Systems and methods for predicting biomarker expression

Abstract

Described herein are methods for predicting presence of a biomarker, comprising obtaining a cell-free sample from a patient, detecting levels of small non-coding ribonucleic acids (smRNAs) in the cell-free sample of the patient, predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs.

Description

Attomey Docket No. 61740-709601SYSTEMS AND METHODS FOR PREDICTING BIOMARKER EXPRESSIONCROSS-REFERENCE

[0001] This application claims the benefit of priority to U.S. Provisional App. No. 63 / 692,006 filed September 6, 2024, which is incorporated by reference herein in its entirety.BACKGROUND

[0002] Biomarkers are commonly used for patient therapy selection and for assessment of pharmacodynamics of target during cancer therapy treatment. Biomarkers can be detected from biopsy specimens, cell-free samples, or other biological samples. Biomarker expression can be correlated with disease or response to a treatment for a disease.SUMMARY

[0003] Applicant has recognized various non-limiting issues in terms of tissue-based biomarker analysis. For example, Applicant has recognized that a possible major challenge for tissue-based biomarker analysis such as H4C can be when patient tissue biopsies are unavailable, particularly in challenging biopsy specimens or when the cancer has metastasized, and there is an urgent need to assess the biomarker status in the tumor.

[0004] Applicant has recognized that, currently, major commercially available liquid biopsy diagnostics generally do not have the capabilities to robustly assess gene expression profiles or pathway activation. Applicant has recognized, for example, non-limiting challenges in using liquid biopsy-derived ctDNA-based approaches. For example, applicant has recognized an issue in that reliance on tumor shedding in the plasma by ctDNA-based approaches can limit the sensitivity of detection and accurate subtyping. Applicant has recognized that therefore, reach of patients who can benefit from the diagnostic test may be limited. In addition, Applicant has recognized that cell-free DNA assays generally require a targeted panel enriched for specific biomarker loci, which Applicant has recognized generally constrains the diagnostic capacity to assess diverse biomarkers in the absence of prior panel coverage. Applicant has recognized that a single, universal assay capturing the entire transcriptome from cell-free small RNAs has the potential to facilitate adaptable, comprehensive exploration of biomarker profiles.

[0005] Provided herein is a method for predicting presence of a biomarker, comprising: (a) obtaining a cell-free sample from a patient, (b) detecting levels of small non-coding ribonucleic acids (smRNAs) in the cell-free sample of the patient, and (c) predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs.

[0006] In some embodiments, (b) comprises subjecting the smRNAs to reverse transcription to generate complementary deoxyribonucleic acid (cDNA) molecules and sequencing the one orAttorney Docket No. 61740-709601 more cDNA molecules or derivatives thereof. In some embodiments, (c) comprises, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with expression of the biomarker.

[0007] In some embodiments, the panel comprises at least 50 smRNAs. In some embodiments, the panel comprises at least 100 smRNAs. In some embodiments, the panel comprises at least 250 smRNAs. In some embodiments, the panel comprises at least 500 smRNAs. In some embodiments, the panel comprises at least 1000 smRNAs.

[0008] In some embodiments, the panel of smRNAs is identified using expression quantitative trait loci (QTL) analysis. In some embodiments, the panel of smRNAs is identified using association of smRNAs with one or more known gene sets, one or more RNA isoform specific start sites, or other fragmentomic features, thereby obtaining the panels of smRNAs. In some embodiments, the smRNAs in the panel directly overlap with the biomarker and a specific locus of expression. In some embodiments, a first portion of the panel of smRNAs are positively correlated with expression of the biomarker, and wherein, a second portion of the panel of smRNAs are negatively correlated with expression of the biomarker. In some embodiments, (c) comprises calculating a gene expression score, wherein, the gene expression score is calculated based on a ratio of: (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell-free sample. In some embodiments, the gene expression score is a log2 of the ratio. In some embodiments, (c) comprises: (i) predicting the patient is positive for the biomarker if the gene expression score is positive and (ii) predicting the patient is negative for the biomarker if the gene expression score is negative. In some embodiments, the panel of smRNAs are positively correlated with the expression of the biomarker. In some embodiments, (c) comprises calculating a gene expression score, wherein the gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample. In some embodiments, (c) comprises: (i) predicting the patient is positive for the biomarker if the gene expression score is above a threshold value and (ii) predicting the patient is negative for the biomarker if the gene expression score is below the threshold value. In some embodiments, the sequencing comprises sequencing by synthesis.

[0009] In some embodiments, the biomarker comprises estrogen receptors (ERs). In some embodiments, (c) comprises, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with ESRI gene expression. In some embodiments, the panel of smRNAs comprises any smRNAs complementary to the cDNA sequences listed in Table 1. In some embodiments, theAttorney Docket No. 61740-709601 panel of smRNAs comprises any smRNA sequence for which at least 50% of the smRNA sequence is complementary to a sequence of any of the chromosomal regions listed in Table 2. In some embodiments, a first portion of the panel of smRNAs are positively correlated with ESRI gene expression, and wherein a second portion of the panel of smRNAs are negatively correlated with ESRI gene expression. In some embodiments, (c) comprises calculating an ESRI gene expression score, wherein the ESRI gene expression score is calculated based on a ratio of: (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell-free sample. In some embodiments, the ESRI gene expression score is a log2 of the ratio. In some embodiments, a positive ESRI gene expression score indicates the patient is positive for ERs. In some embodiments, a negative ESRI gene expression score indicates the patient is negative for ERs. In some embodiments, the panel of smRNAs are positively correlated with ESRI gene expression. In some embodiments, (c) comprises calculating an ESRI gene expression score, wherein the ESRI gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample. In some embodiments, (c) comprises: (i) predicting the patient is positive for ERs if the ESRI gene expression score is above a threshold value and (ii) predicting the patient is negative for ERs if the ESRI gene expression score is below the threshold value.

[0010] In some embodiments, the biomarker comprises human epidermal growth factor receptor- 2 (HER2). In some embodiments, (c) comprises, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with ERBB2 gene expression. In some embodiments, the panel of smRNAs comprises any smRNAs complementary to the cDNA sequences listed in Table 3. In some embodiments, the panel of smRNAs comprises any smRNA sequence for which at least 50% of the smRNA sequence is complementary to a sequence of any of the chromosomal regions listed in Table 4. In some embodiments, the panel of smRNAs are positively correlated with ERBB2 gene expression. In some embodiments, (c) comprises calculating an ERBB2 gene expression score, wherein the ERBB2 gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample. In some embodiments, (c) comprises: (i) predicting the patient is HER2 positive if the ERBB2 gene expression score is above a threshold value and (ii) predicting the patient is HER2 negative if the ERBB2 gene expression score is below the threshold value.

[0011] In some embodiments, the biomarker comprises messenger RNA (mRNA) gene expression. In some embodiments, the biomarker comprises pathway activation. In someAttorney Docket No. 61740-709601 embodiments, the biomarker comprises trophoblast cell-surface antigen 2 (TR0P2). In some embodiments, the biomarker comprises human epidermal growth factor receptor-3 (HER3). In some embodiments, the biomarker comprises carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5).

[0012] In some embodiments, (a) comprises obtaining plasma from the patient, wherein the plasma comprises the cell-free sample. In some embodiments, (a) comprises obtaining serum from the patient, wherein the serum comprises the cell-free sample. In some embodiments, in (a), a tissue sample from the patient is not obtained. In some embodiments, (a) comprises obtaining cerebrospinal fluid (CSF) from the patient, wherein the CSF comprises the cell-free sample. In some embodiments, (a) comprises obtaining urine from the patient, wherein the urine comprises the cell-free sample. In some embodiments, (a) comprises obtaining lymph from the patient, wherein the lymph comprises the cell-free sample. In some embodiments, (a) comprises obtaining saliva from the patient, wherein the saliva comprises the cell-free sample. In some embodiments, a total volume of the cell-free sample is between about 20 microliters and about 2 milliliters. In some embodiments, a total volume of the cell-free sample is between about 100 microliters to about 500 microliters.

