments, this disclosure provides methods for obtaining a classifier, for example, a valid classifier that can be used to determine HER2 status. In some embodiments, based on analysis of cell-free DNA (cfDNA) from a biological sample obtained from or derived from a subject, optionally from a liquid biopsy sample, the presence of a valid epigenetic characteristic indicating HER2-positive cancer is determined in the subject, wherein the presence of the valid epigenetic characteristic has been determined using a valid classifier.

[0174] For illustrative purposes and not limited thereto, in one exemplary embodiment of this disclosure, a validated classifier can be obtained in the following manner: (a) Identify genomic features of (i) one or more HER2-positive cell lines or (ii) biological samples obtained from a first group of subjects who have been previously identified as having HER2-positive cancer, optionally HER2-3+ cancer, HER2-2+ cancer or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing, including histone modifications, chromatin accessibility, binding of one or more transcription factors and / or DNA methylation. (b) Identify genomic features of (i) one or more HER2-negative cell lines or (ii) biological samples obtained from a second group of healthy subjects or subjects previously identified as having HER2-negative cancer, optionally based on IHC testing, including histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation. (c) Compare the genomic features identified in step (a) with those identified in step (b) to identify genomic sites (“difference sites”) that show statistical differences in histone modification, chromatin accessibility, transcription factor binding and / or DNA methylation levels. (d) Using histone modification, chromatin accessibility, transcription factor binding, and / or DNA methylation level training at the differential sites, a classifier is used to distinguish (i) samples from one or more HER2-positive cell lines or biological samples obtained from the first cohort, and (ii) samples from one or more HER2-negative cell lines or biological samples obtained from the second cohort, to identify samples with histone modification, chromatin accessibility, transcription factor binding, and / or DNA methylation level characteristics (“epigenetic features”) indicating that the sample may have been obtained from a HER2-positive cell line or from the first cohort; and (e) The validated classifier is obtained by validating the classifier in step (d) on a third cohort comprising independent, blinded subjects with HER2-positive and HER2-negative cancers, and a threshold is selected such that the validated classifier predicts HER2-positive cancers, optionally based on HER2-3+, HER2-2+, or HER2-1+ cancers based on IHC testing, or HER2-low cancers based on IHC / ISH testing, with an area under the recipient operating characteristic curve (AUROC) greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95), wherein subjects falling into the predicted HER2-positive cancer group exhibit the validated epigenetic feature, and subjects not falling into the HER2-positive cancer group lack the validated epigenetic feature.

[0175] Those skilled in the art will understand that other methods can be used to obtain classifiers, such as classifiers that can be used to determine the verification of HER2 status, and this disclosure is not limited to classifiers obtained according to this method.

[0176] Exemplary genomic sites This disclosure includes the identification of exemplary genomic loci that are differentially modified and / or differentially accessible in HER2-positive cancers (e.g., HER2-3+ cancers based on IHC testing) and HER2-negative cancers (e.g., HER2-0 cancers based on IHC testing). See Tables 1 through 3, which show the chromosomal coordinates of each genomic locus and its observed log2 (fold change) (HER2-positive / HER2-negative), values ​​indicating whether the modification of the locus is associated with HER2-positive status (log2(FC) > 0) or with HER2-negative status (log2(FC) < 0). Genomic loci are ordered based on their chromosomal coordinates, which are based on human genome version hg19.

[0177] This invention is not limited to methods using the exact same chromosome coordinates listed in Tables 1 through 3. This disclosure covers methods using any genomic locus and its subregions as listed in Tables 1 through 3. Right nowThe methods mentioned in this article concerning the detection and / or quantification of one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic loci listed in Tables 1 to 3 cover methods for detecting these markers at any location (including any subregion) within these genomic loci. For example, Table 2 lists chr1:1097716-1101022 as genomic loci for detecting and / or quantifying H3K27ac modifications, which covers methods for detecting and / or quantifying H3K27ac modifications at any location or subregion within chr1:1097716-1101022, such as methods for detecting and / or quantifying H3K27ac modifications within chr1:1097816-1100522, and so on. In some embodiments, the subregion may span at least 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1500, 2000, 2500, or at least 3000 consecutive base pairs located between the lower and upper coordinates of the genomic sites listed in Tables 1 to 3. In some embodiments, the subregion may span fewer than 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1500, 2000, 2500, or at least 3000 consecutive base pairs located between the lower and upper coordinates of the genomic sites listed in Tables 1 to 3. In some embodiments, the subregion may have the same center coordinates as the genomic sites listed in Tables 1 to 3. In some embodiments, the subregion may have different center coordinates than the genomic sites listed in Tables 1 to 3. It should also be understood that the lower / upper bound coordinates of the genomic loci in Tables 1 to 3 are approximate values, and this disclosure covers methods for expanding any one or more genomic loci by increasing the size of the genomic loci by 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, or up to 50% in one or both directions.

[0178] In some embodiments, the classifier is generated using a set of differentially modified and / or differentially accessible genomic loci associated with HER2-positive status and a set of differentially modified and / or differentially accessible loci associated with HER2-negative status. Sequence reads falling into each selected genomic locus are analyzed and counted, for example, as described herein, including examples. In some embodiments, the counts of genomic loci associated with HER2-positive status are aggregated, and the counts of genomic loci associated with HER2-negative status are aggregated. In some embodiments, the ratio of aggregated HER2-positive to HER2-negative counts is used to determine HER2 status. Other methods described herein and known in the art for generating and applying classifiers to determine HER2 status using genomic loci and associated sequencing data include, but are not limited to, methods using learned statistical classifier systems or combinations of learned statistical classifier systems.

[0179] In some embodiments, exemplary genomic loci from Tables 1, 2, or 3 are used for a unimodal classifier, for example, using a single histone modification (e.g., H3K4me3 or H3K27ac) or DNA methylation at one or more genomic loci to determine the HER2 state. In some embodiments, exemplary genomic loci from Tables 1, 2, or 3 are combined for a multimodal classifier, for example, using more than one histone modification (e.g., H3K4me3 and H3K27ac) or one or more histone modifications (e.g., H3K4me3 and / or H3K27ac) and DNA methylation at one or more genomic loci to determine the HER2 state.

[0180] Difference H3K4me3 modification Table 1 provides genomic loci exhibiting differential H3K4 methylation (particularly H3K4 trimethylation, H3K4me3) in HER2-positive and HER2-negative cancers, showing the chromosomal coordinates of each locus and its observed log2 (fold change) (HER2-positive / HER2-negative). Genomic loci are ordered based on their chromosomal coordinates, which are based on the human genome version hg19.

[0181] Those skilled in the art will understand that the methods disclosed herein do not require evaluating H3K4me3 modifications for every genomic locus listed in Table 1. Instead, H3K4me3 modifications for a subset of loci can be evaluated. A subset of genomic loci in Table 1 can be selected based on various performance criteria (e.g., for determining HER2 status), such as selecting genomic loci exhibiting differential modifications at a specific statistical significance level and / or a specific difference threshold (e.g., measured log2 (fold change)) between relevant statuses. A subset of genomic loci can also be selected based on algorithms, such as during the process of obtaining a classifier. Those skilled in the art will understand that such subsets of loci in Table 1, and the loci included in such subsets, whether present individually or in randomly selected subsets, have at least equivalent information content (e.g., statistical significance and / or reliability) for the purposes disclosed herein (e.g., for determining HER2 status). See also the embodiments of this disclosure illustrating experiments demonstrating that information-rich classifiers can be produced using many different combinations of loci. Among other things, this disclosure specifically includes subsets of genomic loci listed in Table 1, which have absolute log2 (fold change) values ​​of 6.0 or higher, 5.5 or higher, 5.0 or higher, 4.5 or higher, 4.0 or higher, 3.5 or higher, 3.0 or higher, 2.5 or higher, 2.0 or higher, 1.9 or higher, 1.8 or higher, 1.7 or higher, 1.6 or higher, 1.5 or higher, 1.4 or higher, 1.3 or higher, 1.2 or higher, 1.1 or higher, 1.0 or higher, 0.9 or higher, 0.8 or higher, 0.7 or higher, 0.6 or higher, or 0.5 or higher. This disclosure also includes subsets of genomic loci listed in Table 1, wherein the absolute log2 (fold change) of these subsets is 6.0 or higher, 5.5 to less than 6.0, 5.0 to less than 5.5, 4.5 to less than 5.0, 4.0 to less than 4.5, 3.8 to less than 4.0, 3.6 to less than 3.8, 3.4 to less than 3.6, 3.2 to less than 3.4, 3.0 to less than 3.2, 2.8 to less than 3.0, 2.6 to less than 2.8, 2.4 to less than 2.6, 2.2 to less than 2.4, 2.0 to less than 2.2, 1.8 to less than 2.0, 1.6 to less than 1.8, 1.4 to less than 1.6, 1.2 to less than 1.4, 1.0 to less than 1.2, 0.8 to less than 1.0, or 0.6 to less than 0.8.

[0182] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive). In some implementations, if at least a certain number are identified in Table 1 (the lower limit is selected from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250 or 300) and the upper limit is selected from 10, 15, 20, 25, 50, 75, 100), then... If a locus (1, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000) is differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the subject from whom the sample was obtained or from which the sample was derived has a specific HER2 status (e.g., HER2-positive). In certain specific implementations, if at least 1, 2, 3, 4, 5, 10, 20, 30, 40, or 50 sites identified in Table 1 (e.g., about 1 to about 1,000, about 5 to about 3,000, about 10 to about 1,000, about 25 to about 200, about 5, about 10, about 20, or about 50 sites) are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample was obtained or from which the sample was derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, 10%, 20%, 30%, 40%, 50%, 75%, or 100% of the sites identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).In some implementations, if at least a certain percentage of the loci identified in Table 1 (lower limit selected from 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, or 10% and upper limit selected from 1%, 2%, 3%, 4%, 5%, 10%, 20%, 30%, 40%, 50%, 75%, or 100%) are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).

[0183] In various implementations, if at least one of the top 3, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci identified in Table 1 (e.g., at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10) is differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive) (where, for example, the “top” 10 loci refer to the 10 loci with the highest absolute log2 (fold change) in Table 1). In some embodiments, if at least one of the first 10 loci identified in Table 1 is differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least one of the first 25 loci identified in Table 1 is differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least one of the first 50 loci identified in Table 1 is differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 10 loci identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 25 loci identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 50 loci identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive).

[0184] In various implementations, if at least one of the top 10 loci identified in Table 1 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or 10) and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least one of the first 25 loci identified in Table 1 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10, at least 15, at least 20, or 25) and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least one of the first 50 sites identified in Table 1 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10, at least 15, at least 20, or at least 25, at least 30, at least 35, at least 40, at least 45, or 50) and the total number of sites identified in Table 1 is at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 8 If 5, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive).In various implementations, if at least five of the first 25 loci identified in Table 1, and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least five of the first 50 loci identified in Table 1, and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 1 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0185] In various implementations, differential H3K4me3 modification refers to a methylation state characterized by an increase or decrease of at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 2-fold, 3-fold, 4-fold, 5-fold, 6-fold, 7-fold, 8-fold, 9-fold, 10-fold, 15-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, 45-fold, 40-fold, 45-fold, 45-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, 45-fold, compared to a reference. 50 times or more, or any range such as 1% to 50%, 50% to 2 times, 25% to 50 times, 25% to 30 times, 25% to 20 times, 25% to 16 times, 30% to 16 times, 50% to 16 times, 70% to 16 times, 2 times to 16 times, 2.2 times to 16 times, 2.6 times to 16 times, 3 times to 16 times, 3.4 times to 16 times, 4 times to 16 times, 4.5 times to 16 times, 5.2 times to 16 times, 6 times to 16 times, 7 times to 16 times or 8 times to 16 times (inclusive), optionally wherein the statistical significance of the increase or decrease is at least 5e-2, 1e-2, 5e-3, 1e-3, 5e-4, 1e-4, 5e-5, 1e-5, 5e-6 or 1e-6. In various implementations, the increase or decrease in the measured methylation value can be or is expressed as log2 (fold change), for example, log2 (fold change) is at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times, 15 times, 20 times or higher, or such as an increase or decrease of 0.1 times to 10 times, 0.2 times to 5 times, 0.2 times to 4.0 times, 0.4 times to 4.0 times, 0.4 times to Any range between 4.0x, 0.6x to 4.0x, 0.8x to 4.0x, 1.0x to 4.0x, 1.2x to 4.0x, 1.4x to 4.0x, 1.6x to 4.0x, 1.8x to 4.0x, 2.0x to 4.0x, 2.2x to 4.0x, 2.4x to 4.0x, 2.6x to 4.0x, 2.8x to 4.0x, or 3.0x to 4.0x (inclusive), optionally wherein the statistical significance of the increase or decrease is at least 5e-2, 1e-2, 5e-3, 1e-3, 5e-4, 1e-4, 5e-5, 1e-5, 5e-6, or 1e-6.

[0186] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 k4 sites (or any subset thereof) identified in Table 5 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).

[0187] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, or 17 k4 sites (or any subset thereof) identified in Table 6 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which the sample was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0188] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, or 17 k4 sites (or any subset thereof) identified in Table 7 are differentially modified with H3K4me3 compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0189] Difference H3K27ac modification Table 2 provides genomic loci exhibiting differential H3K27ac modification in HER2-positive and HER2-negative cancers, showing the chromosomal coordinates of each locus and its observed log2 (fold change) (HER2-positive / HER2-negative). Genomic loci are ordered based on their chromosomal coordinates, which are based on the human genome version hg19.

[0190] Those skilled in the art will understand that the methods disclosed herein do not require evaluation of H3K27ac modifications for every genomic locus listed in Table 2. Instead, H3K27ac modifications for a subset of loci can be evaluated. A subset of genomic loci in Table 2 can be selected based on various performance criteria (e.g., for determining HER2 status), such as selecting genomic loci exhibiting differential modifications at a specific statistical significance level and / or a specific difference threshold (e.g., measured log2 fold change) between relevant statuses. A subset of genomic loci can also be selected based on algorithms, such as during the process of obtaining a classifier. Those skilled in the art will understand that such subsets of loci in Table 2, and the loci included in such subsets, whether present individually or in randomly selected subsets, have at least equivalent information content (e.g., statistical significance and / or reliability) for the purposes disclosed herein (e.g., for determining HER2 status). See also the embodiments of this disclosure illustrating experiments demonstrating that information-rich classifiers can be produced using many different combinations of loci. Among other things, this disclosure specifically includes subsets of genomic loci listed in Table 2, which have absolute log2 (fold change) values ​​of 6.0 or higher, 5.5 or higher, 5.0 or higher, 4.5 or higher, 4.0 or higher, 3.5 or higher, 3.0 or higher, 2.5 or higher, 2.0 or higher, 1.9 or higher, 1.8 or higher, 1.7 or higher, 1.6 or higher, 1.5 or higher, 1.4 or higher, 1.3 or higher, 1.2 or higher, 1.1 or higher, 1.0 or higher, 0.9 or higher, 0.8 or higher, 0.7 or higher, 0.6 or higher, or 0.5 or higher. This disclosure also includes subsets of genomic loci listed in Table 2, wherein the absolute log2 (fold change) of these subsets is 6.0 or higher, 5.5 to less than 6.0, 5.0 to less than 5.5, 4.5 to less than 5.0, 4.0 to less than 4.5, 3.8 to less than 4.0, 3.6 to less than 3.8, 3.4 to less than 3.6, 3.2 to less than 3.4, 3.0 to less than 3.2, 2.8 to less than 3.0, 2.6 to less than 2.8, 2.4 to less than 2.6, 2.2 to less than 2.4, 2.0 to less than 2.2, 1.8 to less than 2.0, 1.6 to less than 1.8, 1.4 to less than 1.6, 1.2 to less than 1.4, 1.0 to less than 1.2, 0.8 to less than 1.0, or 0.6 to less than 0.8.

[0191] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample was obtained or from which the sample was derived is determined to have a specific HER2 status (e.g., HER2-positive). In some implementations, if at least a certain number of loci identified in Table 2 (lower limit selected from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, or 300 and upper limit selected from 10, 15, 20, 25, 50, 75, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000) are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In certain specific implementations, if at least 1, 2, 3, 4, 5, 10, 20, 30, 40, or 50 sites identified in Table 2 (e.g., about 1 to about 1,000, about 5 to about 3,000, about 10 to about 1,000, about 25 to about 200, about 5, about 10, about 20, or about 50 sites) are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which the sample was derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, 10%, 20%, 30%, 40%, 50%, 75%, or 100% of the sites identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).In some implementations, if at least a certain percentage of the loci identified in Table 2 (lower limit selected from 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, or 10% and upper limit selected from 1%, 2%, 3%, 4%, 5%, 10%, 20%, 30%, 40%, 50%, 75%, or 100%) are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).

[0192] In various implementations, if at least one of the top 3, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci identified in Table 2 (e.g., at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10) is modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive) (where, for example, the “top” 10 loci refer to the 10 loci with the highest absolute log2 (fold change) in Table 2). In some embodiments, if at least one of the first 10 loci identified in Table 2 is modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least one of the first 25 loci identified in Table 2 is modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least one of the first 50 loci identified in Table 2 is modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 10 loci identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 25 loci identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 50 loci identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive).