[0013] In some embodiments, the method further comprises (d) predicting a response of the patient to a treatment based at least in part on the predicted presence of the biomarker obtained in (c). In some embodiments, the biomarker comprises estrogen receptors (ERs). In some embodiments, the treatment comprises administration of a selective estrogen receptor modulator. In some embodiments, the treatment comprises administration of an aromatase inhibitor. In some embodiments, the treatment comprises administration of a selective estrogen receptor degrader. In some embodiments, the treatment comprises administration of a cyclic-dependent kinase (CDK) inhibitor. In some embodiments, the CDK inhibitor comprises a CDK4 inhibitor. In some embodiments, the CDK inhibitor comprises a CDK6 inhibitor. In some embodiments, the CDK inhibitor comprises palbociclib. In some embodiments, the CDK inhibitor comprises ribociclib. In some embodiments, the CDK inhibitor comprises abemaciclib. In some embodiments, the treatment comprises administration of a mTOR inhibitor. In some embodiments, the mTOR inhibitor comprises everolimus. In some embodiments, the treatment comprises administration of a PI3K inhibitor. In some embodiments, the PI3K inhibitor comprises alpelisib. In some embodiments, the PI3K inhibitor comprises inavolisib. In some embodiments, the treatment comprises administration of an AKT inhibitor. In some embodiments, the AKT inhibitor comprises capivasertib. In some embodiments, the treatment comprises administration of sacituzumab govitecan. In some embodiments, the treatment comprises administration ofAttorney Docket No. 61740-709601 datopotamab deruxtecan. In some embodiments, the treatment comprises administration of a PARP inhibitor.

[0014] In some embodiments, the biomarker comprises human epidermal growth factor receptor- 2 (HER2). In some embodiments, the treatment comprises administration of trastuzumab. In some embodiments, the treatment comprises administration of pertuzumab. In some embodiments, the treatment comprises administration of margetuximab. In some embodiments, the treatment comprises administration of a kinase inhibitor. In some embodiments, the kinase inhibitor comprises neratinib. In some embodiments, the kinase inhibitor comprises tucatinib. In some embodiments, the kinase inhibitor comprises lapatinib. In some embodiments, the treatment comprises administration of ado-trastuzumab emtansine. In some embodiments, the treatment comprises administration of fam-trastuzumab deruxtecan.

[0015] In some embodiments, the smRNAs have a length of less than 200 nucleotides. In some embodiments, the smRNAs have a length from 50 to 100 nucleotides.INCORPORATION BY REFERENCE

[0016] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The novel features of the disclosure are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present disclosure will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the disclosure are utilized, and the accompanying drawings of which:

[0018] FIG. 1A shows a Pearson correlation between the predicted ESRI gene expression (using smRNA-QTLs) against known ESRI gene expression, as determined by TCGA.

[0019] FIG. IB shows Mann-Whitney-Wilcoxon (MWW) analysis (p = 1.341e-90) of the predicted ESRI gene expression (using smRNA-QTLs) compared to the known ESRI gene expression, as determined by TCGA.

[0020] FIG. 2A shows a Pearson correlation (P = 0.63, p-value = 3.39e-120) between the predicted ERBB2 gene expression (using smRNA-QTLs) against known ERBB2 gene expression, as determined by TCGA.Attorney Docket No. 61740-709601

[0021] FIG. 2B shows Mann-Whitney-Wilcoxon (MWW) analysis of the predicted ERBB2 gene expression (using smRNA-QTLs) compared to the known ERBB2 gene expression, as determined by TCGA.

[0022] FIG. 3 shows intracellular and extracellular ERBB2 (HER2) status correlated to ERBB2 smRNA CPM, for the cell lines described herein.

[0023] FIG. 4 shows ERBB2 (HER2) status from the extracellular fraction, correlated to ERBB2 smRNA CPM, for the cell lines presented in FIG. 3, as determined by the Cancer Dependency Map Project (DepMap).

[0024] FIG. 5 shows intracellular and extracellular ERBB2 (HER2) status from the cell lines presented in FIG. 3, as determined by DepMap, stratified by negative status positive status from 0-85thpercentile expression, and positive status from 85- 100thpercentile expression, against ERBB2 smRNA CPM, generated as described herein.

[0025] FIG. 6 shows intracellular and extracellular ER status correlated to ER smRNA CPM, generated as described herein (using ER exons), for the cell lines described herein.

[0026] FIG. 7A shows ER expression scores in HC1143 cells across titration levels of smRNAs from MCF7 and ZR7530 conditioned media.

[0027] FIG. 7B shows HER2 expression scores in HC1143 cells across titration levels of smRNAs from HCC1954 and SKBR3 conditioned media.

[0028] FIGs. 8A-8B show ESRI smRNA CPM scores from TCGA plasma samples and stratified by known ER status from H4C on these patients’ tumors, patients across Stage I / I I tumor status (FIG. 8A) and Stage III / IV tumor status (FIG. 8B).

[0029] FIGs. 9A-9B show ERBB2 smRNA CPM scores from TCGA plasma samples and stratified by known HER2 status from H4C on these patients’ tumors, across Stage I / I I tumor status (FIG. 9 A) and Stage III / IV tumor status (FIG. 9B).

[0030] FIGs. 10A-10B shows a correlation of ESRI smRNA CPM from two replicate plasma samples from each patient, as provided by TCGA, across tumor Stage VII (FIG. 10A) and tumor Stage III / IV (FIG. 10B).

[0031] FIGs. 11A-11B shows a correlation of ERBB2 smRNA CPM from two replicate plasma samples from each patient, as provided by TCGA, across tumor Stage VII (FIG. 11 A) and tumor Stage III / IV (FIG. 11B).

[0032] FIG. 12 illustrates an example workflow schematic for sample collection and smRNA library generation.Attorney Docket No. 61740-709601

[0033] FIG. 13 shows EBBR2 smRNA CPM scores from subject-derived plasma stratified by HER2 expression status, across tumor stages.

[0034] FIG. 14 shows, for each subject, discrimination of ESRI (ER) status by ESRI VV gene score in HER2- tumor tissue at tumor stage 1 (n=<100), stage II (n=317), stage III (n=123), and stage IV (n=6).

[0035] FIG. 15 shows Pearson’s correlation (R = 0.689) of ESRI expression z-scores to ESRI VV gene scores.

[0036] FIG. 16 shows intracellular and extracellular ESRI (ER) status correlated to ESRI VV gene score, for the cell lines described herein.

[0037] FIG. 17 shows intracellular and extracellular ESRI expression from the cell lines presented in FIG. 16, as determined by the Cancer Dependency Map Project (DepMap), stratified by negative status, positive status from 0-85thpercentile expression, and positive status from 85-100thpercentile expression, against ESRI VV gene scores.

[0038] FIG. 18 shows ESRI VV gene scores stratified by negative ER status and positive ER status across tumor stages.DETAILED DESCRIPTION

[0039] Disclosed herein are methods for predicting a presence of a biomarker. The method can comprise obtaining a cell-free sample from a patient. The method can comprise detecting levels of small non-coding ribonucleic acids (smRNAs) in the cell-free sample of the patient. The method can comprise predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs.

[0040] The detecting levels of smRNAs in the cell-free sample of the patient can comprise subjecting the smRNAs to reverse transcription to generate complementary deoxyribonucleic acid (cDNA) molecules and sequencing the one or more cDNA molecules or derivatives thereof. The predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs can comprise, based on the sequencing, determining whether the panel of smRNAs are associated with expression of the biomarker. The panel of smRNAs can comprise from about 20 smRNAs to about 80 smRNAs, from about 40 smRNAs to about 100 smRNAs, from about 60 smRNAs to about 120 smRNAs, about 80 smRNAs to about 140 smRNAs, or from about 100 smRNAs to about 160 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 50 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 100 smRNAs. The panel of smRNAs can comprise from about 200 smRNAs to about 300 smRNAs, from about 350 smRNAs to about 450 smRNAs, from about 500 smRNAs to about 650Attorney Docket No. 61740-709601 smRNAs, about 700 smRNAs to about 850 smRNAs, or from about 900 smRNAs to about 1000 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 250 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 500 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 1000 smRNAs. The panel of smRNAs can comprise from about 500 to about 1,500 smRNAs, from about 2,000 to about 3,000 smRNAs, from about 3,500 to about 4,500 smRNAs, from about 5,000 to about 6,000 smRNAs, from about 6,500 to about 7,000 smRNAs, from about 7,500 to about 8,500 smRNAs, or from about 9,000 to about 10,000 smRNAs.

[0041] The panel of smRNAs can be identified using expression quantitative trait loci (QTL) analysis. The panel of smRNAs can be identified using an association of a plurality of smRNAs detected in a cell-free sample, as described herein, with one or more known gene sets, one or more RNA isoform specific start sites, or one or more other fragmentomic features.

[0042] In certain instances, one or more of the smRNAs in the panel directly overlap with a portion of a gene encoding for a biomarker. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, each of the smRNAs in the panel directly overlap with a portion of a gene encoding for a biomarker. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, one or more of the smRNAs in the panel directly overlap with a locus of expression of a biomarker. In certain instances, each of the smRNAs in the panel directly overlap with one or more loci of expression of a biomarker. In certain instances, one or more of the smRNAs in the panel partially overlap with a portion of a gene encoding for a biomarker. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, each of the smRNAs in the panel partially overlap with a portion of a gene encoding for a biomarker. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, one or more of the smRNAs in the panel partially overlap with a locus of expression of a biomarker. In certain instances, each of the smRNAs in the panel partially overlap with one or more loci of expression of a biomarker. Partially overlapping can comprise at least or about 50% alignment, at least or about 55% alignment, at least or about 60% alignment, at least or about 70% alignment, at least or about 75% alignment, at least or about 80% alignment, at least or about 85% alignment, at least or about 90% alignment, at least or about 95% alignment, or at least or about 98% alignment of an smRNA of the panel of smRNAs with a portion of a gene encoding a biomarker or a locus of expression of a biomarker. Partially overlapping can comprise at most or about 50% alignment, at most or about 55% alignment, at most or about 60% alignment, at most or about 70% alignment, at most or about 75% alignment, at most or about 80% alignment, at most or about 85% alignment, at mostAttorney Docket No. 61740-709601 or about 90% alignment, at most or about 95% alignment, or at most or about 98% alignment of an smRNA of the panel of smRNAs with a portion of a gene encoding a biomarker or a locus of expression of a biomarker. In certain instances, each of a plurality of the smRNAs in the panel can directly overlap with a portion of a different gene encoding a biomarker or a different locus of expression of the gene. In certain instances, each of a plurality of the smRNAs in the panel can partially overlap with a portion of a different gene encoding a biomarker or a different locus of expression of the gene.

[0043] In certain instances, a first portion of the panel of smRNAs are positively correlated with expression of the biomarker and a second portion of the panel of smRNAs are negatively correlated with expression of the biomarker. Predicting a presence of a biomarker in a patient based on the detected levels of the smRNAs described herein can comprise calculating a gene expression score. The gene expression score can be calculated based on a ratio of: (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell-free sample The gene expression score can be a log2 of the ratio. The predicting can comprise (i) predicting the patient is positive for the biomarker if the gene expression score is positive and (ii) predicting the patient is negative for the biomarker if the gene expression score is negative.

[0044] The panel of smRNAs can be positively correlated with the expression of the biomarker. The predicting can comprise calculating a gene expression score, wherein the gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample. The predicting can comprise: (i) predicting the patient is positive for the biomarker if the gene expression score is above a threshold value and (ii) predicting the patient is negative for the biomarker if the gene expression score is below the threshold value. The sequencing described herein can comprise sequencing by synthesis.

[0045] The smRNAs described herein can have a length of at least about 250 nucleotides, at least about 200 nucleotides, at least about 150 nucleotides, at least about 100 nucleotides, at least about 50 nucleotides, at least about 40 nucleotides, at least about 30 nucleotides, at least about 20 nucleotides, or at least about 10 nucleotides. The smRNAs described herein can have a length of at most about 250 nucleotides, at most about 200 nucleotides, at most about 150 nucleotides, at most about 100 nucleotides, at most about 50 nucleotides, at most about 40 nucleotides, at most about 30 nucleotides, at most about 20 nucleotides, or at most about 10 nucleotides. In certain instances, the smRNAs have a length from about 20 to about 200 nucleotides. In certain instances, the smRNAs have a length from about 30 to about 150 nucleotides. In certain instances, the smRNAs have a length form about 50 to about 100 nucleotides.Attorney Docket No. 61740-709601Biomarkers

[0046] The biomarker can comprise estrogen receptors (ERs). The detecting levels of small noncoding ribonucleic acids (smRNAs) in the cell-free sample of the patient can comprise subjecting the smRNAs to reverse transcription to generate complementary deoxyribonucleic acid (cDNA) molecules and sequencing the one or more cDNA molecules or derivatives thereof.

[0047] The predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs can comprise, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs. The panel of smRNAs can be associated with ESRI gene expression. The panel of smRNAs can comprise any smRNAs complementary to the cDNA sequences listed in Table 1 (SEQ ID NOs: 1-1763). The panel of smRNAs can comprise any smRNA sequence for which at least 50% of the smRNA sequence is complementary to a sequence of any one of the chromosomal regions listed in Table 2.

[0048] The panel of smRNAs can be associated with expression of estrogen receptor alpha (ESRI). The panel of smRNAs can comprise from about 20 smRNAs to about 80 smRNAs, from about 40 smRNAs to about 100 smRNAs, from about 60 smRNAs to about 120 smRNAs, about 80 smRNAs to about 140 smRNAs, or from about 100 smRNAs to about 160 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 50 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 100 smRNAs. The panel of smRNAs can comprise from about 200 smRNAs to about 300 smRNAs, from about 350 smRNAs to about 450 smRNAs, from about 500 smRNAs to about 650 smRNAs, about 700 smRNAs to about 850 smRNAs, or from about 900 smRNAs to about 1000 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 250 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 500 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 1000 smRNAs. The panel of smRNAs can comprise from about 500 to about 1,500 smRNAs, from about 2,000 to about 3,000 smRNAs, from about 3,500 to about 4,500 smRNAs, from about 5,000 to about 6,000 smRNAs, from about 6,500 to about 7,000 smRNAs, from about 7,500 to about 8,500 smRNAs, or from about 9,000 to about 10,000 smRNAs. The panel of smRNAs can be identified using expression quantitative trait loci (QTL) analysis. The panel of smRNAs can be identified using an association of a plurality of smRNAs detected in a cell-free sample, as described herein, with one or more known gene sets, one or more RNA isoform specific start sites, or one or more other fragmentomic features. In certain instances, one or more of the smRNAs in the panel directly overlap with a portion of a gene encoding for ESRI. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, each of the smRNAs in the panel directly overlap with a portion of a gene encodingAttorney Docket No. 61740-709601 for ESRI. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, one or more of the smRNAs in the panel directly overlap with a locus of expression of ESRI. In certain instances, each of the smRNAs in the panel directly overlap with one or more loci of expression of ESRI. In certain instances, one or more of the smRNAs in the panel partially overlap with a portion of a gene encoding for ESRI. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, each of the smRNAs in the panel partially overlap with a portion of a gene encoding for ESRI . In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, one or more of the smRNAs in the panel partially overlap with a locus of expression of ESRI. In certain instances, each of the smRNAs in the panel partially overlap with one or more loci of expression of ESRI. Partially overlapping can comprise at least or about 50% alignment, at least or about 55% alignment, at least or about 60% alignment, at least or about 70% alignment, at least or about 75% alignment, at least or about 80% alignment, at least or about 85% alignment, at least or about 90% alignment, at least or about 95% alignment, or at least or about 98% alignment of an smRNA of the panel of smRNAs with a portion of a gene encoding ESRI or a locus of expression of ESRI. Partially overlapping can comprise at most or about 50% alignment, at most or about 55% alignment, at most or about 60% alignment, at most or about 70% alignment, at most or about 75% alignment, at most or about 80% alignment, at most or about 85% alignment, at most or about 90% alignment, at most or about 95% alignment, or at most or about 98% alignment of an smRNA of the panel of smRNAs with a portion of a gene encoding ESRI or a locus of expression of ESRI. In certain instances, each of a plurality of the smRNAs in the panel can directly overlap with a portion of a different gene encoding a biomarker or a different locus of expression of the gene. In certain instances, each of a plurality of the smRNAs in the panel can partially overlap with a portion of a different gene encoding a biomarker or a different locus of expression of the gene.