[0193] In various implementations, if at least one of the top 10 loci identified in Table 2 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or 10) and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least one of the first 25 loci identified in Table 2 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10, at least 15, at least 20, or 25) and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least one of the first 50 sites identified in Table 2 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10, at least 15, at least 20, or at least 25, at least 30, at least 35, at least 40, at least 45, or 50) and the total number of sites identified in Table 2 is at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, or 80) If 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive).In various implementations, if at least five of the first 25 loci identified in Table 2, and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from whom the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least five of the first 50 loci identified in Table 2, and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 2 are modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0194] In various implementations, differential H3K27ac modification refers to an acetylation state characterized by an increase or decrease of at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 2-fold, 3-fold, 4-fold, 5-fold, 6-fold, 7-fold, 8-fold, 9-fold, 10-fold, 15-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, 45-fold, 45-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, 45-fold, 45-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, or 45-fold compared to a reference. 50 times or more, or any range such as 1% to 50%, 50% to 2 times, 25% to 50 times, 25% to 30 times, 25% to 20 times, 25% to 16 times, 30% to 16 times, 50% to 16 times, 70% to 16 times, 2 times to 16 times, 2.2 times to 16 times, 2.6 times to 16 times, 3 times to 16 times, 3.4 times to 16 times, 4 times to 16 times, 4.5 times to 16 times, 5.2 times to 16 times, 6 times to 16 times, 7 times to 16 times or 8 times to 16 times (inclusive), optionally wherein the statistical significance of the increase or decrease is at least 5e-2, 1e-2, 5e-3, 1e-3, 5e-4, 1e-4, 5e-5, 1e-5, 5e-6 or 1e-6. In various implementations, the increase or decrease in the measured value of acetylation can be or is expressed as log2 (fold change), for example, log2 (fold change) is at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times, 15 times, 20 times or higher, or such as an increase or decrease of 0.1 times to 10 times, 0.2 times to 5 times, 0.2 times to 4.0 times, 0.4 times to 4.0 times, 0.4 times to Any range between 4.0x, 0.6x to 4.0x, 0.8x to 4.0x, 1.0x to 4.0x, 1.2x to 4.0x, 1.4x to 4.0x, 1.6x to 4.0x, 1.8x to 4.0x, 2.0x to 4.0x, 2.2x to 4.0x, 2.4x to 4.0x, 2.6x to 4.0x, 2.8x to 4.0x, or 3.0x to 4.0x (inclusive), optionally wherein the statistical significance of the increase or decrease is at least 5e-2, 1e-2, 5e-3, 1e-3, 5e-4, 1e-4, 5e-5, 1e-5, 5e-6, or 1e-6.

[0195] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, or 29 k27 sites (or any subset thereof) identified in Table 5 are differentially modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).

[0196] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, or 43 k27 sites (or any subset thereof) identified in Table 6 are differentially modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0197] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, or 37 k27 sites (or any subset thereof) identified in Table 7 are differentially modified with H3K27ac compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0198] Differential DNA methylation Table 3 provides genomic sites exhibiting differential DNA methylation in HER2-positive and HER2-negative cancers, showing the chromosomal coordinates of each site and its observed log2 (fold change) (HER2-positive / HER2-negative). Genomic sites are ordered based on their chromosomal coordinates, which are based on the human genome version hg19.

[0199] Those skilled in the art will understand that the methods disclosed herein do not require evaluating DNA methylation at every genomic site listed in Table 3. Instead, DNA methylation at a subset of sites can be evaluated. A subset of genomic sites in Table 3 can be selected based on various performance criteria (e.g., for determining HER2 status), such as selecting genomic sites exhibiting differential modifications at a specific statistical significance level and / or a specific difference threshold (e.g., measured log2 fold change) between relevant states. A subset of genomic sites can also be selected based on algorithms, such as during the process of obtaining a classifier. Those skilled in the art will understand that such subsets of sites in Table 3, and the sites included in such subsets, whether present individually or in randomly selected subsets, have at least equivalent information content (e.g., statistical significance and / or reliability) for the purposes disclosed herein (e.g., for determining HER2 status). See also the embodiments of this disclosure illustrating experiments demonstrating that using many different combinations of sites can produce information-rich classifiers. Among other things, this disclosure specifically includes subsets of genomic loci listed in Table 3, which have absolute log2 (fold change) values ​​of 6.0 or higher, 5.5 or higher, 5.0 or higher, 4.5 or higher, 4.0 or higher, 3.5 or higher, 3.0 or higher, 2.5 or higher, 2.0 or higher, 1.9 or higher, 1.8 or higher, 1.7 or higher, 1.6 or higher, 1.5 or higher, 1.4 or higher, 1.3 or higher, 1.2 or higher, 1.1 or higher, 1.0 or higher, 0.9 or higher, 0.8 or higher, 0.7 or higher, 0.6 or higher, or 0.5 or higher. This disclosure also includes subsets of genomic loci in Table 3 with absolute log2 (fold change) values ​​of 6.0 or higher, 5.5 to less than 6.0, 5.0 to less than 5.5, 4.5 to less than 5.0, 4.0 to less than 4.5, 3.8 to less than 4.0, 3.6 to less than 3.8, 3.4 to less than 3.6, 3.2 to less than 3.4, 3.0 to less than 3.2, 2.8 to less than 3.0, 2.6 to less than 2.8, 2.4 to less than 2.6, 2.2 to less than 2.4, 2.0 to less than 2.2, 1.8 to less than 2.0, 1.6 to less than 1.8, 1.4 to less than 1.6, 1.2 to less than 1.4, 1.0 to less than 1.2, 0.8 to less than 1.0, or 0.6 to less than 0.8.

[0200] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) identified in Table 3 are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive). In some implementations, if at least a certain number of sites identified in Table 3 (lower limit selected from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, or 300 and upper limit selected from 10, 15, 20, 25, 50, 75, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000) are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In certain specific implementations, if at least 1, 2, 3, 4, 5, 10, 20, 30, 40, or 50 sites identified in Table 3 (e.g., about 1 to about 1,000, about 5 to about 3,000, about 10 to about 1,000, about 25 to about 200, about 5, about 10, about 20, or about 50 sites) are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, 10%, 20%, 30%, 40%, 50%, 75%, or 100% of the sites identified in Table 3 are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).In some implementations, if at least a certain percentage of the sites identified in Table 3 (lower limits selected from 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, or 10% and upper limits selected from 1%, 2%, 3%, 4%, 5%, 10%, 20%, 30%, 40%, 50%, 75%, or 100%) are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from is determined to have a specific HER2 status (e.g., HER2-positive).

[0201] In various implementations, if at least one of the first 3, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci identified in Table 3 (e.g., at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10) is differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive) (where, for example, the “first” 10 loci refer to the 10 loci with the highest absolute log2 (fold change) in Table 3). In some embodiments, if at least one of the first 10 loci identified in Table 3 is differentially DNA-methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least one of the first 25 loci identified in Table 3 is differentially DNA-methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least one of the first 50 loci identified in Table 3 is differentially DNA-methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 10 loci identified in Table 3 are differentially DNA-methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 25 loci identified in Table 3 are differentially DNA-methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive). In some embodiments, if at least five of the first 50 loci identified in Table 3 are differentially DNA-methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), the subject from whom or from which the sample was obtained is determined to have a specific HER2 status (e.g., HER2-positive).

[0202] In various implementations, if at least one of the first 10 loci identified in Table 3 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or 10) and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 3 are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least one of the first 25 loci identified in Table 3 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10, at least 15, at least 20, or 25) and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 3 are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least one of the first 50 sites identified in Table 3 (e.g., at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, or at least 10, at least 15, at least 20, or at least 25, at least 30, at least 35, at least 40, at least 45, or 50) and the total number of sites identified in Table 3 is at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, or 80) If 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) are differentially methylated with DNA compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was derived is determined to have a specific HER2 status (e.g., HER2-positive).In various implementations, if at least five of the first 25 loci identified in Table 3, and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 loci (or any subset thereof) identified in Table 3 are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive). In various implementations, if at least five of the first 50 sites identified in Table 3, and a total of at least 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 150, 200, 250, 300, 350, 400, 450, 500, 750, 1000, 1500, 2000, 2500, or 3000 sites (or any subset thereof) identified in Table 3 are differentially DNA methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample is derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0203] In various implementations, differential DNA methylation refers to a methylation state characterized by an increase or decrease of at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 2-fold, 3-fold, 4-fold, 5-fold, 6-fold, 7-fold, 8-fold, 9-fold, 10-fold, 15-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, 45-fold, ...2-fold, 3-fold, 4-fold, 5-fold, 6-fold, 7-fold, 8-fold, 9-fold, 10-fold, 15-fold, 20-fold, 25-fold, 30-fold, 35-fold, 40-fold, 45-fold, 5-fold, 5-fold, 2-fold, 3-fold, 40-fold, 45-fold, 5-fold, 2-fold, 3-fold, 3-fold, 40-fold, 45-fold, 5-fold, 2-fold, 3-fold, 4-fold, 5-fold, 2-fold, 3-fold, 3-fold, 40-fold, 45-fold, 5-fold, 2-fold, 3-fold, 4-fold, 5-fold, 2-fold, 3-fold, 3-fold, 40-fold, 45-fold, 0 times or greater, or any range such as 1% to 50%, 50% to 2 times, 25% to 50 times, 25% to 30 times, 25% to 20 times, 25% to 16 times, 30% to 16 times, 50% to 16 times, 70% to 16 times, 2 times to 16 times, 2.2 times to 16 times, 2.6 times to 16 times, 3 times to 16 times, 3.4 times to 16 times, 4 times to 16 times, 4.5 times to 16 times, 5.2 times to 16 times, 6 times to 16 times, 7 times to 16 times, or 8 times to 16 times (inclusive), optionally wherein the statistical significance of the increase or decrease is at least 5e-2, 1e-2, 5e-3, 1e-3, 5e-4, 1e-4, 5e-5, 1e-5, 5e-6, or 1e-6. In various implementations, the increase or decrease in the measured methylation value can be or is expressed as log2 (fold change), for example, log2 (fold change) is at least 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 75%, 100%, 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, 10 times, 15 times, 20 times or higher, or such as an increase of 0.1 times to 10 times, 0.2 times to 5 times, 0.2 times to 4.0 times, 0.4 times to 4.0 times, 0.4 times to 4.0 times. Any range between 0.0x, 0.6x to 4.0x, 0.8x to 4.0x, 1.0x to 4.0x, 1.2x to 4.0x, 1.4x to 4.0x, 1.6x to 4.0x, 1.8x to 4.0x, 2.0x to 4.0x, 2.2x to 4.0x, 2.4x to 4.0x, 2.6x to 4.0x, 2.8x to 4.0x, or 3.0x to 4.0x (inclusive), optionally wherein the statistical significance of the increase or decrease is at least 5e-2, 1e-2, 5e-3, 1e-3, 5e-4, 1e-4, 5e-5, 1e-5, 5e-6, or 1e-6.

[0204] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15 mbd sites (or any subset thereof) identified in Table 5 are differentially methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which the sample was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0205] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, or 24 mbd sites (or any subset thereof) identified in Table 6 are differentially methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0206] In various implementations, if at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, or 24 mbd sites (or any subset thereof) identified in Table 7 are differentially methylated compared to a reference (e.g., a sample from a HER2-negative or healthy subject), then the sample or the subject from which the sample was obtained or from which it was derived is determined to have a specific HER2 status (e.g., HER2-positive).

[0207] Differential chromatin accessibility or transcription factor binding The genomic loci provided in Tables 1 through 3 can also demonstrate differential chromatin accessibility or transcription factor binding in HER2-positive cancers (e.g., HER2-3+ cancers based on IHC testing) and HER2-negative cancers (e.g., HER2-0 cancers based on IHC testing).

[0208] In various embodiments, while not wishing to be bound by any particular scientific theory, histone methylation (e.g., H3K4me3) corresponds to and / or is related to chromatin accessibility. In various embodiments, while not wishing to be bound by any particular scientific theory, histone acetylation (e.g., H3K27ac) corresponds to and / or is related to chromatin accessibility. In various embodiments, while not wishing to be bound by any particular scientific theory, DNA methylation corresponds to and / or is related to chromatin accessibility.

[0209] In some implementations, while not intended to be limited to any particular scientific theory, chromatin accessibility corresponds to and / or is related to H3K4me3 modification. Therefore, in some implementations, HER2 status can be determined by detecting and quantifying chromatin accessibility at one or more genomic sites in Table 1, based on the section above discussing exemplary genomic sites with differential H3K4me3 modification.

[0210] In some implementations, while not intended to be limited to any particular scientific theory, chromatin accessibility corresponds to and / or is related to H3K27ac modifications. Therefore, in some implementations, HER2 status can be determined by detecting and quantifying chromatin accessibility at one or more genomic sites in Table 2, based on the section above discussing exemplary genomic sites with differential H3K27ac modifications.

[0211] In some implementations, while not intended to be limited to any particular scientific theory, chromatin accessibility corresponds to and / or is related to DNA methylation. Therefore, in some implementations, based on the section above discussing exemplary genomic sites with differential DNA methylation, HER2 status can be determined by detecting and quantifying chromatin accessibility at one or more genomic sites in Table 3.

[0212] In various embodiments, while not wishing to be bound by any particular scientific theory, histone methylation (e.g., H3K4me3) corresponds to and / or is associated with transcription factor binding. In various embodiments, while not wishing to be bound by any particular scientific theory, histone acetylation (e.g., H3K27ac) corresponds to and / or is associated with transcription factor binding. In various embodiments, while not wishing to be bound by any particular scientific theory, DNA methylation corresponds to and / or is associated with transcription factor binding.

[0213] In some implementations, while not intended to be limited to any particular scientific theory, the binding of RNA pol II corresponds to and / or is associated with H3K4me3 modification. Therefore, in some implementations, based on the section above discussing exemplary genomic sites with differential H3K4me3 modification, HER2 status can be determined by detecting and quantifying the binding of RNA pol II at one or more genomic sites in Table 1.

[0214] In some implementations, while not intended to be limited to any particular scientific theory, the binding of p300, the mediator complex, the cohesin complex, or RNA pol II corresponds to and / or is associated with H3K27ac modification. Therefore, in some implementations, based on the section above discussing exemplary genomic sites with differential H3K27ac modification, HER2 status can be determined by detecting and quantifying the binding of p300, the mediator complex, the cohesin complex, or RNA pol II at one or more genomic sites in Table 2.

[0215] In some embodiments, while not intended to be limited to any particular scientific theory, the binding of FOXA1, ESR1, PR, MYC, EN1, FOXM1, KLF4, AP-2, RARA, or RUNX1 corresponds to and / or is associated with histone methylation (e.g., H3K4me3), histone acetylation (e.g., H3K27ac), or DNA methylation. Therefore, in some embodiments, based on the section above discussing exemplary genomic sites with differential histone methylation (e.g., H3K4me3), histone acetylation (e.g., H3K27ac), or DNA methylation, HER2 status can be determined by detecting and quantifying the binding of FOXA1, ESR1, PR, MYC, EN1, FOXM1, KLF4, AP-2, RARA, or RUNX1 at one or more genomic sites in Tables 1 through 3.

[0216] application The methods, kits, and systems disclosed herein include analyzing differentially modified and / or differentially accessible genomic sites to determine the HER2 status of cancer. The methods, kits, and systems disclosed herein can be used in any of a variety of applications. For example, the methods, kits, and systems disclosed herein can be used to detect and / or treat cancer based on HER2 status. The methods, kits, and systems disclosed herein can also be used to detect or identify cancers, such as resistance to therapy in breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer, or the transformation of cancer from one cancer subtype to another.

[0217] In various implementations, the methods, kits, and systems of this disclosure can be applied to asymptomatic human subjects. As used herein, a subject may be described as “asymptomatic” if they do not report, and / or demonstrate, by noninvasive, observable indicators (e.g., no device-based detection, tissue sample analysis, body fluid analysis, surgery, or cancer screening, one, several, or all of which) sufficient cancer characteristics to support a medically reasonable suspicion that the subject may have cancer, such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer. The methods, kits, and systems of this disclosure enable early cancer detection, leading to medical benefits, including the possibility of early treatment and improved treatment outcomes.

[0218] In various implementations, the methods, kits, and systems of this disclosure can be applied to human subjects who are highly susceptible to cancer.

[0219] In various embodiments, the methods, kits, and systems of this disclosure can be applied to symptomatic human subjects. As used herein, a subject may be described as “symptomatic” if the subject reports, and / or demonstrates by non-invasive, observable indicators (e.g., without one, several, or all of device-based detection, tissue sample analysis, body fluid analysis, surgery, or cancer screening) that there are sufficient cancer characteristics to support a medically reasonable suspicion that the subject may have cancer, such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer. For example, in various embodiments, a sample from a subject (optionally, wherein the subject has cancer but their HER2 status is unknown) may be measured according to one or more embodiments of this disclosure to determine whether the cancer is HER2 positive or HER2 negative. In various embodiments, a sample from a subject with cancer, known or suspected to be HER2 positive (or HER2 negative), may be measured according to one or more embodiments of this disclosure to determine whether the cancer is actually HER2 positive (or HER2 negative).

[0220] In some embodiments, the methods, kits, and systems of this disclosure can be used to determine if a subject has HER2-positive cancer, optionally HER2-positive cancer associated with an IHC-based HER2-3+, HER2-2+, or HER2-1+ score, or HER2-low cancer based on an IHC / ISH test. In some embodiments, the methods, kits, and systems of this disclosure can be used to determine if a subject has HER2-negative cancer, optionally HER2-negative cancer associated with an IHC-based HER2-0 score.

[0221] In some embodiments, the methods, kits, and systems of this disclosure can be used to verify or confirm that a subject has HER2-positive cancer, optionally based on prior identification of HER2-3+, HER2-2+, or HER2-1+ cancer by IHC testing, or HER2-low cancer by IHC / ISH testing. In some embodiments, the methods, kits, and systems of this disclosure can be used to verify or confirm that a subject has HER2-negative cancer, optionally based on prior identification of HER2-0 cancer by IHC testing.

[0222] In some embodiments, the methods, kits, and systems of this disclosure are used to identify and detect novel HER2-related categories independent of IHC or ISH scores. For example, instead of training a classifier on samples from a cohort defined based on HER2 IHC or ISH testing, it is trained on samples from a cohort defined based on whether or not a subject responds to a specific HER2 target. The resulting classifier is then used to identify subjects more likely to respond to a specific HER2 target without any IHC or ISH score. Therefore, it should be understood that the term “HER2 status” as used herein is not limited to HER2-positive and HER2-negative or conventional HER2 scores based on IHC or ISH testing, but can also encompass any HER2-related category, including whether a subject will respond to a specific HER2 target.