[0049] In certain instances, a first portion of the panel of smRNAs are positively correlated with ESRI gene expression and a second portion of the panel of smRNAs are negatively correlated with ESRI gene expression. Predicting a presence of a biomarker in a patient based on the detected levels of the smRNAs described herein can comprise calculating an ESRI gene expression score. The ESRI gene expression score can be calculated based on a ratio of: (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell-free sample The ESRI gene expression score can be a log2 of the ratio. The positive ESRI gene expression score can indicate the patient is positive for ERs. A negative ESRI gene expression score can indicate the patient is negative for ERs.Attorney Docket No. 61740-709601

[0050] The panel of smRNAs can be positively correlated with ESRI gene expression. Predicting a presence of a biomarker in a patient based on the detected levels of the smRNAs described herein can comprise calculating an ESRI gene expression score, wherein the score is calculated based on a count of smRNAs from the panel present in the cell-free sample. The predicting can comprise (i) predicting the patient is positive for ERs if the ESRI gene expression score is above a threshold value and (ii) predicting the patient is negative for ERs if the ESRI gene expression score is below the threshold value.

[0051] A biomarker described herein can comprise human epidermal growth factor receptor-2 (HER2 / ERBB2). Predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs can comprise, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with ERBB2 gene expression. The panel of smRNAs can comprise any smRNAs complementary to the cDNA sequences listed in Table 3. The panel of smRNAs can comprise any smRNA sequence for which at least 50% of the smRNA sequence is complementary to a sequence of any of the chromosomal regions listed in Table 4. The panel of smRNAs can be positively correlated with ERBB2 gene expression. The predicting can comprise calculating an ERBB2 gene expression score, wherein the ERBB2 gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample. The predicting can comprise (i) predicting the patient is HER2 positive if the ERBB2 gene expression score is above a threshold value and (ii) predicting the patient is HER2 negative if the ERBB2 gene expression score is below the threshold value.

[0052] The panel of smRNAs can be associated with expression of ERBB2. The panel of smRNAs can comprise from about 20 smRNAs to about 80 smRNAs, from about 40 smRNAs to about 100 smRNAs, from about 60 smRNAs to about 120 smRNAs, about 80 smRNAs to about 140 smRNAs, or from about 100 smRNAs to about 160 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 50 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 100 smRNAs. The panel of smRNAs can comprise from about 200 smRNAs to about 300 smRNAs, from about 350 smRNAs to about 450 smRNAs, from about 500 smRNAs to about 650 smRNAs, about 700 smRNAs to about 850 smRNAs, or from about 900 smRNAs to about 1000 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 250 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 500 smRNAs. In certain instances, the panel of smRNAs comprises at least or about 1000 smRNAs. The panel of smRNAs can comprise from about 500 to about 1,500 smRNAs, from about 2,000 to about 3,000 smRNAs, from about 3,500 to about 4,500 smRNAs, from aboutAttorney Docket No. 61740-7096015,000 to about 6,000 smRNAs, from about 6,500 to about 7,000 smRNAs, from about 7,500 to about 8,500 smRNAs, or from about 9,000 to about 10,000 smRNAs. The panel of smRNAs can be identified using expression quantitative trait loci (QTL) analysis. The panel of smRNAs can be identified using an association of a plurality of smRNAs detected in a cell-free sample, as described herein, with one or more known gene sets, one or more RNA isoform specific start sites, or one or more other fragmentomic features. In certain instances, one or more of the smRNAs in the panel directly overlap with a portion of a gene encoding for ERBB2. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, each of the smRNAs in the panel directly overlap with a portion of a gene encoding for ERBB2. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, one or more of the smRNAs in the panel directly overlap with a locus of expression of ERBB2. In certain instances, each of the smRNAs in the panel directly overlap with one or more loci of expression of ERBB2. In certain instances, one or more of the smRNAs in the panel partially overlap with a portion of a gene encoding for ERBB2. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, each of the smRNAs in the panel partially overlap with a portion of a gene encoding for ERBB2. In certain instances, the portion of the gene is one or more exons of the gene. In certain instances, one or more of the smRNAs in the panel partially overlap with a locus of expression of ERBB2. In certain instances, each of the smRNAs in the panel partially overlap with one or more loci of expression of ESRI. Partially overlapping can comprise at least or about 50% alignment, at least or about 55% alignment, at least or about 60% alignment, at least or about 70% alignment, at least or about 75% alignment, at least or about 80% alignment, at least or about 85% alignment, at least or about 90% alignment, at least or about 95% alignment, or at least or about 98% alignment of an smRNA of the panel of smRNAs with a portion of a gene encoding ERBB2or a locus of expression of ERBB2. Partially overlapping can comprise at most or about 50% alignment, at most or about 55% alignment, at most or about 60% alignment, at most or about 70% alignment, at most or about 75% alignment, at most or about 80% alignment, at most or about 85% alignment, at most or about 90% alignment, at most or about 95% alignment, or at most or about 98% alignment of an smRNA of the panel of smRNAs with a portion of a gene encoding ERBB2 or a locus of expression of ERBB2. In certain instances, each of a plurality of the smRNAs in the panel can directly overlap with a portion of a different gene encoding a biomarker or a different locus of expression of the gene. In certain instances, each of a plurality of the smRNAs in the panel can partially overlap with a portion of a different gene encoding a biomarker or a different locus of expression of the gene.Attorney Docket No. 61740-709601

[0053] In certain instances, a first portion of the panel of smRNAs are positively correlated with ERBB2 gene expression and a second portion of the panel of smRNAs are negatively correlated with ERBB2 gene expression. Predicting a presence of a biomarker in a patient based on the detected levels of the smRNAs described herein can comprise calculating an ERBB2 gene expression score. The ERBB2gene expression score can be calculated based on a ratio of (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell-free sample The ERBB2 gene expression score can be a log2 of the ratio. The positive ERBB2 gene expression score can indicate the patient is positive for ERBB2. A negative ERBB2 gene expression score can indicate the patient is negative for ERBB2.

[0054] The panel of smRNAs can be positively correlated with ERBB2 gene expression. Predicting a presence of a biomarker in a patient based on the detected levels of the smRNAs described herein can comprise calculating an ERBB2gene expression score, wherein the score is calculated based on a count of smRNAs from the panel present in the cell-free sample. The predicting can comprise (i) predicting the patient is positive for ERBB2 if the ERBB2 gene expression score is above a threshold value and (ii) predicting the patient is negative for ERBB2 if the ERBB2gene expression score is below the threshold value.

[0055] A biomarker described herein can comprise messenger RNA (mRNA) gene expression. The biomarker can comprise pathway activation. The biomarker can comprise trophoblast cellsurface antigen 2 (TROP). The biomarker can comprise human epidermal growth factor receptor-3 (HER3). The biomarker can comprise carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5).Treatment Response

[0056] The methods described herein can further comprise predicting a response of a patient to a treatment based at least in part on the predicted presence of the biomarker. In certain instances, the biomarker comprises ESRI. The treatment can comprise administration of a selective estrogen receptor modulator. The treatment can comprise administration of an aromatase inhibitor. The treatment can comprise administration of a selective estrogen receptor degrader. The treatment can comprise administration of a cyclic-dependent kinase (CDK) inhibitor. The CDK inhibitor can comprise a CDK4 inhibitor. The CDK inhibitor can comprise a CDK6 inhibitor. The CDK inhibitor can comprise palbociclib. The CDK inhibitor can comprise ribociclib. The CDK inhibitor can comprise abemaciclib. The treatment can comprise administration of a mTOR inhibitor. The mTOR inhibitor can comprise everolimus. The treatment can comprise administration of a PI3K inhibitor. The PI3K inhibitor can compriseAttomey Docket No. 61740-709601 alpelisib. The PI3K inhibitor can comprise inavolisib. The treatment can comprise administration of an AKT inhibitor. The AKT inhibitor can comprise capivasertib. The treatment can comprise administration of sacituzumab govitecan. The treatment can comprise administration of datopotamab deruxtecan. The treatment can comprise administration of a PARP inhibitor.