[0223] Those skilled in the art will understand that regular, preventative, and / or disease-preventive screening to determine HER2 status improves the diagnosis of cancer, including and / or particularly early-stage cancer. Therefore, this disclosure specifically provides methods, kits, and systems particularly useful for the diagnosis and treatment of early-stage cancer. Generally, the methods, kits, and systems of this disclosure are particularly likely to detect early-stage HER2-positive cancer, especially in embodiments where HER2-positive cancer testing is performed annually according to this disclosure, and / or where the subject is asymptomatic at the time of testing. In various embodiments, testing according to the methods, kits, and systems of this disclosure reduces cancer mortality, for example, through early cancer diagnosis.

[0224] In various implementations, HER2 status determination according to this disclosure may be performed once or multiple times on a given subject. In various implementations, HER2 status determination is performed periodically according to this disclosure, such as every six months, annually, every two years, every three years, every four years, every five years, or every ten years.

[0225] In various embodiments, the methods, kits, and systems disclosed herein provide for the determination of HER2 status. In other cases, the methods, kits, and systems disclosed herein will indicate HER2 status but cannot determine it. In all cases where the methods, kits, and systems of this disclosure are used to determine HER2 status, further confirmatory assays may then be performed to confirm, support, refute, or reject the previously determined status, for example, the determination according to this disclosure. As used herein, confirmatory assays may be currently medically recognized HER2 tests, such as HER2 scores based on IHC or ISH tests.

[0226] In various embodiments, cancer treatment is performed after determining the HER2 status according to one or more methods, kits, and / or systems disclosed herein. In various embodiments, cancer treatment includes administering a treatment regimen comprising one or more cancer therapies provided herein, including but not limited to one or more of HER2-targeted therapy, surgery, radiation, endocrine therapy, chemotherapy, and / or immunotherapy. In various embodiments, cancer treatment includes administering a treatment regimen comprising one or more treatments provided herein, which are available, appropriate, and / or preferred for a specific HER2 status.

[0227] In various implementations, methods, kits, and systems can be used to determine whether a particular subject and / or cancer is likely and / or can be characterized as responsive to HER2-targeted therapy. In some such implementations, the methods, kits, and systems can subsequently be used to treat the subject with HER2-targeted therapy.

[0228] In various implementations, methods, kits, and systems can be used to determine whether a particular subject and / or cancer is likely and / or characterized as resistant to, unresponsive to, or not recommended for treatment with HER2-targeted therapy. In some such implementations, the methods, kits, and systems can be followed by treatment with one or more of surgery and / or radiation, endocrine therapy (if hormone receptors such as estrogen receptors are positive), chemotherapy, and immunotherapy, instead of HER2-targeted therapy.

[0229] Response can refer to the ability or likelihood of a therapy to shrink tumor size or inhibit tumor growth or metastasis. Response can refer to improved prognosis (e.g., prolonged time to cancer recurrence or extended life expectancy, such as extended overall survival, recurrence-free survival, metastasis-free survival, or disease-free survival). Response can refer to the realization of treatment benefits, including, for example, improvement in one or more symptoms of cancer, such as breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer. Response can be measured quantitatively (e.g., tumor size; measurements of histone modifications, chromatin accessibility, transcription factor binding, or DNA methylation at one or more genomic loci; or clinical benefit (CBR) calculations) or qualitatively (e.g., by metrics such as “pathological complete response” (pCR), “clinical complete response” (cCR), “clinical partial response” (cPR), “clinical stable disease” (cSD), “clinical progressive disease” (cPD), or other qualitative criteria). Resistance can refer to the inability or improbability of a therapy to achieve the desired therapeutic effect in the subject and / or cancer (e.g., shrinking tumor size, improving prognosis, or other treatment benefits, such as, for example, improvement in one or more cancer symptoms). Resistance includes acquired resistance and natural resistance. In some embodiments, resistance includes the extent to which one or more desired therapeutic benefits resulting from administration of a therapy to a subject and / or cancer are less than those anticipated and / or achieved in a reference (e.g., less than 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20%, or 10% of the benefits achieved in a reference).

[0230] In various implementations, methods, kits, and systems can be used to detect the clinical efficacy of cancer treatments (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer). For example, the methods and / or compositions of this disclosure can be used to determine the presence of cancer or HER2 status of cancer in a subject during treatment. The methods and / or compositions disclosed herein may be used in conjunction with or confirmed by other methods for determining the presence of cancer or the HER2 status of cancer, such as measuring the size or characteristics of the tumor by techniques such as CT, PET, mammography, ultrasound, palpation, histology, biopsy, or caliper measurement after surgical resection, or measuring the size or characteristics of the tumor by various qualitative, quantitative, or semi-quantitative scoring systems (including, but not limited to, those based on IHC or ISH tests, residual cancer burden (Symmans et al., J Clin Oncol (2007) 25:4414-4422, incorporated herein by reference in its entirety) or Miller-Payne score (Ogston et al., Breast (2003) 12:320-327, incorporated herein by reference in its entirety), in a qualitative manner (e.g., “pathological complete response” (pCR), “clinical complete response” (cCR), “clinical partial response” (cPR), “clinical stable disease” (cSD), or “clinical progressive disease” (cPD)).

[0231] In some embodiments, the methods, kits, and systems described herein can be used to monitor disease progression in a subject. In some embodiments, monitoring progression requires obtaining and characterizing samples from the subject at least at a first time point and a second time point. In some embodiments, at the first time point, the subject has been diagnosed with cancer (e.g., HER2-positive or HER2-negative cancer). In some embodiments, at the first time point, the subject has been diagnosed with cancer and treatment has been administered before or near the first time point (e.g., on the same day as the first time point) or between the first and second time points; in such embodiments, determining HER2 status at at least the first and second time points can be used to monitor treatment efficacy and / or determine when treatment should be changed. For example, in some embodiments, the subject has been diagnosed with cancer at the first time point, is receiving or will receive treatment, and disease status can be monitored, which may be used, for example, to determine whether treatment should be changed. In some embodiments, treatment efficacy can be monitored, for example, by using the methods described herein to determine a decrease or increase in disease status signals, which may be useful in determining whether the administered treatment is effective and / or whether treatment should be changed. In some embodiments, at the first time point, the subject's cancer has remitted (e.g., the subject has very little residual disease). In implementations where cancer has been alleviated, the methods, kits, and systems described herein can be used, for example, to detect cancer recurrence, and may be faster, less costly, and / or less invasive compared to methods that rely on, for example, tissue biopsy and / or imaging techniques.

[0232] In some implementations, the methods, kits, and systems provided herein for HER2 status determination may inform treatment and / or payment (e.g., medical expense reimbursement or reduction, such as testing or treatment) decisions and / or actions, for example, by individuals, healthcare institutions, healthcare practitioners, health insurance providers, government agencies, or other parties interested in healthcare costs.

[0233] In some implementations, the methods, kits, and systems provided herein for determining HER2 status may inform decisions by health insurance providers regarding whether to reimburse healthcare payers or recipients, for example, for (1) the HER2 status determination itself (e.g., reimbursement for tests not otherwise available, reimbursement only for regular / routine tests, or reimbursement only for ad hoc and / or incidental tests); and / or for (2) treatment, including initiating, maintaining, and / or changing therapy, for example, based on the determined HER2 status. For example, in some implementations, the methods, kits, and systems provided herein for determining HER2 status are used as a basis, supplementary basis, or supporting basis for determining whether to provide reimbursement or cost reduction to healthcare payers or recipients. In some cases, a party seeking reimbursement or cost reduction may provide the results of a HER2 status determination made according to this disclosure and submit such a request for healthcare reimbursement or reduction. In some cases, a party deciding whether to provide healthcare reimbursement or reduction will make its decision wholly or in part based on receiving and / or reviewing the results of a HER2 status determination made according to this disclosure.

[0234] In various embodiments, HER2 status determination using the methods, kits, and systems disclosed herein can be used to classify subjects, samples, and / or tumors (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, or lung cancer subjects, samples, and / or tumors). In various embodiments, the methods, kits, and systems disclosed herein can be used to generate a set of subjects, samples, and / or tumors identified according to the methods, kits, and systems of the present invention, each subject, sample, and / or tumor being classified to correspond to a specific HER2 status, and optionally two or more such classifications of subjects, samples, and / or tumors are used to identify and distinguish these categories. Right now Biomarkers that differentiate subjects, samples, and / or tumors based on their HER2 status (e.g., by the type of subject, sample, and / or tumor).

[0235] For illustrative purposes, but not limited to, in one exemplary assay of this disclosure, one or more samples (e.g., liquid biopsy samples containing cfDNA, e.g., plasma samples containing cfDNA) obtained from a subject are analyzed by a method comprising enriching cfDNA containing specific histone modifications, wherein enrichment is performed by incubating the sample with a reagent that specifically binds to the histone modifications to be enriched, and sequencing the enriched cfDNA. An example of such an assay is ChIP-seq for histone modifications (e.g., H3K4me3 and / or H3K27ac). Sequence reads (e.g., ChIP-seq sequence reads) can be aligned to human genome version hg19, for example using Burrows-WheelerAligner (BWA). Non-unique mappings and redundant reads are optionally discarded.

[0236] For example, MACS v2.1.1.20140616 can be used for sequence (e.g., ChIP-seq) peak recall with a q-value (FDR) threshold of 0.01. Sequence (e.g., ChIP-seq) data quality can optionally be assessed using one or more of a variety of metrics, including total number of peaks, FRiP (fraction of reads in a peak) score, number of high-confidence peaks (e.g., >10-fold enrichment over background), and percentage overlap of peaks with “blacklisted” DHS peaks from the ENCODE project (Amemiya et al., Sci Rep (2019) 9(1):9354). If sequence (e.g., ChIP-seq) data quality falls below a certain threshold, the data may be discarded and the assay repeated. Sequence (e.g., ChIP-seq) peaks overlapping with selected genomic sites that are differentially modified for relevant histone modifications (e.g., Tables 1-2 and / or Tables 5-7) as described herein can be used to determine HER2 status. The number of reads overlapping with selected genomic sites with relevant histone modifications can be summed, for example, in some embodiments, selecting all differentially modified genomic sites with an absolute log2 (fold change) ≥ 4.0. In some embodiments, the average number of reads in the local background of each ChIP-seq peak is subtracted to improve the signal-to-noise ratio. In some embodiments, the read density of one or more histone modifications can be calculated by including: (1) summing the background adjustment sequence counts at one or more genomic sites and dividing the sum by the total number of kilobases at one or more genomic sites; or (2) for each genomic site, determining the ratio of the background adjustment fragment count to the number of kilobases at the genomic site and then summing the ratios for each site. In some embodiments, the method includes determining a HER2 positive / HER2 negative ratio score, for example, by including: (a) calculating the HER2 positive read density, calculating the HER2 negative read density, and dividing the HER2 positive read density by the HER2 negative read density. In some implementations, the HER2-positive sequence read density can be determined by a method comprising: calculating the sequence read density using one or more genomic loci, which have increased levels of one or more epigenetic biomarkers in samples obtained from one or more subjects with HER2-negative cancer compared to samples obtained from one or more subjects with HER2-positive cancer.In some embodiments, the HER2-negative sequence read density can be determined by methods including: calculating the sequence read density using one or more genomic loci, which have increased levels of one or more epigenetic biomarkers in samples obtained from one or more subjects with HER2-positive cancer compared to samples obtained from one or more subjects with HER2-negative cancer. A HER2-positive / HER2-negative ratio score is used to determine one or more histone modifications. In some embodiments, a HER2-positive / HER2-negative ratio score is determined for H3K4me3 modification. In some embodiments, a HER2-positive / HER2-negative ratio score is determined for H3K27ac modification. In some embodiments, a HER2-positive / HER2-negative ratio score is determined for methylated DNA. In some embodiments, a HER2-positive / HER2-negative ratio score is determined for H3K4me3 modification and H3K27ac modification, H3K4me3 and methylated DNA, or H3K27ac and methylated DNA. In some embodiments, a HER2-positive / HER2-negative ratio score is determined for each of H3K4me3 modification, H3K27ac modification, and methylated DNA. In some implementations, HER2-positive / ER-negative ratio scores for two or more different epigenetic biomarkers can be combined. In some implementations, fitted values ​​determined by logistic regression can be used to combine each ratio score.

[0237] The data can then be log2 transformed and quantile normalized to match the data distribution used to train the classifier. The normalized data can be used as input to a classifier trained using the same histone modifications and selected genomic sites. The classifier can then use the input data to determine the HER2 status of a subject's cancer. It should be understood that this or similar methods can be applied to the assays of quantifying chromatin accessibility, transcription factor binding, and / or DNA methylation as described in this disclosure.

[0238] In some embodiments, multiple epigenetic biomarkers (e.g., one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation) can be quantified in a single sample. In such embodiments, two or more assessments of epigenetic biomarkers can be performed sequentially (meaning each modification can be detected sequentially in a single sample) or in parallel (meaning a single sample can be split into multiple fractions and each fraction can be analyzed to quantify the epigenetic biomarkers). In some embodiments, H3K4me3 and H3K27ac histone modifications; H3K4me3 modification and DNA methylation; H3K27ac modification and DNA methylation; or H3K4me3 modification, H3K27ac histone modification, and DNA methylation are quantified in a single sample.

[0239] To avoid any doubt, those skilled in the art will understand from this disclosure that the methods, kits, and systems for determining HER2 status disclosed herein are for at least in vitro use. Therefore, all aspects and embodiments of this disclosure can be at least... exist Performed and / or used in vitro.

[0240] Those skilled in the art will also understand that, in some embodiments, the methods of this disclosure may be implemented on and / or in conjunction with computer programs and computer systems. In some embodiments, the methods of this disclosure may be implemented on and / or in conjunction with a non-transient computer-readable storage medium encoded with a computer program, wherein the program contains instructions that, when executed by one or more processors, cause the one or more processors to perform operations to execute the methods. The computer system may also store and manipulate data generated by the methods of this disclosure, including multiple genomic site modification states and / or accessibility state changes / profiles, which the computer system may use to implement the methods disclosed herein. In some embodiments, the computer system (i) receives modification state and / or accessibility state data; (ii) stores the data; and (iii) compares the data in any of the various ways described herein (e.g., analysis relative to a suitable reference), for example, to determine HER2 status. In some embodiments, the computer system (i) compares the genomic site modification and / or accessibility status with a reference; and (ii) outputs an indication indicating whether the genomic site modification and / or accessibility status differs significantly from the reference, and / or provides information regarding the determination of HER2 status.

[0241] Based on the knowledge possessed by those skilled in the art of bioinformatics and / or computer science, various types of computer systems can be used to implement the methods of this disclosure. During the operation of such a computer system, several software components can be loaded into memory. These software components may include standard software components in the art and components specific to this disclosure (e.g., the dCHIP software described in Lin et al., Bioinformatics (2004) 20:1233-1240, which is incorporated herein by reference in its entirety; radial basis function (RBM) algorithms known in the art). The methods of this disclosure can also be programmed or modeled in mathematical software packages that allow symbolic input equations and advanced processing specifications, including specific algorithms to be used, thereby eliminating the need for users to programmatically write individual equations and algorithms. Such software packages include, for example, Matlab from Mathworks (Natick, MA), Mathematica from Wolfram Research (Champaign, IL), S-Plus from MathSoft (Seattle, WA), R from the R Foundation for Statistical Computation (Vienna, Austria), Python from the Python Software Foundation (Wilmington, DE), or Perl from the Perl Foundation (Holland, MI). In some embodiments, the computer system includes a database for storing data on the status and / or accessibility of genomic site modifications. Such stored files can then be accessed and used for comparisons of interest. Other alternative program structures and computer systems, besides those exemplary described herein, will readily be apparent to those skilled in the art.

[0242] As illustrated in the embodiments, various algorithms can be applied to compare the modification status and / or accessibility status of genomic loci between a sample and a reference sample, where the genomic loci are differentially modified in different HER2 states. In various embodiments, the algorithm may be a single learned statistical classifier system. Other suitable statistical algorithms are well known to those skilled in the art. For example, a learned statistical classifier system includes a machine learning algorithm technique capable of adapting to complex datasets (e.g., a set of genomic loci of interest) and making decisions based on such datasets. In some embodiments, a single learned statistical classifier system, such as a classification tree (e.g., a random forest), is used. In other embodiments, combinations of 2, 3, 4, 5, 6, 7, 8, 9, 10, or more learned statistical classifier systems are preferably used in tandem. Examples of learning statistical classifier systems include, but are not limited to, those systems described in the embodiments, as well as systems using: inductive learning (e.g., decision / classification trees, such as random forests, classification and regression trees (C&RT), boosting trees, etc.), probabilistic approximate correct (PAC) learning, connectionist learning (e.g., neural networks (NN), artificial neural networks (ANN), neural fuzzy networks (NFN), network structures, perceptrons such as multilayer perceptrons, multilayer feedforward networks, applications of neural networks, Bayesian learning in belief networks, etc.), reinforcement learning (e.g., passive learning in known environments such as naive learning, adaptive dynamic learning and temporal difference learning, passive learning in unknown environments, active learning in unknown environments, learning action-value functions, applications of reinforcement learning, etc.), and genetic algorithms and evolutionary programming. Other learned statistical classifier systems include support vector machines (e.g., kernel methods), multivariate adaptive regression splines (MARS), Levenberg-Marquardt algorithms, Gauss-Newton algorithms, mixtures of Gaussians, gradient descent, and learned vector quantization (LVQ). In some implementations, the methods of this disclosure may include sending classification results to medical practitioners, such as oncologists.

[0243] In various implementations, the area under the recipient operating characteristic curve (AUROC) used to determine whether a subject has a specific HER2 cancer status (e.g., HER2-positive cancer versus HER2-negative cancer) is greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95).

[0244] Preparation and administration of therapeutic agents This disclosure includes methods for administering therapeutic agents or regimens to subjects based on HER2 status of cancer (e.g., breast cancer, gastric / gastric esophageal cancer, colorectal cancer, lung cancer, etc.). Generally, the therapeutic agents or regimens provided herein are available, suitable, and / or preferred for a given HER2 status. Those skilled in the art will understand the recommended and / or government-approved formulations and / or dosages of the various therapeutic agents provided herein.

[0245] This disclosure includes pharmaceutical compositions for delivering one or more therapeutic agents to a subject. As disclosed herein, the pharmaceutical composition may be any form known in the art, including formulations administered via any route known in the art. An appropriate method of administration may be selected based on the subject's age and medical condition.

[0246] The pharmaceutical compositions disclosed herein may be in the form of, for example, liquid, semi-solid, and solid dosage forms. The pharmaceutical compositions disclosed herein may be in the form of, for example, liquid solutions (e.g., injectable and infusionable solutions), dispersions or suspensions, tablets, pills, powders, and liposomes. The choice or use of any particular form may depend in part on the intended route of administration and therapeutic application. Therefore, the compositions may be formulated for administration via parenteral (e.g., intravenous, subcutaneous, intraperitoneal, or intramuscular injection) or non-parenteral routes. As used herein, parenteral administration refers to a route of administration other than enteral and local administration, typically by injection or infusion.

[0247] In some embodiments, the compositions provided herein are available in unit dosage forms suitable for self-administration. Such unit dosage forms may be provided in containers such as tablets, vials, cartridges, pre-filled syringes, or disposable injection pens.

[0248] The pharmaceutical compositions disclosed herein may be in injectable or infusion-ready forms. For example, this disclosure includes sterile formulations for injection or infusion that can be formulated according to conventional pharmaceutical practices. A sterile solution can be prepared by incorporating the desired amount of the composition described herein with one of the ingredients listed above or a combination thereof into a suitable solvent, followed by filtration sterilization as needed. For example, an isotonic solution containing glucose and other supplements (such as D-sorbitol, D-mannose, D-mannitol, or sodium chloride) can be used as an aqueous solution for injection, optionally combined with a suitable solubilizer (e.g., alcohols such as ethanol and / or polyols such as propylene glycol or polyethylene glycol and / or nonionic surfactants such as polysorbate 80™ or HCO-50, etc.). In the case of sterile powders used to prepare sterile injectable solutions, the preparation methods include vacuum drying and freeze-drying, which produce a powder of the composition described herein plus any additional desired ingredients (see below) from its previously sterile filtered solution. Appropriate solution fluidity can be achieved, for example, by using coatings such as lecithin, by maintaining the desired particle size in the case of dispersions, and by using surfactants. Extended absorption of injectable compositions can be achieved by including, for example, monostearate and gelatin as agents that extend absorption. In certain cases, pharmaceutical compositions can be formulated into buffer solutions, for example, at suitable concentrations and suitable for storage (e.g., at 2-8°C (e.g., 4°C)).

[0249] In various embodiments, the pharmaceutical compositions of this disclosure can be formulated as solutions, microemulsions, dispersions, liposomes, or other ordered structures suitable for stable storage at high concentrations. Typically, dispersions are prepared by incorporating the compositions described herein into a sterile medium containing an alkaline dispersion medium and the other desired components listed above.

[0250] In various cases, pharmaceutical compositions may be formulated to include pharmaceutically acceptable carriers or excipients. Pharmaceutically acceptable carriers include, but are not limited to, any and all physiologically compatible solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic agents, and absorption delay agents.

[0251] In some embodiments, compositions can be formulated with a carrier that prevents rapid release of the therapeutic agent, such as controlled-release formulations comprising an implant and a microencapsulated delivery system. Biodegradable, biocompatible polymers such as ethylene vinyl acetate, polyanhydride, polyglycolic acid, collagen, polyorthoester, and polylactic acid can be used. Many methods for preparing such formulations are known in the art. See, for example, JR Robinson (1978) “Sustained and Controlled Release Drug Delivery Systems,” Marcel Dekker, Inc., New York.

[0252] The route of administration can be parenteral, such as by injection. Injection administration can be via intravenous injection, intramuscular injection, intraperitoneal injection, or subcutaneous injection. Administration can be systemic or local. In some embodiments, the compositions described herein can be therapeutically delivered to a subject via local administration. As used herein, “local administration” or “local delivery” means delivery that does not rely on the delivery of the composition or therapeutic agent to its intended target tissue or site via the vascular system. For example, the composition can be delivered by injection or implantation of the composition or therapeutic agent or by injection or implantation of a device containing said composition or therapeutic agent. In some embodiments, after local administration near a target tissue or site, the composition or therapeutic agent or one or more components thereof may diffuse to the intended target tissue or site, which is not the site of administration.

[0253] Pharmaceutical compositions can be administered parenterally in the form of injectable formulations comprising a sterile solution or suspension in water or another pharmaceutically acceptable liquid. For example, a pharmaceutical composition can be formulated by suitably combining a therapeutic molecule with a pharmaceutically acceptable medium or medium (such as sterile water and saline, vegetable oil, emulsifier, suspending agent, surfactant, stabilizer, flavoring excipient, diluent, carrier, preservative, binder) and then mixing them in a unit dose form required by generally accepted pharmaceutical practice. Examples of oily liquids include sesame oil and soybean oil, and it can be combined with benzyl benzoate or benzyl alcohol as a solubilizer. Other items that may be included are buffers (such as phosphate buffers or sodium acetate buffers), analgesics (such as procaine hydrochloride), stabilizers (such as benzyl alcohol or phenol), and antioxidants. The formulated injectable can be packaged in suitable ampoules.

[0254] In various implementations, subcutaneous administration can be achieved by devices such as syringes, pre-filled syringes, autoinjectors (e.g., disposable or reusable), injection pens, patch syringes, wearable syringes, portable syringe infusion pumps with subcutaneous infusion devices, or other devices that combine with therapeutic agents for subcutaneous injection.

[0255] The injection system disclosed herein may employ a delivery pen as described in U.S. Patent No. 5,308,341. Insulin pens are most commonly used for self-delivering insulin to diabetic patients and are well known in the art. Such devices may include at least one injection needle, typically pre-filled with one or more therapeutic unit doses of solution containing a therapeutic agent, and may be used to rapidly deliver the solution to the subject with minimal pain. A drug delivery pen includes a vial holder into which a vial containing a therapeutic agent or other medication may be housed. The injection pen may be a fully mechanical device or may be combined with electronic circuitry to precisely set and / or indicate the dosage of medication injected into the user. See, for example, U.S. Patent No. 6,192,891. In some embodiments, the needle of the pen device is disposable, and the kit includes one or more disposable replacement needles. Pen devices suitable for delivering any of the currently characteristic compositions are also described, for example, in U.S. Patent Nos. 6,277,099; 6,200,296; and 6,146,361, the disclosure of each of which is incorporated herein by reference in its entirety. For example, U.S. Patent No. 7,556,615 describes a microneedle-based pen device, the disclosure of which is incorporated herein by reference in its entirety. See also the MOLLY precision pen injector (PPI) device manufactured by Scandinavian Health Ltd. TM .

[0256] In some embodiments, administration of the therapeutic agent as described herein is achieved by administering to a subject a nucleic acid encoding the therapeutic agent described herein. The nucleic acid encoding the therapeutic agent described herein may be incorporated into a gene construct as part of a gene therapy regimen to deliver the nucleic acid, which may be used for intracellular expression and production of the therapeutic agent. Expression constructs of such components may be administered in any therapeutically effective carrier, such as any formulation or composition capable of effectively delivering the component gene in vivo to cells. Methods include inserting the subject gene into a viral vector, including recombinant retroviruses, adenoviruses, adeno-associated viruses, lentiviruses, and herpes simplex virus-1 (HSV-1), or recombinant bacterial or eukaryotic plasmids. The viral vector may be directly transfected into cells; plasmid DNA may be delivered using, for example, cationic liposomes (liposomal transfection agents) or derived polylysine conjugates, Gramin S, artificial viral envelopes, or other such intracellular carriers, or directly injected with the gene construct or CaPO4 precipitation. Examples of suitable retroviruses include adenovirus-derived vectors, adeno-associated viruses (AAV), pLJ, pZIP, pWE, and pEM, which are known to those skilled in the art.

[0257] In some embodiments, the composition may be formulated for storage at temperatures below 0°C (e.g., -20°C or -80°C). In some embodiments, the composition may be formulated for storage at 2-8°C (e.g., 4°C) for up to 2 years (e.g., one month, two months, three months, four months, five months, six months, seven months, eight months, nine months, ten months, eleven months, one year, or two years). Therefore, in some embodiments, the composition described herein is stable when stored at 2-8°C (e.g., 4°C) for at least one year.

[0258] Pharmaceutical compositions may contain a therapeutically effective amount of the therapeutic agent described herein. Such effective amounts can be readily determined by those skilled in the art. A therapeutically effective amount is the amount by which any toxic or harmful effects of the composition outweigh the beneficial therapeutic effects. In some embodiments, the dosage may also be selected to reduce or avoid the production of antibodies or other host immune responses against the therapeutic agent. Those skilled in the art will understand that data obtained from cell culture assays and animal studies can be used to formulate dosage ranges for human use. In various embodiments, the amount of active ingredient contained in the pharmaceutical composition allows for the administration of an appropriate dose within a specified range to a subject. The dosage and method of administration may depend on the patient's weight, age, condition, and other characteristics, and may be appropriately selected as needed by those skilled in the art.

[0259] Pharmaceutical compositions, including certain therapeutic agents such as therapeutic antibodies, may be administered at a fixed dose or at a dose of milligrams per kilogram (mg / kg). While not intended to be limiting, exemplary single doses of certain pharmaceutical compositions described herein may include certain therapeutic agents as described herein, in amounts equivalent to, for example, 0.001 mg / kg to 1000 mg / kg, 1-1000 mg / kg, 1-100 mg / kg, 0.5-50 mg / kg, 0.1-100 mg / kg, 0.5-25 mg / kg, 1-20 mg / kg, and 1-10 mg / kg body weight. Exemplary doses of the compositions described herein include, but are not limited to, 0.1 mg / kg, 0.5 mg / kg, 1 mg / kg, 2 mg / kg, 4 mg / kg, 8 mg / kg, or 20 mg / kg. This disclosure is not limited to such ranges or doses.

[0260] This disclosure further includes methods for preparing the pharmaceutical compositions of this disclosure and kits comprising the pharmaceutical compositions of this disclosure.

[0261] In various embodiments, the therapeutic agents of this disclosure may be administered to a subject, and the treatment process may further include the administration of one or more other therapeutic agents or non-therapeutic agents (e.g., surgery or radiation). Combination therapies of this disclosure may include therapeutic agents that simultaneously expose a subject to two or more treatment options.

[0262] In some embodiments, the therapeutic agent as described herein may be administered together with other agents or therapies (e.g., simultaneously and / or with the same composition). In some embodiments, the therapeutic agent of this disclosure may be administered separately from additional therapeutic agents or therapies (e.g., at a different time or with a different composition than the additional therapeutic agent or therapy). The dosing regimen of the therapeutic agent and one or more additional therapeutic agents administered therewith may be determined in a coordinated manner or independently. In various embodiments, as described herein, additional therapeutic agents or therapies administered in combination with the therapeutic agent may be administered simultaneously with the therapeutic agent, on the same day as the therapeutic agent, or in the same week as the therapeutic agent. In various embodiments, additional therapeutic agents or therapies may be administered in combination with the therapeutic agent as described herein, such that the interval between administration of the therapeutic agent and the additional therapeutic agent or therapy is one or more hours before and after administration of the therapeutic agent, one or more days before and after administration of the therapeutic agent, one or more weeks before and after administration of the therapeutic agent, or one or more months before and after administration of the therapeutic agent. In various embodiments, the frequency and / or dosage of administration of one or more additional therapeutic agents may be the same as, similar to, or different from the frequency of administration of the therapeutic agent. In some embodiments, two or more regimens may be administered simultaneously; in some embodiments, such regimens may be administered sequentially (e.g., all “doses” of the first regimen are administered before any dose of the second regimen is administered); in some embodiments, such therapeutic agents are administered in an overlapping dosing regimen.

[0263] In some implementations, the subject to which the therapeutic agent is administered may be a subject who has previously received, is scheduled to receive, or is currently receiving a treatment regimen including additional cancer therapies. In some cases, combining the administration of one therapeutic agent may improve the delivery or efficacy of another therapeutic agent or therapy.

[0264] It is believed that combination therapy can exhibit a synergistic and / or greater additive effect when a therapeutic agent is administered in combination with one or more additional therapeutic agents. The therapeutic agent can be administered at any independently determined effective amount, or at any effective amount determined by the combined effect of the administered therapeutic agent with one or more additional therapeutic agents or therapies. In some embodiments, administration of the therapeutic agent may reduce the therapeutically effective dose, required dose, or administered dose of the additional therapeutic agent or therapy relative to a reference administration regimen of the additional therapeutic agent or therapy or a therapy without the therapeutic agent. In some embodiments, the compositions described herein may replace or enhance other previously or currently administered therapies. For example, after treatment with the therapeutic agent, administration of one or more additional therapeutic agents or therapies may be stopped or reduced, for example, administered at a lower level.

[0265] Reagent test kit This disclosure includes kits for detecting modifications and / or accessibility at one or more genomic loci. In some embodiments, this disclosure provides kits for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic loci. Kits of this disclosure may include, for example, reagents for detecting and quantifying histone modifications, such as buffers and / or antibodies. In some embodiments, kits of this disclosure may include at least one antibody that selectively binds to histone modifications selected from H3K9ac, H3K14ac, H3K18ac, H3K23ac, H3K27ac, H3K4me1, H3K4me2, or H3K4me3, or panacetylation. In some embodiments, kits of this disclosure may include at least one antibody that selectively binds to H3K4me3 modifications. In some embodiments, kits of this disclosure may include at least one antibody that selectively binds to H3K27ac modifications. Kits of this disclosure may include explanatory material that discloses or describes the use of the kit in methods for determining the HER2 status and / or treatment disclosed herein. In various embodiments, the kits disclosed herein may include one or more therapeutic agents that can be used to treat cancer, for example, optionally in combination with illustrative materials for the treatment of HER2-based cancers (e.g., breast cancer, gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc.), as disclosed herein.

[0266] In some embodiments, the kits disclosed herein include reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic sites, wherein one or more genomic sites are selected from Tables 1 to 3, and optionally one or more genomic sites are not derived from HER2 amplicon.

[0267] In some embodiments, the kit includes reagents for quantifying H3K4me3 at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 1, optionally one or more of these genomic loci are not derived from HER2 amplicons. In some embodiments, the kit includes reagents for quantifying H3K27ac at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 2, optionally one or more of these genomic loci are not derived from HER2 amplicons. In some embodiments, the kit includes one or more antibodies for ChIP-seq, optionally wherein said one or more antibodies specifically bind to histones modified with H3K4me3 or H3K27ac.

[0268] In some embodiments, the kit includes reagents for quantifying DNA methylation at at least 5, 10, 20, 30, 40, or 50 genomic sites in Table 3, optionally one or more of which are not derived from the HER2 amplicon. In some embodiments, the kit includes one or more methyl-binding domains for MBD-seq.

[0269] In some embodiments, the kit includes reagents for isolating cell-free DNA (cfDNA) from liquid biopsy samples. In some embodiments, the kit includes reagents for sequencing library preparation. In some embodiments, the kit includes reagents for sequencing. In some embodiments, the kit includes instructions for determining whether a subject has HER2-positive cancer.

[0270] system This disclosure includes systems for detecting modifications and / or accessibility at one or more genomic loci. In some embodiments, this disclosure provides systems for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic loci. Systems of this disclosure may include: a sequencer configured to generate sequencing datasets from a sample; and a non-transient computer-readable storage medium and / or a computer system.

[0271] In some embodiments, a non-transient computer-readable storage medium is encoded with a computer program containing instructions that, when executed by one or more processors, cause the one or more processors to perform operations to execute the methods of this disclosure.

[0272] In some implementations, the computer system includes a memory and one or more processors coupled to the memory, wherein the one or more processors are configured to perform the methods of this disclosure.

[0273] In some embodiments, the sequencer is configured to generate whole-genome sequencing (WGS) datasets from a sample. In some embodiments, the system also includes a sample preparation device configured to prepare a sample for sequencing from a biological sample (optionally, a liquid biopsy sample). The sample preparation device may include reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic sites in cell-free DNA (cfDNA) from the biological sample (optionally, a liquid biopsy sample).

[0274] The systems disclosed herein may include, for example, reagents such as buffers and / or antibodies, for detecting and quantifying histone modifications. In some embodiments, the systems disclosed herein may include at least one antibody that selectively binds to histone modifications selected from H3K9ac, H3K14ac, H3K18ac, H3K23ac, H3K27ac, H3K4me1, H3K4me2, or H3K4me3, or panacetylation. In some embodiments, the systems disclosed herein may include at least one antibody that selectively binds to H3K4me3 modifications. In some embodiments, the systems disclosed herein may include at least one antibody that selectively binds to H3K27ac modifications. The systems disclosed herein may include explanatory material that discloses or describes the use of the system in methods for determining the HER2 status and / or treatment disclosed herein.

[0275] In some embodiments, the system disclosed herein includes reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic sites, wherein one or more genomic sites are selected from Tables 1 to 3, and optionally one or more genomic sites are not derived from HER2 amplicon.

[0276] In some embodiments, the system includes reagents for quantifying H3K4me3 at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 1, optionally one or more of these genomic loci are not derived from HER2 amplicons. In some embodiments, the system includes reagents for quantifying H3K27ac at at least 5, 10, 20, 30, 40, or 50 genomic loci in Table 2, optionally one or more of these genomic loci are not derived from HER2 amplicons. In some embodiments, the system includes one or more antibodies for ChIP-seq, optionally wherein said one or more antibodies specifically bind to histones modified with H3K4me3 or H3K27ac.