[0057] In certain instances, the biomarker comprises human epidermal growth factor receptor-2 (HER2). The treatment can comprise administration of trastuzumab. The treatment can comprise administration of pertuzumab. The treatment can comprise administration of margetuximab. The treatment can comprise administration of a kinase inhibitor. The kinase inhibitor can comprise neratinib. The kinase inhibitor can comprise tucatinib. The kinase inhibitor can comprise lapatinib. The treatment can comprise administration of ado-trastuzumab emtansine. The treatment can comprise administration of fam-trastuzumab deruxtecan.Cell-Free Sample

[0058] A cell-free sample obtained from a patient can comprise an extracellular fluid. In certain instances, the extracellular fluid comprises blood. In certain instances, the extracellular fluid comprises plasma. In certain instances, the extracellular fluid comprises serum. In certain instances, the extracellular fluid comprises plasma and serum. In certain instances, the extracellular fluid comprises urine. In certain instances, the extracellular fluid comprises cerebrospinal fluid. In certain instances, the extracellular fluid comprises lymph. In certain instances, the extracellular fluid comprises saliva.

[0059] In certain instances, obtaining a cell-free sample from a patient comprises obtaining plasma from a patient, wherein the plasma comprises the cell-free sample. In certain instances, obtaining a cell-free sample from a patient comprises obtaining serum from a patient, wherein the serum comprises the cell-free sample. In certain instances, obtaining a cell-free sample from a patient comprises obtaining cerebrospinal fluid (CSF) from a patient, wherein the CSF comprises the cell-free sample. In certain instances, obtaining a cell-free sample from a patient comprises obtaining lymph from a patient, wherein the lymph comprises the cell-free sample. In certain instances, obtaining a cell-free sample from a patient comprises obtaining saliva from a patient, wherein the saliva comprises the cell-free sample. In certain instances, obtaining a cell- free sample from a patient does not comprise obtaining a tissue sample from the patient. In certain instances, a cell-free sample described herein does not comprise a tissue sample.

[0060] Obtaining a cell-free sample from a patient comprises obtaining a sufficient volume of the cell-free sample. The sufficient volume of the cell-free sample can allow for detecting levels of small non-coding ribonucleic acids (smRNAs) in the cell-free sample of the patient. In certain instances, the volume of the cell free sample is at least or about 20 microliters (pL), at least orAttorney Docket No. 61740-709601 about 50 pL, at least or about 0.1 milliliters (mL), at least or about 0.5 mL, at least or about 1 mL, at least or about 1.5 mL, at least or about 2 mL, at least or about 2.5 mL, at least or about 3 mL, at least about 3.5 mL, at least or about 4 mL, at least or about 4.5 mL, at least or about 5 mL, at least or about 5.5 mL, at least or about 6 mL, at least or about 6.5 mL, at least or about 7 mL, at least or about 7.5 mL, at least or about 8 mL, at least or about 8.5 mL, at least or about 9 mL, at least or about 9.5 mL, or at least or about 10 mL. In certain instances, the volume of the cell free sample is at most or about 0.5 mL, at most or about 1 mL, at most or about 1.5 mL, at most or about 2 mL, at most or about 2.5 mL, at most or about 3 mL, at most about 3.5 mL, at most or about 4 mL, at most or about 4.5 mL, at most or about 5 mL, at most or about 5.5 mL, at most or about 6 mL, at most or about 6.5 mL, at most or about 7 mL, at most or about 7.5 mL, at most or about 8 mL, at most or about 8.5 mL, at most or about 9 mL, at most or about 9.5 mL, or at most or about 10 mL. In certain instances, the volume of the cell-free sample is between about 20 pL to about 2 mL. In certain instances, the volume of the cell-free sample is between about 20 pL to about 0.5 mL.Nucleic Acid Sequencing

[0061] Provided herein are methods for detecting levels of small non-coding ribonucleic acids (smRNAs) in a cell-free sample of a patient. The detecting can comprise subjecting the smRNAs to reverse transcription to generate complementary deoxyribonucleic acid (cDNA) molecules and sequencing the one or more cDNA molecules or derivatives thereof. In certain instances, the detecting can comprise sequencing the smRNAs. The sequencing can comprise whole transcriptome sequencing (RNASeq), whole genome sequencing, whole exome sequencing, or any combination thereof. The sequencing can comprise sequencing by synthesis.

[0062] Disclosed herein are various methods and systems to predict gene expression from smRNAs. In some embodiments, methods and systems can comprise feature discovery methods. In some embodiments, methods and systems can comprise modeling of expression methods.

[0063] Disclosed herein are methods comprising a cell-free prediction of biomarker expression using small RNAs representing the tumor’s transcriptome from plasma or serum. In some embodiments, the method does not require prior enrichment of targeted regions. In some embodiments, the methods and systems can comprise using small, non-coding RNA, or smRNAs, which are stable and abundant in blood due to protection by extracellular vesicles and interactions with protein-lipid complexes, thereby preventing degradation by extracellular nucleases. In some embodiments, the methods and systems can comprise identifying the specific smRNAs, for example orphan non-coding RNAs, associated with the reference biomarker. In some embodiments, the reference biomarker can comprise messenger RNA (mRNA) geneAttorney Docket No. 61740-709601 expression, or pathway activation, or both. In some embodiments, specific smRNA features can be identified. In some embodiments, methods for identifying specific smRNA features can comprise expression quantitative trait loci (QTL) analysis with the mRNA. In some embodiments, methods for identifying specific smRNA features can comprise association with known gene sets. In some embodiments, methods for identifying specific smRNA features can comprise RNA isoform specific start sites. In some embodiments, methods for identifying specific smRNA features can comprise fragmentomic features. In some embodiments, the relationship between smRNA features and the mRNA or pathway expression can be subsequently modeled. In some embodiments, the relationship between smRNA features and the mRNA or pathway expression can be subsequently modeled based on a ratio between expression positively-associated to negatively-associated smRNAs. In some embodiments, the relationship between smRNA features and the mRNA or pathway expression can be subsequently modeled based on a generalized linear model or a neural net. In some embodiments, the model can be fine-tuned in a biofluid, for example plasma or serum. In some embodiments, the model can be calibrated to predict a quantitative or binary biomarker. In some embodiments, the model can be calibrated to predict one or more pathway expression read-outs. In some embodiments, the method can be applied to cancer subtypes. In some embodiments, the method can be applied to phenotypes that have been determined by an orthogonal method in, for example, tissue, cell line or plasma.

[0064] In some embodiments, the methods and systems can be used when the tumor has metastasized or when the patient cannot tolerate an invasive biopsy. In some embodiments, the methods and systems comprise a blood-based gene expression biomarker. In some embodiments, a positive read-out would make a patient eligible for an approved targeted therapy. In some embodiments, the methods and systems can comprise facilitating targeted therapies for metastatic patients, or patients with hard to reach biopsies. . In some embodiments, the methods and systems can comprise longitudinal monitoring of target engagement or pharmacodynamics, where repeat biopsies are impractical. In some embodiments, the methods and systems can comprise monitoring changes in blood-based gene expression associated with treatment strategies or intervention if the tumor dependence on gene expression has changed.

[0065] In some embodiments, the methods and systems can comprise targeted therapies including therapy based on trastuzumab deruxtecan (Enhertu), an antibody drug conjugate (ADC) HER2 targeted therapy for lung cancer using a circulating tumor DNA based test that detects HER2 mutations from blood. In some embodiments, the methods and systems canAttorney Docket No. 61740-709601 comprise a liquid biopsy test that could identify patients with high HER2 tumor expression from cell-free RNA and can identify additional patients to benefit from the therapy treatment.