[0277] In some embodiments, the system includes reagents for quantifying DNA methylation at at least 5, 10, 20, 30, 40, or 50 genomic sites in Table 3, optionally one or more of which are not derived from the HER2 amplicon. In some embodiments, the system includes one or more methyl-binding domains for MBD-seq.

[0278] In some embodiments, the system includes reagents for isolating cell-free DNA (cfDNA) from liquid biopsy samples. In some embodiments, the sequencer includes reagents for preparing sequencing libraries. In some embodiments, the sequencer includes reagents for sequencing. In some embodiments, the system includes instructions for determining whether a subject has HER2-positive cancer.

[0279] The foregoing description of illustrative embodiments of the systems and methods disclosed herein refers to computations performed locally by a computing device. However, computations performed over a network are also contemplated. Figure 16 An illustrative network environment 1600 is shown for use in the methods and systems described herein. In a brief overview, reference is now made to... Figure 16 A block diagram illustrating an illustrative cloud computing environment 1600 is shown and described. The cloud computing environment 1600 may include one or more resource providers 1602a, 1602b, 1602c (collectively referred to as 1602). Each resource provider 1602 may include computing resources. In some implementations, computing resources may include any hardware and / or software for processing data. For example, computing resources may include hardware and / or software capable of executing algorithms, computer programs, and / or computer applications. In some embodiments, the illustrative computing resources may include application servers and / or databases with storage and retrieval capabilities. Each resource provider 1602 may be connected to any other resource provider 1602 in the cloud computing environment 1600. In some implementations, resource providers 1602 may be connected via a computer network 1608. Each resource provider 1602 may be connected via the computer network 1608 to one or more computing devices 1604a, 1604b, 1604c (collectively referred to as 1604).

[0280] The cloud computing environment 1600 may include a resource manager 1606. The resource manager 1606 can be connected to resource providers 1602 and computing devices 1604 via a computer network 1608. In some implementations, the resource manager 1606 may facilitate one or more resource providers 1602 to provide computing resources to one or more computing devices 1604. The resource manager 1606 may receive requests for computing resources from a particular computing device 1604. The resource manager 1606 may identify one or more resource providers 1602 capable of providing the computing resources requested by the computing device 1604. The resource manager 1606 may select a resource provider 1602 to provide computing resources. The resource manager 1606 may facilitate a connection between the resource provider 1602 and the particular computing device 1604. In some implementations, the resource manager 1606 may establish a connection between a particular resource provider 1602 and the particular computing device 1604. In some implementations, resource manager 1606 may redirect a particular computing device 1604 to a particular resource provider 1602 that has requested computing resources.

[0281] Figure 17Examples of computing devices 1700 and mobile computing devices 1750 that can be used in the methods and systems described in this disclosure are shown. Computing device 1700 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Mobile computing device 1750 is intended to represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are intended only as examples and not as limitations.

[0282] The computing device 1700 includes a processor 1702, a memory 1704, a storage device 1706, a high-speed interface 1708 connected to the memory 1704 and a plurality of high-speed expansion ports 1710, and a low-speed interface 1712 connected to a low-speed expansion port 1714 and the storage device 1706. Each of the processor 1702, memory 1704, storage device 1706, high-speed interface 1708, high-speed expansion port 1710, and low-speed interface 1712 is interconnected using various buses and may be mounted on a common motherboard or otherwise mounted as appropriate. The processor 1702 can process instructions (including instructions stored in the memory 1704 or storage device 1706) for execution within the computing device 1700 to display graphical information of a GUI on an external input / output device, such as a display 1716 coupled to the high-speed interface 1708. In other embodiments, multiple processors and / or multiple buses may be used together with multiple memories and various types of memory as appropriate. Additionally, multiple computing devices can be connected, each providing a portion of the necessary operation (e.g., as a server group, a set of blade servers, or a multiprocessor system). Therefore, as used herein, when multiple functions are described as being performed by a “processor,” this covers implementations where the multiple functions are performed by any number of processors (e.g., one or more processors) of any number of computing devices (e.g., one or more computing devices). Furthermore, when a function is described as being performed by a “processor,” this covers implementations where said function is performed by any number of processors (e.g., one or more processors) of any number of computing devices (e.g., one or more computing devices) (e.g., in a distributed computing system).

[0283] Memory 1704 stores information within computing device 1700. In some implementations, memory 1704 is one or more volatile memory cells. In some embodiments, memory 1704 is one or more non-volatile memory cells. Memory 1704 may also be another form of computer-readable medium, such as a magnetic disk or optical disk.

[0284] Storage device 1706 provides mass storage for computing device 1700. In some implementations, storage device 1706 may be or contain computer-readable media, such as hard disk drives, optical disk drives, flash memory or other similar solid-state storage devices, or device arrays, including those in a storage area network or other configuration. Instructions may be stored in an information carrier. When executed by one or more processing devices (e.g., processor 1702), the instructions are performed in one or more ways, such as those described above. Instructions may also be stored by one or more storage devices, such as computer or machine-readable media (e.g., memory 1704, storage device 1706, or memory on processor 1702).

[0285] High-speed interface 1708 manages bandwidth-intensive operations of computing device 1700, while low-speed interface 1712 manages lower bandwidth-intensive operations. This allocation of functions is merely an example. In some embodiments, high-speed interface 1708 is coupled to memory 1704, display 1716 (e.g., via a graphics processor or accelerator), and to high-speed expansion port 1710, which accepts various expansion cards (not shown). In embodiments, low-speed interface 1712 is coupled to storage device 1706 and low-speed expansion port 1714. Low-speed expansion port 1714, which may include various communication ports (e.g., USB, Bluetooth®, Ethernet, Wireless Ethernet), may be coupled to one or more input / output devices, such as keyboards, pointing devices, scanners, or networking devices (such as switches or routers), for example, via a network adapter.

[0286] The computing device 1700 can be implemented in a variety of different forms, as shown in the figure. For example, it can be implemented as a standard server 1720 or multiple times in a group of such servers. Furthermore, it can be implemented in a personal computer, such as a laptop computer 1722. It can also be implemented as part of a rack server system 1724. Alternatively, components from the computing device 1700 can be combined with other components in a mobile device (not shown) (such as a mobile computing device 1750). Each of such devices may contain one or more of the computing device 1700 and the mobile computing device 1750, and the entire system may consist of multiple computing devices communicating with each other.

[0287] Mobile computing device 1750 includes a processor 1752, memory 1764, input / output devices (such as a display 1754), a communication interface 1766, and a transceiver 1768, as well as other components. Mobile computing device 1750 may also be provided with storage devices, such as microdrives or other devices, to provide additional storage. Each of the processor 1752, memory 1764, display 1754, communication interface 1766, and transceiver 1768 is interconnected using various buses, and some of the components may be mounted on a common motherboard or otherwise mounted as appropriate.

[0288] Processor 1752 executes instructions within mobile computing device 1750, including instructions stored in memory 1764. Processor 1752 may be implemented as a chipset comprising individual and multiple analog and digital processors. Processor 1752 may provide coordination, for example, with other components of mobile computing device 1750, such as control of the user interface, application execution performed by mobile computing device 1750, and wireless communication performed by mobile computing device 1750.

[0289] Processor 1752 can communicate with the user via control interface 1758 and display interface 1756 coupled to display 1754. Display 1754 can be, for example, a TFT (Thin Film Transistor Liquid Crystal Display) or OLED (Organic Light Emitting Diode) display, or other suitable display technology. Display interface 1756 may include appropriate circuitry for driving display 1754 to present graphics and other information to the user. Control interface 1758 can receive commands from the user and translate them for submission to processor 1752. Furthermore, external interface 1762 can provide communication with processor 1752 to enable near-area communication between mobile computing device 1750 and other devices. External interface 1762 may provide, in some embodiments, wired communication, or in other embodiments, wireless communication, and multiple interfaces may be used.

[0290] Memory 1764 stores information within mobile computing device 1750. Memory 1764 may be implemented as one or more computer-readable media, one or more volatile memory cells, or one or more non-volatile memory cells. Extended memory 1774 may also be provided and connected to mobile computing device 1750 via an extended interface 1772, which may include, for example, a SIMM (Single In-line Memory Module) card interface. Extended memory 1774 may provide additional storage space for mobile computing device 1750, or it may store applications or other information for mobile computing device 1750. Specifically, extended memory 1774 may include instructions for executing or supplementing the processes described above, and may also include security information. Thus, for example, extended memory 1774 may be provided as a security module of mobile computing device 1750 and may be programmed with instructions that allow secure use of mobile computing device 1750. Furthermore, secure applications and additional information, such as placing identification information on the SIMM card in an unbreakable manner, may be provided via a SIMM card.

[0291] The memory may include, for example, flash memory and / or NVRAM (non-volatile random access memory), as discussed below. In some embodiments, instructions are stored in an information carrier and, when executed by one or more processing devices (e.g., processor 1752), perform one or more methods, such as those described above. Instructions may also be stored by one or more storage devices, such as one or more computer or machine-readable media (e.g., memory 1764, extended memory 1774, or memory on processor 1752). In some embodiments, instructions may be received in a propagated signal, for example, via transceiver 1768 or external interface 1762.

[0292] Mobile computing device 1750 can wirelessly communicate via communication interface 1766, which may include digital signal processing circuitry if necessary. Communication interface 1766 can provide communication under various modes or protocols, such as GSM voice calls (Global System for Mobile Communications), SMS (Short Message Service), EMS (Enhanced Messaging Service) or MMS messages (Multimedia Messaging Service), CDMA (Code Division Multiple Access), TDMA (Time Division Multiple Access), PDC (Personal Digital Cellular), WCDMA (Wideband Code Division Multiple Access), CDMA2000, or GPRS (General Packet Radio Service). For example, such communication can occur using radio frequency via transceiver 1768. Additionally, short-range communication may occur, such as using Bluetooth®, Wi-Fi™, or other transceivers (not shown). Furthermore, GPS (Global Positioning System) receiver module 1770 can provide additional navigation and location-related wireless data to mobile computing device 1750, which can be used by applications running on mobile computing device 1750 as appropriate.

[0293] The mobile computing device 1750 can also use an audio codec 1760 for audible communication, which receives voice information from a user and converts it into usable digital information. The audio codec 1760 can also generate audible sounds for the user, such as through a speaker, for example, in the handset of the mobile computing device 1750. Such sounds may include sounds from voice telephone calls, recorded sounds (e.g., voice messages, music files, etc.), and sounds generated by applications running on the mobile computing device 1750.

[0294] Mobile computing device 1750 can be implemented in many different forms, as shown in the figure. For example, it can be implemented as a cellular phone 1780. It can also be implemented as part of a smartphone 1782, a personal digital assistant, or other similar mobile device.

[0295] Various implementations of the systems and techniques described herein can be implemented in digital electronic circuits, integrated circuits, specially designed ASICs (Application-Specific Integrated Circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementations in one or more computer programs executable and / or interpretable on a programmable system including at least one programmable processor, which may be dedicated or general-purpose, coupled to receive data and instructions therefrom, and to transfer data and instructions to a storage system, at least one input device, and at least one output device.

[0296] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and can be implemented using high-level programming and / or object-oriented programming languages ​​and / or assembly / machine language. As used herein, the terms machine-readable medium and computer-readable medium refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) used to provide machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term machine-readable signal refers to any signal used to provide machine instructions and / or data to a programmable processor.

[0297] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and pointing device (e.g., a mouse or trackball) for the user to provide input to the computer. Other types of devices may also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including sound, speech, or tactile input.

[0298] The systems and technologies described herein can be implemented in computing systems that include back-end components (e.g., data servers), middleware components (e.g., application servers), front-end components (e.g., client computers with graphical user interfaces or web browsers through which clients can interact with implementations of the systems and technologies described herein), or any combination of such back-end, middleware, or front-end components. Components of the system can be interconnected via digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0299] A computing system may include clients and servers. Clients and computers are generally remote to each other and typically interact via communication networks. The client-server relationship is established by computer programs running on their respective computers that establish a client-server relationship between them.

[0300] Some embodiments described herein utilize computer algorithms in the form of software instructions executed by a computer processor, such as in a classifier. In some embodiments, the software instructions include machine learning (ML) modules, for example, as a classifier. As used herein, a machine learning module refers to a computer implementation process (e.g., software function) that implements one or more specific machine learning techniques, such as artificial neural networks (ANNs), such as convolutional neural networks (CNNs), random forests, decision trees, support vector machines, etc., to determine one or more output values ​​for a given input. In some embodiments, the input includes image data and / or alphanumeric data, which may include, for example, 2D and / or 3D datasets, numbers, words, phrases, or longer strings. In some embodiments, the one or more output values ​​include image data (e.g., 2D and / or 3D datasets) and / or values ​​representing numerical values, words, phrases, or other alphanumeric strings.

[0301] In some implementations, a machine learning module implementing machine learning techniques is trained, for example, using a dataset that includes the data categories described herein. Such training can be used to determine various parameters of the machine learning algorithm implemented by the machine learning module, such as weights associated with layers in a neural network. In some implementations, once the machine learning module is trained, for example, to perform a specific task, such as identifying certain response strings, the values ​​of the parameters are fixed (e.g., immutable, static), and the machine learning module is used to process new data (e.g., different from the training data) and complete its trained task without further updating its parameters (e.g., the machine learning module does not receive feedback and / or updates). In some implementations, the available input data includes training data and validation data, for example, validation data is separate from and does not overlap with training data. For example, in some implementations, training data is used to optimize the model during the training process, while validation data is used to check the accuracy of the model when working with previously unseen data. In some implementations, the training data is divided into batches (e.g., portions) and then used sequentially (e.g., in random order) as the input set for training the model. In some implementations, the model is trained multiple times (e.g., multiple iterations) on the entire training dataset. In some implementations, the machine learning module may receive feedback, such as user reviews of accuracy, and this feedback can be used as additional training data to dynamically update the machine learning module. In some implementations, two or more machine learning modules may be combined and implemented as a single module and / or a single software application. In some implementations, two or more machine learning modules may also be implemented separately, for example, as separate software applications. Machine learning modules can be software and / or hardware. For example, a machine learning module may be implemented entirely as software, or certain functionalities of an ANN module may be implemented via dedicated hardware (e.g., via application-specific integrated circuits (ASICs) and / or field-programmable gate arrays (FPGAs)).

[0302] In some implementations, a machine learning module implementing machine learning techniques may consist of individual nodes (e.g., units, neurons). A node may receive a set of inputs, which may include at least a portion of the given input data for the machine learning module and / or at least one output from another node. A node may have at least one parameter to apply and / or a set of instructions to execute on the set of inputs (e.g., a mathematical function to be performed). In some implementations, node instructions may include a step of assigning various relative importances to a set of inputs using various parameters (such as weights). Weights can be applied by performing a scalar multiplication (e.g., or other mathematical function) on a set of input values ​​with the parameters, resulting in a weighted set of inputs. In some implementations, a node may have a transfer function that combines a set of weighted inputs into an output value. The transfer function can be implemented by summing all the weighted inputs and adding an offset (e.g., bias) value. In some implementations, a node may have an activation function to introduce non-linearity into the output value. Non-limiting examples of activation functions include modified linear activation (ReLU), logistic (e.g., sigmoid), hyperbolic tangent (tanh), and softmax. In some implementations, nodes may have the ability to remember previous states (e.g., cyclic nodes). A set of learning parameters can be used to apply the previous state to the input and output values.

[0303] In some implementations, the machine learning module comprises a deep learning architecture consisting of nodes organized into layers. For example, a layer is a set of nodes that receives data input (e.g., weighted or unweighted input), transforms it (e.g., by executing instructions, such as applying a set of functions, e.g., linear and / or nonlinear functions), and passes the transformed value as output (e.g., to the next layer). In some implementations, a set of nodes within a particular layer may share the same parameters and instructions without interacting with each other. The machine learning module may consist of at least one layer (e.g., an ordered layer). Instances of layer types include convolutional layers (e.g., layers with kernels, where the kernel is a parameter matrix that slides over the input and multiplies multiple input values, thus simplifying them to a single output value); fully connected (FC) layers (e.g., all nodes are connected to all outputs of the previous layer); recurrent layers, long short-term memory (LSTM) layers, gated recurrent unit (GRU) layers (e.g., nodes with various abilities to remember and apply their previous inputs and / or outputs); batch normalization (BN) layers (e.g., layers that normalize a set of outputs from another layer, allowing individual layers to learn more independently); activation layers (e.g., layers where nodes contain only activation functions); and / or (non-)pooling layers [e.g., layers that reduce (increase) the input dimension by summing (splitting) the input values ​​into defined blocks].

[0304] In some implementations, the performance of a machine learning module can be characterized by its ability to produce output data with a specific accuracy. To achieve this specific accuracy, a training process is performed to find the optimal parameters (such as weights) for each node in each layer of the machine learning module. In some implementations, the training process of the machine learning module may involve using the output data to compute an objective function (e.g., a cost function, loss function, error function) that needs to be optimized (e.g., minimized, maximized). For example, the machine learning objective function can be a combination of a loss function and regularization parameters. The loss function relates to how well the output can predict the input. Loss functions can take many forms, such as mean squared error, mean absolute error, binary cross-entropy, and classification cross-entropy. Regularization terms may be needed to prevent overfitting and improve the generalization of the training process. Examples of regularization techniques include L1 regularization or Lasso regression, L2 regularization or Ridge regression, and Dropout (e.g., randomly discarding layer outputs during the training process).

[0305] In some implementations, the objective function optimization of the machine learning module may involve finding at least one (e.g., all) current global optima (e.g., rather than local optima). In some implementations, the objective function optimization algorithm follows the mathematical optimization principles of multivariable functions and depends on the specific precision of the implementation process. Examples of objective function optimization algorithms include gradient descent, nonlinear conjugate gradient, stochastic search, Levenberg-Marquardt method, limited-memory Broyden-Fietcher-Goldfarb-Shanno algorithm, pattern search, basin jump, Krylov method, Adam method, genetic algorithm, particle swarm optimization, surrogate optimization, and simulated annealing.

[0306] The methods disclosed in this paper can use one or more machine learning models as classifiers. Machine learning models can be or include artificial neural networks. Machine learning models can take the form of, for example: attention-based models (e.g., transformer models, such as visual transformers), regression-based models (e.g., logistic regression models), regularization-based models (e.g., elastic network models or ridge regression models), instance-based models (e.g., support vector machines or k-nearest neighbors models), Bayesian-based models (e.g., naive-based models or Gaussian naive models), clustering-based models (e.g., expectation-maximization models), ensemble-based models (e.g., adaptive boosting models, random forest models, bootstrap aggregation models, or gradient boosting machine models), or neural network-based models (e.g., convolutional neural networks, recurrent neural networks, autoencoders, backpropagation networks, or stochastic gradient descent networks).

[0307] In some implementations, the machine learning model used as the classifier is derived from decision tree methods, neural boosting methods, bootstrap forest methods, boosting tree methods, k-nearest neighbor methods, generalized regression forward selection methods, generalized regression pruning forward selection methods, stepwise fitting methods, generalized regression lasso methods, generalized regression elastic network methods, generalized regression ridge methods, nominal logistic methods, support vector machine methods, discriminative methods, naive Bayes methods, or combinations thereof. In some implementations, the machine learning model is derived from decision tree methods, neural boosting methods, bootstrap forest methods, boosting tree methods, generalized regression lasso methods, generalized regression elastic network methods, generalized regression ridge methods, nominal logistic methods, support vector machine methods, discriminative methods, or combinations thereof. In some implementations, the machine learning model is derived from decision tree methods, neural boosting methods, bootstrap forest methods, boosting tree methods, support vector machine methods, or combinations thereof.

[0308] definition "One (A or An)": The articles “a” and “an” are used in this text to refer to one or more ( Right now (This refers to at least one) grammatical object of an article. For example, "one element" means one element or more elements.

[0309] about:When used herein to refer to a value, the term "about" refers to a value similar to the value being referred to in the context. Generally, those skilled in the art will understand the degree of difference covered by "about" in that context. For example, in some embodiments, the term "about" may cover a range of values ​​up to a fraction of 25%, 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, or one percent of the value being referred to.

[0310] "Accessibility status" or "chromatin accessibility status": As used herein, the “accessibility status” or “chromatin accessibility status” of a genomic locus refers to the frequency at which a DNA sequence corresponding to a genomic locus is identified in an accessibility chromatin assay. Accessibility status can be determined by various assays known in the art, including but not limited to ChIP-seq as an example. Differences in the chromatin accessibility status of a genomic locus can be detected when two samples are analyzed individually using the same or comparable assays to detect accessible DNA sequences. Accessibility status can be compared to a standard or reference. If the accessibility status of a sample differs from that of a standard or reference, it can be termed differentially modified. Various assays suitable for determining chromatin accessibility are known in the art. Exemplary assays include ATAC-seq (transposon accessible chromatin sequencing assay), NOMe-seq (nucleosome occupancy and methylome sequencing), FAIRE-seq (formaldehyde-assisted separation of regulatory elements sequencing), MNase-seq (micrococcal nuclease digestion sequencing), and / or DNase hypersensitivity assays.

[0311] Application: As used herein, the term "application" generally refers to the application of appropriate (e.g., appropriate for HER2-positive cancer) treatment to a disease. In some embodiments, appropriate disease treatment may include the administration of a composition to a subject, for example, to achieve the delivery of an agent that is a component of the composition, an agent contained in the composition, or an agent otherwise delivered by the composition. In some embodiments, appropriate disease treatment may include the application of appropriate surgical or radiation therapy, optionally in combination with the application of the composition.

[0312] Agent: As used herein, the term "agent" may refer to any chemical or physical entity, including but not limited to atoms. ( For example, any one or more of the following: radioactive atoms, molecules, compounds, conjugates, polypeptides, polynucleotides, polysaccharides, lipids, cells, or combinations or complexes thereof.

[0313] AntibodyAs used herein, the term "antibody" refers to a polypeptide (e.g., a heavy chain variable domain, a light chain variable domain, and / or one or more CDRs) comprising one or more typical immunoglobulin sequence elements sufficient to confer specific binding to a particular antigen. Therefore, the term antibody includes, but is not limited to, human antibodies, non-human antibodies, synthetic and / or engineered antibodies, fragments thereof, and agents comprising the aforementioned antibodies. Antibodies can be naturally occurring immunoglobulins (e.g., immunoglobulins produced by an organism in response to an antigen). Synthetic, non-natural, or engineered antibodies can be produced by recombinant engineering, chemical synthesis, or other artificial systems or methods known to those skilled in the art.

[0314] As is well known in the art, a typical human immunoglobulin is a tetramer of approximately 150 kDa, comprising two identical heavy (H) chain polypeptides (each approximately 50 kDa) and two identical light (L) chain polypeptides (each approximately 25 kDa), which associate with each other to form a structure commonly referred to as a “Y-shape.” Typically, each heavy chain contains a heavy chain variable domain (VH) and a heavy chain constant domain (CH). The heavy chain constant domain contains three CH domains: CH1, CH2, and CH3. Short regions called “switches” connect the heavy chain variable and constant regions. “Hinges” connect the CH2 and CH3 domains to the rest of the immunoglobulin. Each light chain contains a light chain variable domain (VL) and a light chain constant domain (CL), separated by another “switch.” Each variable domain contains three hypervariable loops (CDR1, CDR2, and CDR3) called “complementarity-determining regions” and four slightly invariant “framework” regions (FR1, FR2, FR3, and FR4). In each VH and VL, the three CDRs and four FRs are arranged from the amino terminus to the terminal terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. It is generally assumed that the variable regions of the heavy and / or light chains provide the binding portions for interaction with antigens. The invariant domains can mediate the binding of antibodies to various immune system cells (e.g., effector cells and / or cells mediating cytotoxicity), receptors, and complement system elements. The heavy and light chains are linked to each other by a single disulfide bond, and two additional disulfide bonds link the heavy chain hinge regions together, allowing the dimers to link together and form a tetramer. When innate immunoglobulins fold, the FR region forms a β-sheet that provides a structural framework for the domain, and the CDR loop regions from both the heavy and light chains aggregate together in three-dimensional space to create a single hypervariable antigen-binding site at the top of the Y structure.

[0315] In some embodiments, the antibody is a polyclonal antibody, a monoclonal antibody, a monospecific antibody, or a multispecific antibody (e.g., a bispecific antibody). In some embodiments, the antibody comprises at least one light chain monomer or dimer, at least one heavy chain monomer or dimer, at least one heavy chain-light chain dimer, or a tetramer comprising two heavy chain monomers and two light chain monomers. Furthermore, the term "antibody" can include (unless otherwise stated or the context clearly indicates) any prior art known construct or form using antibody structural and / or functional characteristics, including but not limited to internal antibodies, domain antibodies, antibody mimics, Zybodies®, Fab fragments, Fab' fragments, F(ab')2 fragments, Fd' fragments, Fd fragments, isolated CDRs or collections thereof, single-chain antibodies, single-chain Fv (scFv), disulfide-linked Fv. (sdFv), peptide-Fc fusion proteins, single-domain antibodies (e.g., shark single-domain antibodies, such as IgNAR or fragments thereof), camel antibodies, camelified antibodies, masking antibodies (e.g., Probodies®), Affybodies, anti-idiotype (anti-Id) antibodies (including, for example, anti-anti-Id antibodies), small modular immunotherapies (SMIPs), single-chain or tandem biantibodies (TandAb®), VHH, Anticalins®, Nanobodies®, minibodies, BiTE®, ankylosing repeats or DARPINs®, Avimers®, DARTs, TCR-like antibodies, Adnectins®, Affilins®, Trans-bodies®, Affibodies®, TrimerX®, MicroProteins, Fynomers®, Centyrins®, KALBITOR®, chimeric antigen receptors (CARs), engineered T-cell receptors (TCRs), and antigen-binding fragments of any of the above.

[0316] In various embodiments, the antibody includes one or more structural elements that are recognized by those skilled in the art as complementarity-determining regions (CDRs) or variable domains. In some embodiments, the antibody may be a covalently modified (“conjugated”) antibody (e.g., an antibody comprising a polypeptide containing one or more typical immunoglobulin sequence elements sufficient to confer specific binding to a particular antigen, wherein the polypeptide is covalently linked to one or more therapeutic agents, a detectable moiety, another polypeptide, a glycan, or a polyethylene glycol molecule). In some embodiments, the antibody sequence elements are humanized, primate-derived, chimeric, etc., as is known in the art.

[0317] Antibodies containing a heavy chain constant domain can be, but are not limited to, any known class of antibodies, including but not limited to IgA, secretory IgA, IgG, IgE, and IgM, classified based on the amino acid sequence of the heavy chain constant domain (e.g., alpha (α), delta (δ), epsilon (ε), gamma (γ), and muon (μ)). IgG subclasses are also well known to those skilled in the art and include, but are not limited to, human IgG1, IgG2, IgG3, and IgG4. "Isotype" refers to an Ab class or subclass (e.g., IgM or IgG1) encoded by a heavy chain constant region gene. As used herein, "light chain" can be classified into different types based on the amino acid sequence of the light chain constant domain, such as kappa (κ) or lamuda (λ). In some embodiments, antibodies have constant region sequences that are characteristic of mouse, rabbit, primate, or human immunoglobulins. Naturally occurring immunoglobulins are typically glycosylated at the CH2 domain. As is known in the art, the affinity and / or other binding properties of the Fc region to the Fc receptor can be modulated by glycosylation or other modifications. In some embodiments, the antibody may lack the covalent modifications (e.g., glycan attachment) that it would have naturally. In some embodiments, the antibody generated and / or utilized according to the invention comprises a glycosylated Fc domain, which includes a modified or engineered glycosylated Fc domain.

[0318] In some embodiments, the antibody may be specific to a particular histone modification (e.g., under conditions commonly used in ChIP-seq experiments, the antibody may bind to a histone modification, such as H3K27ac, with a higher affinity than other histone modifications). In some embodiments, the antibody is specific to H3K9ac, H3K14ac, H3K18ac, H3K23ac, H3K27ac, H3K4me1, H3K4me2, or H3K4me3 modifications. In some embodiments, the antibody is specific to the H3K27ac modification. In some embodiments, the antibody is specific to the H3K4me3 modification.

[0319] In some embodiments, the antibody is a "pan-" antibody. As used herein, the term pan-antibody refers to an antibody capable of binding to a group of histone modifications having one or more similar characteristics. In some embodiments, the pan-antibody is a pan-methylating antibody (e.g., an antibody capable of binding histones, such as an H3 containing at least one methylated lysine, wherein the at least one methylated lysine can be located at any of a plurality of amino acid positions; for example, in some embodiments, a pan-methylating antibody can bind to an H3 protein containing methylated lysine at any position). In some embodiments, the pan-antibody is a pan-acetylated antibody (e.g., an antibody capable of binding histones, such as an H3 containing at least one acetylated lysine, wherein the at least one acetylated lysine can be located at any of a plurality of amino acid positions; for example, a pan-acetylated antibody can bind to an H3 protein containing acetylated lysine at any position). In some embodiments, the pan-antibody can bind to one or more histone modifications associated with transcriptional activation. In some embodiments, the pan-antibody can bind to one or more histone modifications associated with transcriptional silencing.

[0320] Antibody fragment: As used herein, "antibody fragment" refers to a portion of an antibody or antibody agent as described herein, and generally refers to a portion containing the antigen-binding portion or their variable regions. Antibody fragments can be generated in any manner. For example, in some embodiments, antibody fragments can be enzymatically or chemically generated by fragmentation of a complete antibody or antibody agent. Alternatively, in some embodiments, antibody fragments can be generated through recombinant synthesis. Right now The antibody fragment is generated through recombination via the expression of engineered nucleic acid sequences. In some embodiments, the antibody fragment can be synthesized in whole or in part. In some embodiments, the antibody fragment (especially the antigen-binding antibody fragment) can have at least about 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190 or more amino acids, and in some embodiments, at least about 200 amino acids in length.

[0321] Related:Two events or entities are “associated” with each other, as used herein, if the presence, level, and / or form of one event or entity is related to the presence, level, and / or form of another event or entity. For example, if the presence, level, and / or form of a particular entity (e.g., an epigenetic signature containing one or more histone modifications at a set of genomic loci, etc.) is associated with the incidence and / or susceptibility to a particular disease, condition, or disorder (e.g., in a relevant population), that particular entity is considered associated with that disease, condition, or disorder. In some embodiments, two or more entities are physically “associated” with each other if they interact directly or indirectly such that they are physically close to each other and / or remain physically close to each other. In some embodiments, two or more entities that are physically associated with each other are covalently linked; in some embodiments, two or more entities that are physically associated with each other are not covalently linked but non-covalently associated, for example, through hydrogen bonds, van der Waals interactions, hydrophobic interactions, magnetism, or a combination thereof.

[0322] "between" or "from": As used herein, the term “between” refers to content falling between the indicated upper and lower boundaries, or between the first and second boundaries, including the boundaries themselves. Similarly, the term “from” when used in the context of a range of values ​​indicates a range that includes content falling between the indicated upper and lower boundaries, or between the first and second boundaries, including the boundaries themselves.

[0323] biological samplesAs used herein, the term "biological sample" typically refers to a sample obtained or derived from a target biological source (e.g., tissue or organism or cells) as described herein. In some embodiments, the biological source is or includes an organism, such as a human subject. In some embodiments, the biological sample is or includes biological tissue or fluid. In some embodiments, the biological sample may be or includes cells, tissues, or body fluids. "Body fluids" refers to fluids expelled or secreted from the body, as well as fluids that are not normally expelled or secreted (e.g., blood, serum, plasma, bulbourethral fluid or pre-ejaculated semen, chyle, chyme, feces, interstitial fluid, intracellular fluid, lymph, menstrual fluid, saliva, sebum, semen, serum, sweat, synovial fluid, tears, urine, vitreous fluid, vomitus). In some embodiments, biological samples may be or include blood, blood components, cell-free DNA (cfDNA), circulating tumor DNA (ctDNA), ascites, biopsy samples, surgical specimens, cell-containing body fluids, sputum, saliva, feces, urine, cerebrospinal fluid, peritoneal fluid, pleural fluid, lymph, gynecological fluids, secretions, excretions, skin swabs, vaginal swabs, oral swabs, nasal swabs, irrigation or lavage solutions (such as catheter lavage or bronchoalveolar lavage), aspirates, scrapings, or bone marrow. In some embodiments, biological samples are liquid biopsy samples obtained from body fluids. In some embodiments, biological samples are or include DNA obtained from a single subject or multiple subjects. Biological samples may be “raw samples” obtained directly from biological sources or “processed samples.” Right now A sample derived from a raw sample, for example, through dilution, purification, mixing with one or more reagents, or any other processing steps as described herein. A biological sample may also be referred to as a "sample".

[0324] Blood components: As used herein, the term "blood components" refers to any component of whole blood, including red blood cells, white blood cells, plasma, platelets, endothelial cells, mesothelial cells, epithelial cells, cell-free DNA (cfDNA), and circulating tumor DNA (ctDNA). Blood components also include plasma components, including proteins, metabolites, lipids, nucleic acids, and carbohydrates, as well as any other cells that may be present in the blood, such as those produced due to pregnancy, organ transplantation, infection, injury, or disease.

[0325] cancer:As used herein, the terms “cancer,” “malignant tumor,” “tumor,” and “carcinoma” are used interchangeably to refer to a disease, condition, or disorder in which cells exhibit or have exhibited relatively abnormal, uncontrolled, and / or autonomous growth, resulting in or having exhibited an abnormally elevated rate of proliferation and / or an abnormal growth phenotype. In some embodiments, cancer may include one or more tumors. In some embodiments, cancer may be or comprise precancerous (e.g., benign), malignant, pre-metastatic, metastatic, and / or non-metastatic cells. In some embodiments, cancer may be or comprise solid tumors. In some embodiments, cancer may be associated with a HER2-positive status, such as HER2-positive breast cancer, gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc.

[0326] Combination therapy: As used herein, the term "combination therapy" refers to the administration of two or more therapeutic agents or regimens to a subject so that these two or more therapeutic agents or regimens together treat the subject's disease, ailment, or condition. In some embodiments, the two or more therapeutic agents or regimens may be administered simultaneously, sequentially, or in an overlapping manner. Those skilled in the art will understand that combination therapy includes, but does not require, the administration of two therapeutic agents or regimens together or simultaneously as a single composition.

[0327] Corresponding toAs used herein, the term "corresponds to" can be used to designate the position / identity of a structural element in a compound or composition by comparison with a suitable reference compound or composition. For example, in some embodiments, monomeric residues in a polymer (e.g., amino acid residues in a polypeptide or nucleic acid residues in a polynucleotide) can be identified as "corresponding to" residues in a suitable reference polymer. For example, those skilled in the art will understand that residues in a provided polypeptide or polynucleotide sequence are typically designated (e.g., numbered or labeled) according to a scheme of the relevant reference sequence (even if, for example, such designation does not reflect the literal numbering of the provided sequence). For example, if a reference sequence contains a specific amino acid motif at positions 100-110, and a second relevant sequence contains the same motif at positions 110-120, then the motif position of the second relevant sequence can be said to "correspond" to position 100-110 of the reference sequence. Those skilled in the art will understand that the corresponding positions can be easily identified, for example, by sequence alignment, which is typically achieved through any of a variety of known tools, strategies and / or algorithms, including but not limited to software programs such as, for example, BLAST, CS-BLAST, CUDASW++, DIAMOND, FASTA, GGSEARCH / GLSEARCH, Genoogle, HMMER, HHpred / HHsearch, IDF, Infernal, KLAST, USEARCH, parasail, PSI-BLAST, PSI-Search, ScalaBLAST, Sequilab, SAM, SSEARCH, SWAPHI, SWAPHI-LS, SWIMM or SWIPE. Two sequences can be identified as corresponding if they are completely identical or substantially identical, for example, at least 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical, for example, in lengths of at least 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, or more residues. In various embodiments, the nucleic acid sequence may correspond to a sequence that is identical or substantially identical (e.g., at least 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical) to a complementary sequence of the nucleic acid sequence, for example, in lengths of at least 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, or more nucleic acid residues.