[0066] In some embodiments, the methods and systems can comprise biomarker detection of at least TROP2, HER3, and CEACAM5. In some embodiments, the methods and systems can comprise cell-free RNA assays. In some embodiments, the methods and systems can comprise cell-free RNA assays comprising assaying TROP2 expression, which may enrich patients who would better respond to TROP2-targeted therapy.

[0067] In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for patient selection and enrichment. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for pharmacodynamics to assess the biomarker targets. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for predicting response to therapy. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for identifying patients resistant to therapy or a time for therapy switching in patients. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for identifying a signature of response or resistance. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for longitudinal monitoring of treatment efficacy. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for identifying patient subtypes. In some embodiments, the methods and systems can comprise various methods and systems to infer gene expression from smRNAs for inferring a single gene expression or a subtype of cancer using one or more gene expressions.Feature Discovery

[0068] Disclosed herein are methods comprising defining features from known gene sets found to be co-expressed with a biomarker in tumor tissue or cell line models. These gene sets may include a set of genes that are upregulated or downregulated in association with the gene of interest. In some embodiments, the gene of interest can be derived from a published study of gene expression in tumor studies, for example arrayCGH or RNAseq, or other published methods. In some embodiments, the methods and systems can comprise selecting smRNAs according to the tumor gene set. In some embodiments, the methods and systems can comprise modeling the features according to the previously described directions. In some embodiments, previously described directions can comprise upregulated or downregulated. In someAttorney Docket No. 61740-709601 embodiments, the features can be a ratio, a difference, for example subtraction, or using one or more machine learning models.

[0069] In some embodiments, the methods and systems can comprise selecting smRNAs that directly overlap one or more biomarkers. In some embodiments, the method can include selecting smRNAs that directly overlap one or more biomarkers and a specific locus of expression. In some embodiments, the methods and systems can comprise expression of RNAs such as smRNAs acting as a surrogate for the biomarker expression.

[0070] In some embodiments, the methods and systems can comprise alignment-free quantification of RNA sequences. In some embodiments, the RNA sequences can be significantly associated with a biomarker of interest. In some embodiments, the methods and systems can comprise modeling tumor tissue biological samples, cell lines, or plasma biological samples. In some embodiments, the methods and systems can comprise utilizing generalized linear models with confounding variables for the modeling. In some embodiments, the confounding variables can comprise, for example, number of guanines and cytosines in the sequence, number of sequences, or another compounding variable.

[0071] In some embodiments, the methods and systems can comprise performing expression quantitative trait loci (eQTL) analysis. In some embodiments, performing eQTL analysis can comprise a de novo discovery of smRNAs that are significantly associated with a biomarker of interest. In some embodiments, the smRNAs or biomarker of interest can be in tumor tissue, cell lines, plasma, or other biological material. In some embodiments, the methods and systems can comprise de-novo forming of unannotated smRNA loci through peak calling. In some embodiments, the methods and systems can comprise a transcriptome-wide analysis testing each de-novo loci for an association with a biomarker utilizing generalized linear models with correction for known confounders. In some embodiments, the methods and systems can comprise generating results comprising a set of smRNAs that are associated positively with expression of the biomarker of interest. In some embodiments, the methods and systems can comprise generating results comprising a set of smRNAs that are associated negatively with the biomarker of interest. In some embodiments, the methods and systems can comprise generating smRNA- QTLs.

[0072] In some embodiments, the methods and systems can comprise counting the number of times smRNA sequences start or end at a specific genomic position in tumor tissue, cell lines, plasma, or other biological sample. In some embodiments, the methods and systems can comprise using statistical tests to link precise start or end positions to the biomarker of interest, in some cases limiting generic cancer-associated signals. In some embodiments, the methods andAttorney Docket No. 61740-709601 systems can comprise quantifying underlying fragmentomic characteristics or mRNA isoforms specific to the cancer subtype.Expression Modeling

[0073] In some embodiments, the methods and systems can comprise modeling expression based on the feature set. In some embodiments, the methods and systems can comprise taking the log transformed ratio of smRNAs associated with gene expression in the positive (upregulated) direction to smRNAs associated with gene expression in the negative (downregulated) direction.

[0074] In some embodiments, the methods and systems can comprise modeling expression based on the feature set. In some embodiments, the methods and systems can comprise scoring each sample using the fraction of reads included in the feature set divided by a normalizing factor. In some embodiments, the normalizing factor can comprise the total number of mapped reads, or the total number of reads of the same type of RNA as was curated, or both.

[0075] In some embodiments, the methods and systems can comprise modeling expression based on the feature set. In some embodiments, the methods and systems can comprise applying a regularized linear regression or neural network optimized by cross validation to predict biomarker expression values.EXAMPLES

[0076] The following examples are provided to further illustrate some embodiments of the present disclosure, but are not intended to limit the scope of the disclosure; it will be understood by their exemplary nature that other procedures, methodologies, or techniques known to those skilled in the art may alternatively be used.Example 1 - smRNA-Derived Scores Correlated with ESRI and ERBB2 Expression in Tumors and Cell Lines

[0077] Treatment selection in advanced breast cancer is strongly dependent upon protein expression of Estrogen receptor (ER) and Human Epidermal receptor 2 (HER2).Provided herein is a method for predicting ER / ESR1 or HER2 / ERBB2 status non-coding small RNAs (smRNAs).Computation of ESR1 / ERBB2 Scores by Expression Quantitative Trait Loci (eQTL) Analysis

[0078] The expression of smRNAs overlapping with 27 ERBB2 exons based on Ensembl transcript ID ENST00000269571.10 was quantified. Table 4 provides a list of the 27 ERBB2 exons. Genes with <20 reads that were expressed in <20% of samples were removed from the RNA-seq data, following GTEx guidelines. Expression data then underwent Trimmed Mean ofAttorney Docket No. 61740-709601M-values (TMM) normalization and rank inverse normal transformation. For small RNA-seq data, 1.2% known and 98.8% de-novo smRNAs within the TCGA cohort were identified. Following quality control, associations between expression of ~530k smRNAs and ERBB2 or ESRI RNA expression were determined. For each tumor tissue, a quantitative trait loci (QTL) analysis was performed using QTLTools [Delaneau O., Ongen H., Brown A. A., et al. A complete tool set for molecular QTL discovery and analysis. Nat Commun 8, 15452 (2017)], for all smRNAs, including in cis smRNAs (e.g., within a 1Mb region of each gene's transcription start site) and in trans smRNAs (e.g., not within 1Mb of each gene's transcription start site), to identify the top smRNA-gene associations, corrected for multiple testing through a permutation schema, within each gene for cis analyses and across genes for trans analyses. Multiple testing across gene tests was corrected for in the cis analysis, reporting those c7.s-smR.NAs and gene pairs with a significant FDR adjusted p-value (q<0.1). For / ra / z.s-QTLs, gene expression was permuted and used to generate a null distribution, with FDR adjusted p-values averaged across 200 permutations and significant (q<0.1) results reported. These resulted in 1,763 smRNAs positively associated with ER expression and 687 smRNAs positively associated with ERBB2 expression, as identified by eQTL analysis. Table 1 provides the 1,763 smRNAs positively associated with ER expression, where each smRNA is presented as its corresponding cDNA sequence (SEQ ID NOs: 1-1763). Table 3 provides the 687 smRNAs positively associates with ERBB2 expression, where each smRNA is presented as its corresponding cDNA sequence (SEQ ID NOs: 1764-2450). From the eQTL analysis, the total expression of smRNAs associated with ER (ER smRNA-QTLs) or ERBB2 (ERBB2 smRNA-QTLs) was used as the predicted expression of that gene.

[0079] FIG. 1A shows a Pearson’s correlation (P = 0.835, p-value = 1.03e-268) between the predicted ESRI gene expression (using smRNA-QTLs) against known ESRI gene expression, as determined by TCGA. Light gray dots show samples confirmed to be ER- by H4C analysis. Black dots show samples confirmed to be ER+ by H4C analysis. FIG. IB shows Mann-Whitney- Wilcoxon (MWW) analysis (p = 1.341e-90) of the predicted ESRI gene expression (using smRNA-QTLs) compared to the known ESRI gene expression, as determined by TCGA. These results show that smRNA-QTLs are positively correlated with ESRI gene expression and are predictive of ER+ status.

[0080] FIG. 2A shows a Pearson correlation (P = 0.63, p-value = 3.39e-120) between the predicted ERBB2 gene expression (using smRNA-QTLs) against known ERBB2 gene expression, as determined by TCGA. Light gray dots show samples confirmed to be HER-, determined by IHC analysis. Mid-gray dots show samples confirmed to have low HERAttorney Docket No. 61740-709601 expression, determined by IHC analysis. Black dots show samples confirmed with overexpression of HER, determined by IHC analysis. FIG. 2B shows Mann-Whitney -Wilcoxon (MWW) analysis of the predicted ERBB2 gene expression (using smRNA-QTLs) compared to the known ERBB2 gene expression, as determined by TCGA. These results show that smRNA- QTLs are positively correlated with ERBB2 gene expression and are predictive of HER+ status.Correlation of ERBB2 smRNA Counts Per Million (CPM) to ERBB2 Status in Various Cancer Cell LinesTo confirm extracellular secretion of smRNAs for ER and HER2 scoring, gene biomarker scores were generated from breast cancer (n=8) and human mammary epithelial (n=2) cell lines and the paired conditioned media.

[0081] Intracellular and extracellular (e.g., secreted) ERBB2 (HER2) expression was determined (via TCGA) by immunohistochemistry (IHC) in the following Luminal A breast cancer cell lines: a human ductal carcinoma epithelial cell line (ZR-75-1), a human infiltrating ductal carcinoma cell line (T47D), and a metastatic adenocarcinoma breast cancer cell line (MCF-7); in the following Luminal B breast cancer cell lines: a mammary gland ductal carcinoma cell line (ZR-75-30), and a solid, invasive ductal carcinoma cell line (BT-474); in the following HER2+ cell lines: an adenocarcinoma cell line (SK-BR-3), a ductal carcinoma epithelial cell line, and another ductal carcinoma epithelial cell line (HCC1945); and in the following two triple-negative breast cancer (TNBC) cell lines: HCC38 and HCC1143, as compared to human mammary epithelial cells (MHEC; normal) and a non-tumorigenic epithelial cell line (MCF10A; immortalized). ERBB2 smRNA counts per million (CPM) were determined by summating the smRNA reads that aligned with an exon of ERBB2. Table 4 provides the 27 ERBB2 exons used in this analysis (SEQ ID NOs: 2451-2477). Alignment for the smRNAs to any one of the 27 ERBB2 exons was thresholded for at least 50% alignment to any portion of an ERBB2 exons, as listed in Table 4. FIG. 3 shows intracellular and extracellular ERBB2 (HER2) status correlated to ERBB2 smRNA CPM, for the cell lines described herein. FIG. 4 shows ERBB2 (HER2) status from the extracellular fraction, correlated to ERBB2 smRNA CPM, for the cell lines presented in FIG. 3, as determined by the Cancer Dependency Map Project (DepMap). FIG. 5 shows intracellular and extracellular ERBB2 (HER2) status from the cell lines presented in FIG. 3, as determined by DepMap, stratified by negative status positive status from 0-85thpercentile expression, and positive status from 85-100thpercentile expression, against ERBB2 smRNA CPM, generated as described herein. Increased median of EBBR2 smRNA CPM in EBBR2- positive cell lines with a higher expression of EBBR2 suggests a strong association between EBBR2 activity and EBBR2 smRNA CPM. These data show that EBBR2 smRNA CPMAttorney Docket No. 61740-709601 correlates with EBBR2 expression both extracellularly and intracellularly in the cell lines tested. FIG. 6 shows intracellular and extracellular ER status correlated to ER smRNA CPM, generated as described herein (using ER exons), for the cell lines described herein, suggesting that ER smRNA CPM also is predictive of ER status.

[0082] To determine the relationship of gene biomarker scores and gene expression levels, smRNAs isolated from cell line conditioned media were titrated into a triple-negative background cell line (HC1143). FIG. 7A shows ER expression scores in HC1143 cells across titration levels of smRNAs from MCF7 and ZR7530 conditioned media. FIG. 7B shows HER2 expression scores in HC1143 cells across titration levels of smRNAs from HCC1954 and SKBR3 conditioned media. These results suggest that ER-associated smRNAs and HER2- associated smRNAs can induce expression of ER and HER2, respectively, in a triple-negative breast cancer cell line.

[0083] FIGs. 8A-8B show ESRI smRNA CPM scores from TCGA plasma samples and stratified by known ER status from H4C on these patients’ tumors, patients across Stage I / I I tumor status (FIG. 8A) and Stage III / IV tumor status (FIG. 8B). FIGs. 9A-9B show ERBB2 smRNA CPM scores from TCGA plasma samples and stratified by known HER2 status from H4C on these patients’ tumors, across Stage I / I I tumor status (FIG. 9A) and Stage III / IV tumor status (FIG. 9B). FIGs. 10A-10B shows a correlation of ESRI smRNA CPM from two replicate plasma samples from each patient, as provided by TCGA, across tumor Stage VII (FIG. 10A) and tumor Stage IIVIV (FIG. 10B). FIGs. 11A-11B shows a correlation of ERBB2 smRNA CPM from two replicate plasma samples from each patient, as provided by TCGA, across tumor Stage VII (FIG. 11 A) and tumor Stage IIVIV (FIG. 11B). These results show that ESRI smRNA CPM and ERBB2 smRNA CPM correlate with known ER or HER2 status, respectively, from plasma-derived smRNA samples.Correlation of EBBR2 smRNA CPM to EBBR2 Status from Plasma Samples

[0084] Plasma samples were collected from treatment-naive donors under IRB-approved protocol, processed according to standard clinical procedures, and stored at -80°C. Blood was drawn into EDTA plasma and Streck plasma tubes. The resulting samples underwent bead-based nucleic acid extraction including on -bead DNase I treatment. Total cell-free RNA (cfRNA) was extracted from 0.7-0.8mL of plasma using the Promega Maxwell® RSC miRNA Plasma and Serum kit. Small RNA (smRNA) libraries were generated using ligation-free SMART er® smRNA-seq kit for Illumina®, using custom reverse transcription (RT) primers. Libraries were indexed by PCR, purified, quantified using the Agilent Tapestation® (High Sensitivity DI 000 Screentape), and sequenced on an Illumina® NextSeq® 2000 or NovaSeq® 6000 platform usingAttorney Docket No. 61740-7096011 OO-bp single-end reads, targeting approximately 50 Million reads per sample. FIG. 12 illustrates an example workflow schematic for sample collection and smRNA library generation.Bioinformatic processing

[0085] Raw binary base call (BCL) files from the generated smRNA libraries were converted / demultiplexed (e.g., using a standard module, such as bclconvert), adapters were trimmed (e.g., using a standard module, such as cutadapt), BCL files were quality-checked (e.g., using a standard module, such as FastQC), Unique Molecular Identifiers (UMIs) were collapsed (e.g., using a standard module, such as UMICollapse), reads were aligned (e.g., using a standard module, such as Bowtie2), and Binary Alignment / Map (BAM) / feature tables were produced (e.g., using a standard module, such as samtools / pysam / bedtools). Versioned tools and exact parameters were maintained in the analysis manifest.