[0328] "Diagnosis", "Test", "Confirmation" or "Screening":As used herein, “diagnosis,” “detection,” “determination,” and “screening” for the presence of a disease or condition (e.g., HER2-positive cancer) or related status (e.g., response of HER2-positive cancer to one or more HER2-targeted therapies) includes the acts, processes, and / or outcomes of determining whether a subject has or will have a disease, condition, or related status, and / or the acts, processes, and / or outcomes of determining their qualitative or quantitative probability. In some cases, diagnosis may include assessment of prognosis and / or the likely response to one or more general or specific therapeutic agents or regimens.

[0329] Difference accessibility: As used herein, the term “differential accessibility” describes a genomic locus whose chromatin accessibility status differs between a first condition or sample and a second condition or sample (e.g., a standard or reference). Differentially accessible genomic loci may include a measured increase or decrease in accessibility under a selected condition of interest (e.g., HER2-positive status) compared to a reference status (e.g., HER2-negative status).

[0330] Modify differently: As used herein, the term “differential modification” describes a genomic site whose histone modification status and / or DNA methylation status differs between a first condition or sample and a second condition or sample (e.g., a standard or reference). Differentially modified genomic sites in a selected state of interest (e.g., a HER2-negative state) may include an increase or decrease in the amount or frequency of histone modifications and / or DNA methylation compared to a reference state (e.g., a HER2-negative state).

[0331] Epigenetic modificationsAs used herein, the term "epigenetic modification" refers to heritable alterations to the genome that are not due to changes in the DNA sequence. Epigenetic modifications include chemical modifications such as, for example, DNA methylation and histone modifications. In some embodiments, epigenetic modifications can cause changes in chromatin structure, DNA accessibility, and / or transcription factor binding. In some embodiments, epigenetic modifications can be detected or measured directly (e.g., by using an agent that binds to the epigenetic modification (e.g., an antibody that binds to H3K4me3 or H3K27ac)). In some embodiments, epigenetic modifications can be measured indirectly, for example, by measuring or detecting one or more properties whose changes indicate changes in epigenetic modifications. For example, in some embodiments, chromatin accessibility and / or transcription factor binding can be used as a measure of epigenetic modification at a given site. As used herein, the term "epigenetic biomarker" refers to an indicator of epigenetic status and includes, for example, epigenetic modifications and assays that measure transcription factor binding or chromatin accessibility. As used herein, the term "epigenetic biomarker" refers to an epigenetic marker that can be used to detect disease or disorder.

[0332] identityAs used herein, the term "identity" refers to the overall correlation between aggregate molecules, such as nucleic acid molecules (e.g., DNA molecules) and / or polypeptide molecules. Methods for calculating the percentage of identity between two provided sequences are known in the art. The term "sequence identity %" refers to the relationship between two or more sequences, as determined by comparing sequences. In the art, "identity" also means the degree of sequence correlation between protein sequences and nucleic acid sequences, as determined by matching these sequence strings. "Identity" (often referred to as "similarity") can be easily calculated using known methods, including those described in: *Computational Molecular Biology* (edited by Lesk, AM) Oxford University Press, NY (1988); *Biocomputing: Informatics and Genome Projects* (edited by Smith, DW) Academic Press, NY (1994); *Computer Analysis of Sequence Data, Part I* (edited by Griffin, AM and Griffin, HG) Humana Press, NJ (1994); *Sequence Analysis in Molecular Biology* (edited by Von Heijne, G.) Academic Press (1987); and *Sequence Analysis Primer* (edited by Gribskov, M. and Devereux, J.) Oxford University Press, NY (1992), each of which is incorporated individually as a whole by reference. Preferred methods for identity determination aim to achieve an optimal match between the sequences being tested. Methods for determining identity and similarity have been incorporated into publicly available computer programs. For example, the percentage of identity between two nucleic acid or polypeptide sequences can be calculated by aligning the two sequences (or complementary sequences of one or both sequences) for optimal comparison purposes (e.g., vacancies can be introduced into one or both of the first and second sequences to achieve optimal alignment, and dissimilar sequences can be ignored for comparison purposes). Nucleotides or amino acids at corresponding positions are then compared. The molecules are identical at said position when a position in the first sequence is occupied by the same residue (e.g., a nucleotide or amino acid) as the corresponding position in the second sequence. Optionally, taking into account the number of gaps that need to be introduced for optimal alignment of the two sequences and the length of each gap, the percentage identity between the two sequences is related to the number of shared positions.Sequence alignment and determination of the percentage of identity between two sequences can be achieved using computational algorithms such as BLAST (Basic Local Alignment Search Tool). Sequence alignment and percentage of identity calculation can be performed using the Megalign program of the LASERGENE Bioinformatics Computation Suite (DNASTAR, Inc., Madison, Wisconsin). Multiple alignment of sequences can also be performed using the Clustal alignment method (Higgins and Sharp, Comp Appl Biosci (1989) 5(2):151-153), which is incorporated herein by reference in its entirety, using the default parameters (GAP PENALTY=10, GAP LENGTH PENALTY=10). Related programs also include the GCG program suite (Wisconsin Package Version 9.0, Genetics Computer Group (GCG), Madison, Wisconsin); BLASTP, BLASTN, BLASTX (Altschul et al., J Mol Biol (1990) 215:403-410); DNASTAR (DNASTAR, Inc., Madison, Wisconsin); and the FASTA program containing the Smith-Waterman algorithm (Pearson, Comput Methods Genome Res [Proc IntSymp] (1994), Meeting Date 1992, 111-120. Eds. Suhai, Sandor. Plenum, New York, NY). In the context of this disclosure, it should be understood that when analysis is performed using sequence analysis software, the results are based on the “default values” of the cited programs. “Default values” will refer to any set of values ​​or parameters initially loaded when the software is first initialized.

[0333] “ improve "", Increase "", inhibition "or" reduce " ： As used herein, the terms “improve,” “increase,” “suppress,” and “reduce,” and their grammatical equivalents, indicate differences in quality or quantity from the reference.

[0334] Methylation status:As used herein, "methylation status at a genomic locus" refers to the frequency of DNA sequences corresponding to a genomic locus identified in assays that detect DNA methylation sequences, and / or the density (e.g., measured density) of DNA methylation at the genomic locus. Methylation status can be determined by a variety of assays known in the art, including but not limited to bisulfite sequencing (BS-Seq), whole-genome bisulfite sequencing (WGBS), methylated DNA immunoprecipitation sequencing (MeDIP-seq), or methyl-CpG-binding domain sequencing (MBD-seq). Differences in methylation status at genomic loci can be detected when two samples are analyzed individually using the same or comparable assays to detect DNA methylation sequences. Methylation status can be compared to a standard or reference. Samples with methylation status different from a standard or reference can be referred to as differentially modified samples.

[0335] "Modification status" or "Histone modification status": As used herein, the “modification status” or “histone modification status” of a genomic locus refers to the frequency of DNA sequences corresponding to a genomic locus identified in assays that detect DNA sequences associated with histones bearing one or more histone modifications (e.g., one or more specific histone modifications), and / or the density (e.g., measured density) of histone modifications (e.g., one or more specific histone modifications) corresponding to a genomic locus. Methylation status can be determined by various assays known in the art, including but not limited to ChIP-seq as an example. Other well-known assays include CUT&RUN (target cleavage and nuclease release) sequencing and CUT&Tag (target cleavage and fragmentation tagging). Differences in the modification status of a genomic locus can be detected if two samples are analyzed separately by the same or comparable assays to detect DNA sequences associated with histones bearing one or more histone modifications (e.g., one or more specific histone modifications). Modification status can be compared to a standard or reference. If the modification status or histone modification status of a sample differs from that of a standard or reference, the sample can be referred to as a differentially modified sample.

[0336] Regulatory sequence: As used herein, in the context of nucleic acid coding sequence expression, a regulatory sequence is a nucleic acid sequence that controls the expression of the coding sequence, such as a promoter sequence or an enhancer sequence. In some implementations, regulatory sequences can control or influence one or more aspects of gene expression (e.g., cell type-specific expression, inducible expression, etc.).

[0337] SubjectsAs used herein, the term "subject" refers to an organism, typically a mammal (e.g., a human). In some embodiments, the subject has a disease, condition, or disorder (e.g., HER2-positive cancer, such as HER2-positive breast cancer, gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc.). In some embodiments, the subject is susceptible to a disease, condition, or disorder. In some embodiments, the subject exhibits one or more symptoms or characteristics of a disease, condition, or disorder. In some embodiments, the subject does not have a disease, condition, or disorder. In some embodiments, the subject does not exhibit any symptoms or characteristics of a disease, condition, or disorder. In some embodiments, the subject has one or more characteristics that define susceptibility to or risk of a disease, condition, or disorder. In some embodiments, the subject is a subject who has been tested for a disease, condition, or disorder and / or has received treatment. In some cases, a human subject may be referred to interchangeably as a "patient" or an "individual."

[0338] Therapeutic agents As used herein, the term "therapeutic agent" means any agent that, when administered to a subject, causes the desired pharmacological effect. In some embodiments, an agent is considered a therapeutic agent if it exhibits a statistically significant effect across an appropriate population. In some embodiments, an appropriate population may be a model organism or a population of individuals. In some embodiments, an appropriate population may be defined by various criteria, such as an age group, sex, genetic background, pre-existing clinical conditions, etc. In some embodiments, a therapeutic agent is a substance that can be used to treat a disease, condition, or disorder (e.g., HER2-positive cancers, such as HER2-positive breast cancer, gastric / gastroesophageal cancer, colorectal cancer, lung cancer, etc.). In some embodiments, a therapeutic...

Claims

1. A method for determining the HER2 status of a subject's cancer, the method comprising: Quantifying one or more epigenetic biomarkers at one or more genomic loci in a liquid biopsy sample obtained from or derived from the subject, wherein the one or more epigenetic biomarkers include: (i) One or more histone modifications, (ii) Chromatin accessibility, (iii) The binding of one or more transcription factors, and / or (iv) DNA methylation, and The HER2 status of the subject's cancer is determined by comparing the levels of one or more epigenetic biomarkers at one or more genomic loci with reference values. The one or more genomic loci mentioned above include (i) one or more genomic loci where the levels of one or more of the aforementioned epigenetic biomarkers are elevated in subjects with HER2-positive cancer compared to subjects with HER2-negative cancer; and / or (ii) one or more epigenetic biomarkers where the levels of one or more of the aforementioned epigenetic biomarkers are elevated in subjects with HER2-negative cancer compared to subjects with HER2-positive cancer; and Optionally, one or more of the quantified genomic sites do not originate from the HER2 amplicon.

2. The method of claim 1, wherein: (a) The liquid biopsy sample is a plasma sample, serum sample, or urine sample; (b) The method includes isolating cfDNA from approximately 1 mL of the liquid biopsy sample (e.g., a plasma sample); and / or (c) The sample contains a detectable amount of ctDNA (e.g., where the estimated tumor fraction of the cfDNA is >3%, e.g., as determined by iChorCNA).

3. The method of claim 1 or 2, wherein the one or more histone modifications are quantified using a histone modification assay, the assay measuring one or more of H3K9ac, H3K14ac, H3K18ac, H3K23ac, H3K27ac, H3K4me1, H3K4me2, H3K4me3 and panacetylation.

4. The method according to any one of claims 1 to 3, wherein: (a) The one or more histone modifications are quantified using assays selected from the following: ChIP-seq (chromatin immunoprecipitation sequencing), CUT&RUN (target cleavage and nuclease release) sequencing and CUT&Tag (target cleavage and fragmentation labeling) sequencing. (b) The chromatin accessibility was quantified using a selection of the following chromatin accessibility assays: ATAC-seq (transposon-accessible chromatin sequencing assay), NOMe-seq (nucleosome occupancy and methylome sequencing), FAIRE-seq (formaldehyde-assisted separation of regulatory elements sequencing), MNase-seq (micrococcal nuclease digestion sequencing), and DNase hypersensitivity assay. (c) The binding of one or more transcription factors is quantified using a transcription factor binding assay that detects the binding of one or more of p300, mediator complex, cohesin complex, RNA pol II, FOXA1, ESR1, PR, MYC, EN1, FOXM1, KLF4, AP-2, RARA, or RUNX1, optionally wherein the transcription factor binding assay is selected from ChIP-seq (chromatin immunoprecipitation sequencing), CUT&RUN (target cleavage and nuclease release) sequencing, and CUT&Tag (target cleavage and fragmentation labeling) sequencing; and / or (d) DNA methylation was quantified using bisulfite sequencing (BS-Seq), whole-genome bisulfite sequencing (WGBS), methylated DNA immunoprecipitation sequencing (MeDIP-seq), or methyl-CpG-binding domain sequencing (MBD-seq).

5. The method of any one of claims 1 to 4, wherein the method comprises: (a) Quantify H3K4me3 modifications at one or more genomic sites using a assay that includes enriching cfDNA containing one or more H3K4me3 modifications (e.g., using a method that includes incubation with an agent that binds to H3K4me3 modifications) and sequencing the cfDNA enriched for H3K4me3 modifications to determine the sequence counts of sequences having one or more H3K4me3 modifications. (b) Quantify H3K27ac modifications at one or more genomic sites using a assay that includes enriching cfDNA containing one or more H3K27ac modifications (e.g., using a method that includes incubation with an agent that binds to H3K27ac modifications) and sequencing the cfDNA enriched for H3K27ac modifications to determine the sequence counts of sequences having one or more H3K27ac modifications. and / or (c) Quantifying methylated DNA at one or more genomic sites using a assay comprising: enriching methylated cfDNA (e.g., using a method comprising incubating with an agent that binds to methylated DNA), and sequencing said enriched cfDNA to determine a sequence count having one or more methylated nucleotides; and Optional location: (d) If the method includes using an agent that binds to H3K4me3 modification, an agent that binds to H3K27ac modification, and / or an agent that binds to methylated DNA, the agent is attached (e.g., via covalent or non-covalent bonds) to a physical support (e.g., beads, magnetic beads, agarose beads, or magnetic epoxy beads) and then incubated with the sample; and / or (e) If the method involves incubation with two or more of the following: an agent that binds to H3K4 modification, an agent that binds to H3K27ac modification, and an agent that binds to methylated DNA, then the sample is incubated with the two or more agents in the following manner: (i) sequentially, or (ii) in parallel (e.g., wherein the sample is divided into several portions and each portion is incubated with a different agent).

6. The method of any one of claims 1 to 5, the method comprising mapping sequence reads to a reference genome, optionally wherein non-unique mappings and redundant sequence reads are discarded and / or peaks in high-noise regions are removed.

7. The method of claim 6, wherein the one or more genomic sites correspond to sequence read peaks, wherein the sequence read peaks correspond to regions in the genome where the number of sequence reads is higher than the local background.

8. The method of any one of claims 5 to 7, wherein quantifying H3K4me3 modification, H3K27ac modification, and / or DNA methylation comprises summing the number of sequence reads that overlap with at least one nucleotide at said one or more genomic sites. Choose one of them: Before summing, the sequence reads are adjusted according to sequencing depth (e.g., normalizing the sequence read quantiles to a common reference distribution) and / or ChIP quality; Sequence counts were normalized to aggregate counts of a set of regions (e.g., 10,000 regions) in a given sample, regions previously identified as having DNAse hypersensitivity in most cell types; and / or Before summing, the estimated value of the local background signal is subtracted from the sequence read at each genomic locus.

9. The method of any one of claims 1 to 8, wherein the reference value is a predetermined threshold, a measurement of a liquid biopsy sample, a measurement from a liquid biopsy sample obtained from a group of subjects, and / or a normalized value, wherein: It has been previously demonstrated that the predetermined threshold and the normalized value distinguish between HER2-positive and HER2-negative subjects (e.g., by using AUROC greater than 0.5); The reference values ​​are measurements obtained from liquid biopsy samples from a group of subjects who had previously been diagnosed with HER2-positive or HER2-negative cancer.

10. The method of any one of claims 5 to 9, wherein the method comprises calculating the sequence read density at the one or more genomic loci, optionally wherein the sequence read density is calculated in the following manner: (a) Summing the background-adjusted sequence counts at each of the one or more genomic loci and dividing by the sum of the kilobases at the one or more genomic loci; or (b) For each genomic locus, divide the background adjustment fragment count by the number of kilobases of the genomic locus, and then sum over each locus.

11. The method of claim 10, wherein the method comprises calculating a HER2 positive / HER2 negative ratio score by means of a method comprising: (a) The HER2 positive sequence read density is calculated by summing the background-adjusted sequence counts at each of the one or more genomic loci where the level of the one or more epigenetic biomarkers is increased in a sample obtained from a subject with HER2-negative cancer compared to a sample obtained from a subject with HER2-positive cancer. (b) Calculating the HER2-negative sequence read density by summing background-adjusted sequence counts at each of the one or more genomic loci where the levels of the one or more epigenetic biomarkers are increased, compared to samples obtained from subjects with HER2-positive cancer; and (c) Divide the HER2 positive sequence read density by the HER2 negative sequence read density.

12. The method of claim 11, wherein the method comprises: (a) Determine the HER2 positive / HER2 negative ratio score for H3K4me3 modification; (b) Determine the H3K27ac modified HER2 positive / HER2 negative ratio score; and / or (c) Determine the HER2-positive / HER2-negative ratio score of methylated DNA; and If each of (a) through (c) is performed, then each of the said ratio scores may optionally be combined (e.g., with a fitted value determined using logistic regression).

13. The method of any one of claims 1 to 12, wherein the method comprises quantifying the one or more epigenetic biomarkers at one or more genomic loci in Tables 1 to 3 or Tables 5 to 7; Optionally, the method described herein includes: (a) Quantify H3K4me3 modifications at at least 5, 10, 20, 30, 40 or 50 genomic loci in Table 1, optionally one or more of the genomic loci are not derived from HER2 amplicon; (b) Quantify H3K27ac modifications at at least 5, 10, 20, 30, 40 or 50 genomic loci in Table 2, optionally one or more of the genomic loci are not derived from HER2 amplicon; (c) Quantify DNA methylation at at least 5, 10, 20, 30, 40 or 50 genomic sites in Table 3, optionally one or more of the genomic sites not originating from the HER2 amplicon; (d) Quantify the H3K4me3 modification of one or more k4 analyte sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (e) Quantify the H3K27ac modification of one or more k27 analyte sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (f) Quantify DNA methylation at one or more mbd sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (g) Quantify the H3K4me3 modification of one or more k4 analyte sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (h) Quantify the H3K27ac modification of one or more k27 analyte sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (i) Quantify DNA methylation at one or more mbd sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (j) Quantify the H3K4me3 modification of one or more k4 analyte sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (k) Quantify the H3K27ac modification of one or more k27 analyte sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (l) Quantify DNA methylation at one or more mbd sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (m) or any combination of (a) to (l).

14. The method of any one of claims 1 to 13, wherein the area under the recipient operating characteristic curve (AUROC) provided by the method for determining whether a subject has HER2-positive or HER2-negative cancer is greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95).

15. The method of any one of claims 1 to 14, wherein the subject has been previously diagnosed with cancer, the subject has increased susceptibility to cancer, and / or wherein the method further comprises determining whether the subject has cancer.

16. The method of any one of claims 1 to 15, wherein the HER2-positive cancer is HER2-3+ cancer based on IHC testing, and the HER2-negative cancer is HER2-0 cancer based on IHC testing.

17. The method of any one of claims 1 to 16, wherein the cancer is breast cancer, gastric / gastric esophageal cancer, colorectal cancer, or lung cancer.

18. A method for treating a subject suffering from cancer, the method comprising: The subject is given cancer therapy based on the HER2 status of the cancer, wherein the HER2 status of the cancer has been determined using the method of any one of claims 1 to 17, and wherein: (a) If the cancer has been identified as HER2-positive, the cancer treatment does not include the administration of a HER2-targeting agent; and (b) If the cancer has been determined to be HER2 negative, the cancer treatment does not include the administration of a HER2-targeting agent.

19. A method for monitoring the HER2 status of a subject's cancer, and optionally treating said cancer, the method comprising determining the HER2 status of said cancer at a first time point and a second time point using the method of any one of claims 1 to 18. Optionally, the HER2-targeting agent is administered to the subject close to the first time point (e.g., on the same day as or a few days before the first time point) or after the first time point and before the second time point.

20. The method of claim 18 or 19, the method further comprising administering a cancer therapy to the subject based on the HER2 status of the cancer at the second time point, optionally a HER2-targeting agent, optionally wherein the type, dose, and / or frequency of the cancer therapy is adjusted based on the HER2 status of the cancer at the second time point.

21. A method for treating a subject suffering from cancer, the method comprising: The subject was administered a HER2-targeting agent, wherein analysis of cell-free DNA (cfDNA) from a biological sample obtained from or derived from the subject, optionally from a liquid biopsy sample, determined that the subject possessed validated epigenetic features indicative of HER2-positive cancer. The presence of the verified epigenetic features was determined using a validated classifier. The classifier used for verification is obtained in the following way: (a) Identify genomic features of (i) one or more HER2-positive cell lines or (ii) biological samples obtained from a first group of subjects who have been previously identified as having HER2-positive cancer, optionally HER2-3+ cancer, HER2-2+ cancer or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing, including histone modifications, chromatin accessibility, binding of one or more transcription factors and / or DNA methylation. (b) Identify genomic features of (i) one or more HER2-negative cell lines or (ii) biological samples obtained from a second group of healthy subjects or subjects previously identified as having HER2-negative cancer, optionally based on IHC testing, including histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation. (c) Compare the genomic features identified in step (a) with those identified in step (b) to identify genomic sites ("difference sites") that show statistical differences in histone modification, chromatin accessibility, transcription factor binding and / or DNA methylation levels. (d) Using histone modification, chromatin accessibility, transcription factor binding, and / or DNA methylation level training at the differential sites, a classifier is used to distinguish (i) samples from one or more HER2-positive cell lines or biological samples obtained from the first cohort, and (ii) samples from one or more HER2-negative cell lines or biological samples obtained from the second cohort, to identify samples with histone modification, chromatin accessibility, transcription factor binding, and / or DNA methylation level characteristics ("epigenetic features") indicating that the sample may have been obtained from a HER2-positive cell line or from the first cohort; and (e) The validated classifier is obtained by validating the classifier in step (d) on a third cohort comprising independent, blinded subjects with HER2-positive and HER2-negative cancers, and a threshold is selected such that the validated classifier predicts HER2-positive cancers, optionally based on HER2-3+, HER2-2+, or HER2-1+ cancers based on IHC testing, or HER2-low cancers based on IHC / ISH testing, with an area under the recipient operating characteristic curve (AUROC) greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95), wherein subjects falling into the predicted HER2-positive cancer group exhibit the validated epigenetic characteristic, and subjects not falling into the HER2-positive cancer group lack the validated epigenetic characteristic.

22. The method of claim 21, wherein: (a) The differential sites in step (c) are determined by comparing the genomic features of one or more histone modifications and / or DNA methylation in (i) one or more HER2-positive cell lines and (ii) one or more HER2-negative cell lines; (b) The classifier in step (d) is for use with Computer simulation Training was performed using sequence data from one or more HER2-positive cell lines and sequence data from liquid biopsy samples from healthy subjects, focusing on histone modifications, chromatin accessibility, transcription factor binding, and / or DNA methylation levels at the differentially expressed sites. (c) Validate the classifier validated in step (e) using liquid biopsy samples from the third queue; (d) The classifier in step (d) is trained against one or more (e.g., two or more) histone modification levels and / or DNA methylation levels at the differential sites, optionally wherein the one or more histone modification levels include H3K4me3 and H3K27ac modification levels; (e) The classifier in step (d) is trained using ridge regression, elastic network regression, or lasso regression; and / or (f) wherein the genomic features of one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation have been identified in one or more HER2-positive cell lines; and wherein, in step (d), sequencing fragments from healthy donor plasma samples and cell lines are mixed in different proportions. Computer simulation The sample was diluted to achieve a simulated ctDNA percentage range of 0.5% to 50%.

23. The method as described in claim 22, The genomic features described herein have been identified in one or more HER2-positive cell lines, including one or more histone modifications, chromatin accessibility, binding to one or more transcription factors, and / or DNA methylation; and The method further includes, optionally, using plasma data, adjusting the classifier using a transfer learning process, the transfer learning process including: (i) Calculate the predicted value of the plasma sample in the form of a probability value using the classifier (e.g., using the formula log2(HER2+ probability / 1 – HER2+ probability)); (ii) The odds value is used as an offset in a new model (e.g., a lasso logistic regression model), using all the same features but trained with plasma data (e.g., using leave-one-out method) to determine new weights and coefficients, which are then added to the coefficients determined during cell line training to obtain the adjusted model.

24. The method of claim 23, wherein the novel model is trained for cancer-specific genomic loci in the plasma data, wherein the cancer-specific genomic loci are regions in HER2-positive subjects that are enriched with H3K4me3 and H3K27ac modifications compared to HER2-negative subjects and are associated with ctDNA at the HER2 loci.

25. A kit comprising reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic loci, wherein the one or more genomic loci are selected from Tables 1 to 3 and Tables 4 to 7, optionally wherein one or more of the genomic loci are not derived from HER2 amplicon, and optionally wherein the kit comprises reagents for quantifying: (a) H3K4me3 modification at least 5, 10, 20, 30, 40 or 50 genomic loci in Table 1, optionally one or more of the genomic loci are not derived from the HER2 amplicon; (b) H3K27ac modification of at least 5, 10, 20, 30, 40 or 50 genomic loci in Table 2, optionally one or more of the genomic loci are not derived from HER2 amplicon; (c) DNA methylation at at least 5, 10, 20, 30, 40 or 50 genomic sites in Table 3, optionally one or more of the genomic sites are not derived from the HER2 amplicon; (d) H3K4me3 modification of one or more k4 analyte sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (e) H3K27ac modification of one or more k27 analyte sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (f) DNA methylation of one or more mbd sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (g) H3K4me3 modification of one or more k4 analyte sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (h) H3K27ac modification of one or more k27 analyte sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (i) DNA methylation of one or more mbd sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (j) H3K4me3 modification of one or more k4 analyte sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (k) H3K27ac modification of one or more k27 analyte sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (l) DNA methylation of one or more mbd sites listed in Table 7, optionally one or more of these genomic sites not originating from a HER2 amplicon; or Any combination of (m) (a) to (l).

26. The kit of claim 25, wherein the kit comprises: (a) One or more antibodies for ChIP-seq, wherein said one or more antibodies specifically bind to histones modified with H3K4me3 or H3K27ac; (b) One or more methyl-binding domains for MBD-seq, or one or more antibodies for binding methylated DNA for MeDIP; (c) Reagents for isolating cell-free DNA (cfDNA) from liquid biopsy samples; (d) Reagents for library preparation used in sequencing; (e) Reagents used for sequencing; (f) Instructions used to determine whether a subject has HER2-positive cancer; or (g) Any combination of (a) to (f).

27. A non-transient computer-readable storage medium encoded with a computer program, wherein the program includes instructions that, when executed by one or more processors, cause the one or more processors to perform operations to perform the method of any one of claims 1 to 26.

28. A computer system comprising a memory and one or more processors coupled to the memory, wherein the one or more processors are configured to perform operations to perform the method of any one of claims 1 to 27.

29. A system for determining the HER2 status of a subject's cancer, the system comprising a sequencer configured to generate a sequencing dataset from a sample; and the non-transient computer-readable storage medium of claim 27 and / or the computer system of claim 28; Optionally, the sequencer is configured to generate a whole-genome sequencing (WGS) dataset from the sample.

30. The system of claim 29, further comprising a sample preparation device configured to prepare the sample for sequencing from a biological sample (e.g., a liquid biopsy sample); Optionally, the sample preparation apparatus includes reagents for quantifying one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation at one or more genomic sites in the cell-free DNA (cfDNA) of the biological sample, optionally the liquid biopsy sample.

31. The system of claim 30, wherein the one or more genomic loci are selected from Tables 1 to 3 or Tables 5 to 7, and optionally wherein the device comprises a reagent for: (a) Quantify H3K4me3 at at least 5, 10, 20, 30, 40 or 50 genomic loci, for example, in Table 1, optionally one or more of the genomic loci not derived from HER2 amplicon; (b) Quantify H3K27ac at at least 5, 10, 20, 30, 40 or 50 genomic loci, for example, in Table 2, optionally one or more of the genomic loci not derived from HER2 amplicon; (c) Quantify DNA methylation at at least 5, 10, 20, 30, 40 or 50 genomic sites, for example, as shown in Table 3, optionally one or more of the genomic sites not derived from the HER2 amplicon; (d) Quantify the H3K4me3 modification of one or more k4 analyte sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (e) Quantify the H3K27ac modification of one or more k27 analyte sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (f) Quantify DNA methylation at one or more mbd sites listed in Table 5, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (g) Quantify the H3K4me3 modification of one or more k4 analyte sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (h) Quantify the H3K27ac modification of one or more k27 analyte sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (i) Quantify DNA methylation at one or more mbd sites listed in Table 6, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (j) Quantify the H3K4me3 modification of one or more k4 analyte sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (k) Quantify the H3K27ac modification of one or more k27 analyte sites listed in Table 7, wherein one or more of the genomic sites are not derived from the HER2 amplicon; (l) Quantify DNA methylation at one or more mbd sites listed in Table 7, optionally one or more of these genomic sites not originating from a HER2 amplicon; or Any combination of (m) (a) to (l).

32. The system as claimed in any one of claims 29 to 31, wherein: (a) The reagent contains one or more antibodies for ChIP-seq, wherein the one or more antibodies specifically bind to histones modified by H3K4me3 or H3K27ac; (b) The reagent contains one or more methyl-binding domains for MBD-seq; (c) The apparatus includes reagents for isolating cell-free DNA (cfDNA) from the biological sample, optionally the liquid biopsy sample; (d) The apparatus includes library preparation reagents for sequencing; and / or (e) The sequencer includes reagents for sequencing.

33. A method for determining the HER2 status of cancer in a subject (e.g., a patient), the method comprising: Receive (e.g., via a processor of a computing device) one or more genomic features of the subject, including one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation; as well as Whether the subject has epigenetic features indicating HER2-positive cancer is determined by classifying the genomic features using a HER2 classifier (e.g., by the processor).

34. The method of claim 33, wherein the HER2 classifier has been trained using one or more genomic features including one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation: (i) one or more HER2-positive cell lines and one or more HER2-negative cell lines and / or (ii) one or more biological samples obtained from one or more subject cohorts previously identified as having HER2-positive cancer, optionally HER2-3+, HER2-2+, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing, and one or more biological samples obtained from one or more subject cohorts previously identified as having HER2-negative cancer, optionally HER2-0 cancer based on IHC testing. Optionally, the one or more genomic features used to train the HER2 classifier include one or more genomic features generated by computer simulation of diluting sequence data from HER2-positive or HER2-negative cell lines with sequence data obtained from healthy donor plasma samples to simulate a ctDNA percentage in the range of 0.5% to 50%.

35. The method of claim 34, wherein the genomic features used to train the HER2 classifier are targeted at differentially expressed genomic sites found to be statistically significantly different in terms of one or more histone modifications, chromatin accessibility, binding to one or more transcription factors, and / or DNA methylation levels: (a) compared to one or more HER2-negative cell lines, one or more HER2-positive cell lines, and / or (b) compared to one or more biological samples obtained from one or more subject cohorts previously identified as having HER2-negative cancer, optionally HER2-0 cancer based on IHC testing, or HER2-positive cancer, optionally HER2-3+ cancer, HER2-2+ cancer, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing. Optionally, the differential sites are determined by comparing genomic features of one or more histone modifications and / or DNA methylation in (i) one or more HER2-positive cell lines and (ii) one or more HER2-negative cell lines.

36. The method of any one of claims 33 to 35, wherein the HER2 classifier has been trained using the following: (a) Genomic features of two or more histone modification levels at the differentially expressed sites; or (b) Genomic characteristics of the levels of one or more histone modifications and DNA methylation at the differential sites.

37. The method of any one of claims 33 to 36, wherein the method comprises receiving: (a) One or more genomic features of two or more histone modifications, wherein the two or more histone modifications optionally include H3K4me3 and H3K27ac modifications; (b) One or more genomic features of histone modifications and DNA methylation, wherein, optionally, the histone modifications include H3K4me3 and / or H3K27ac modifications.

38. The method of any one of claims 33 to 37, wherein the HER2 classifier has been adjusted using sequencing data obtained from plasma samples using a transfer learning process, the transfer learning process comprising: (i) Using a HER2 classifier trained with genomic features such as one or more histone modifications, chromatin accessibility, binding of one or more transcription factors, and / or DNA methylation in one or more HER2-positive and one or more HER2-negative cell lines, a predicted value for a plasma sample is calculated in the form of a probability value (e.g., using the formula log2(HER2+ probability / 1 – HER2+ probability)); and (ii) In a new model (e.g., a lasso logistic regression model), the odds value is used as an offset term, using all the same features as the HER2 classifier trained with cell line data, but trained with plasma data (e.g., using leave-one-out method) to determine new weights and coefficients, and these weights and coefficients are then added to the coefficients determined during cell line training to obtain the adjusted model. Optionally, the novel model has been adjusted using cancer-specific genomic loci from the plasma data, wherein the cancer-specific genomic loci are regions enriched with H3K4me3 and / or H3K27ac modifications in HER2-positive subjects compared to HER2-negative subjects, and associated with ctDNA at the HER2 locus.

39. The method of any one of claims 33 to 38, wherein the HER2 classifier is a validated classifier, wherein the HER2 classifier has been validated by selecting a threshold such that the validated classifier predicts HER2-positive cancer, optionally based on HER2-3+, HER2-2+, or HER2-1+ cancer based on IHC testing, or HER2-low cancer based on IHC / ISH testing, with an area under the recipient operating characteristic curve (AUROC) greater than 0.5 (e.g., greater than 0.55, greater than 0.6, greater than 0.65, greater than 0.7, greater than 0.75, greater than 0.8, greater than 0.85, greater than 0.9, or greater than 0.95), and Choose one of them: (a) The HER2 classifier has been validated against a cohort of subjects with HER2-positive and HER2-negative cancers, wherein subjects falling into the predicted HER2-positive cancer group exhibit the validated epigenetic characteristics, and subjects not falling into the HER2-positive cancer group lack the validated epigenetic characteristics; and / or (b) The HER2 classifier has been validated using liquid biopsy sample data.

40. A non-transient computer-readable storage medium encoded with a computer program, wherein the program includes instructions that, when executed by one or more processors, cause the one or more processors to perform operations to perform the method of any one of claims 33 to 39.

41. A computer system comprising a memory and one or more processors coupled to the memory, wherein the one or more processors are configured to perform operations to perform the method of any one of claims 33 to 39.

42. A method for treating a subject suffering from cancer, the method comprising: The subject was administered a HER2-targeting agent, wherein analysis of cell-free DNA (cfDNA) from a biological sample obtained from or derived from the subject, optionally from a liquid biopsy sample, determined that the subject possessed validated epigenetic features indicative of HER2-positive cancer. The presence of the verified epigenetic feature has been determined using a classifier (e.g., a verified classifier) ​​according to any one of claims 33 to 39.

Images