[0086] EBBR2 smRNA CPM, generated as described herein, from patients’ plasma samples (HR- status) as described herein, were correlated to EBBR2 expression status in stage II, II, and IV HER2- breast cancer tumor tissue, as determined by IHC, from the same patients. FIG. 13 shows that negative EBBR2 smRNA CPM from these patients correlated with negative EBBR2 status, and positive EBBR2 smRNA CPM scores correlated with positive ERBB2 status across tumor stages.Example 2 - Exon-Mapping of smRNAs is Predictive of ESRI StatusESRI Expression Scores

[0087] A list of 372 genes associated with breast cancer clinical outcomes were identified according to Van't Veer, Laura J., et al. Nature 415.6871 (2002): 530-536. These 372 genes were determined to have 4,396 exons, for which 1,966 showed positive correlation with ESRI expression, and for which 2,403 showed negative correlation with ESRI expression. smRNAs as described herein were aligned to the 4,396 exons, wherein alignment was thresholded as at least 50% of the smRNA sequence having alignment to any portion of an exon. Table 2 provides a list of the 4,396 exons, with 1,966 exons positively correlated with ESRI expression and 2,403 exons negatively correlated to ESRI expression. A Van’t Veer gene score (VV gene score) was defined as the log2 of the average fold change in expression of smRNAs that aligned with an exon positively correlated with ESRI expression versus the average fold change in expression of smRNAs expression of those that aligned with an exon negatively correlated with ESRI expression. FIG. 14 shows, for each subject, discrimination of ESRI (ER) status by ESRI VV gene score in HER2- tumor tissue at tumor stage 1 (n=<100), stage II (n=317), stage III (n=123), and stage IV (n=6). ER status and smRNA data were provided by TCGA. These results showAttorney Docket No. 61740-709601 that ESRI VV gene scores were lower in ESRI (ER) negative tumor tissue, while VV ERpositive gene scores were higher in ER positive tumor tissue across tumor stages [area under the curve (AUC)=0.957 across all stages], in each subject. FIG. 15 shows Pearson’s correlation (R = 0.689) of ESRI expression z-scores to ESRI W gene scores. The z-score of ESRI gene expression was based on TCGA normalized gene expression values (trimmed mean of M-values (TMM) normalization). These results suggest that ESRI z-scores and VV ESRI gene score are positively correlated, and are able to discriminate ER-negative status from ER-positive status.Correlation of ESRI VV Gene Score to ER Status in Various Cancer Cell Lines

[0088] Intracellular and extracellular (e.g., secreted) ER expression was determined by immunohistochemistry (IHC) in the following Luminal A breast cancer cell lines: a human ductal carcinoma epithelial cell line (ZR-75-1), a human infiltrating ductal carcinoma cell line (T47D), and a metastatic adenocarcinoma breast cancer cell line (MCF-7); in the following Luminal B breast cancer cell lines: a mammary gland ductal carcinoma cell line (ZR-75-30), and a solid, invasive ductal carcinoma cell line (BT-474); in the following HER2+ cell lines: an adenocarcinoma cell line (SK-BR-3), a ductal carcinoma epithelial cell line, and another ductal carcinoma epithelial cell line (HCC1945); and in the following triple-negative breast cancer (TNBC) cell lines: two mammary gland epithelial cell line (HCC38 and HCC1143), as compared to human mammary epithelial cells (MHEC; normal) and a non-tumorigenic epithelial cell line (MCF10A; immortalized). FIG. 16 shows intracellular and extracellular ESRI (ER) status correlated to ESRI VV gene score, for the cell lines described herein. FIG. 17 shows intracellular and extracellular ESRI expression from the cell lines presented in FIG. 16, as determined by the Cancer Dependency Map Project (DepMap), stratified by negative status, positive status from 0-85thpercentile expression, and positive status from 85-100thpercentile expression, against ESRI VV gene scores, generated as described herein. Increased median of VV gene score in ER-positive cell lines with a higher expression of ESRI suggests a strong association between ESRI activity and VV gene score. These data show that ESRI VV gene scores correlate with ER expression both extracellularly and intracellularly in the cell lines tested.Correlation of ESRI VV Gene Score to ER Status in Human Tumor Samples

[0089] ESRI VV gene scores, generated as described herein, using smRNAs derived from patients’ plasma samples, were correlated to ER expression status in stage II, II, and IV HER2- breast cancer tumor tissue, as determined by IHC, from the same patients. FIG. 18 shows that negative ESRI VV gene scores from these patients correlated with negative ER status, and positive ESRI VV gene scores correlated with positive ER status across tumor stages.Attorney Docket No. 61740-709601

[0090] While preferred embodiments of the present disclosure have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the disclosure. It should be understood that various alternatives to the embodiments of the present disclosure may be employed in practicing the present disclosure. It is intended that the following claims define the scope of the present disclosure and that methods and structures within the scope of these claims and their equivalents be covered thereby.Table 1: smRNAs Positively Associated with ESRI Expression, as Determined by eQTL AnalysisAttorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Table 2: ESRI Exons used to Determine ER Van’t Veer Gene ScoresAttorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Table 3: smRNAs Positively Associated with ERBB2 Expression, as Determined by eQTL AnalysisAttorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601Table 4: ERBB2 Exons used for ERBB2 smRNA CPM AnalysisAttorney Docket No. 61740-709601Attorney Docket No. 61740-709601Attorney Docket No. 61740-709601

Claims

Attorney Docket No. 61740-709601CLAIMSWhat is claimed is:

1. A method for predicting presence of a biomarker, comprising:(a) obtaining a cell-free sample from a patient;(b) detecting levels of small non-coding ribonucleic acids (smRNAs) in the cell-free sample of the patient; and(c) predicting the presence of the biomarker in the patient based on the detected levels of the smRNAs.

2. The method of claim 1, wherein (b) comprises subjecting the smRNAs to reverse transcription to generate complementary deoxyribonucleic acid (cDNA) molecules and sequencing the one or more cDNA molecules or derivatives thereof.

3. The method of claim 2, wherein (c) comprises, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with expression of the biomarker.

4. The method of claim 3, wherein the panel comprises at least 50 smRNAs.

5. The method of claim 3, wherein the panel comprises at least 100 smRNAs.

6. The method of claim 3, wherein the panel comprises at least 250 smRNAs.

7. The method of claim 3, wherein the panel comprises at least 500 smRNAs.

8. The method of claim 3, wherein the panel comprises at least 1000 smRNAs.

9. The method of claim 3, wherein the panel of smRNAs is identified using expression quantitative trait loci (QTL) analysis.

10. The method of claim 3, wherein the panel of smRNAs is identified using association of smRNAs with one or more known gene sets, one or more RNA isoform specific start sites, or other fragmentomic features, thereby obtaining the panels of smRNAs.

11. The method of claim 3, wherein the smRNAs in the panel directly overlap with the biomarker and a specific locus of expression.

12. The method of claim 3, wherein a first portion of the panel of smRNAs are positively correlated with expression of the biomarker, and wherein a second portion of the panel of smRNAs are negatively correlated with expression of the biomarker.

13. The method of claim 12, wherein (c) comprises calculating a gene expression score, wherein the gene expression score is calculated based on a ratio of: (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell-free sample.Attorney Docket No. 61740-70960114. The method of claim 13, wherein the gene expression score is a log2 of the ratio.

15. The method of claim 14, wherein (c) comprises: (i) predicting the patient is positive for the biomarker if the gene expression score is positive and (ii) predicting the patient is negative for the biomarker if the gene expression score is negative.

16. The method of claim 3, wherein the panel of smRNAs are positively correlated with the expression of the biomarker.

17. The method of claim 16, wherein (c) comprises calculating a gene expression score, wherein the gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample.

18. The method of claim 17, wherein (c) comprises: (i) predicting the patient is positive for the biomarker if the gene expression score is above a threshold value and (ii) predicting the patient is negative for the biomarker if the gene expression score is below the threshold value.

19. The method of any one of claims 2-18, wherein the sequencing comprises sequencing by synthesis.

20. The method of any of the preceding claims, wherein the biomarker comprises estrogen receptors (ERs).

21. The method of claim 20, wherein (c) comprises, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with ESRI gene expression.

22. The method of claim 21, wherein the panel of smRNAs comprises any smRNAs complementary to the cDNA sequences listed in Table 1.

23. The method of claim 21, wherein the panel of smRNAs comprises any smRNA sequence for which at least 50% of the smRNA sequence is complementary to a sequence of any of the chromosomal regions listed in Table 2.

24. The method of claim 21, wherein a first portion of the panel of smRNAs are positively correlated with ESRI gene expression, and wherein a second portion of the panel of smRNAs are negatively correlated with ESRI gene expression.

25. The method of claim 24, wherein (c) comprises calculating an ESRI gene expression score, wherein the ESRI gene expression score is calculated based on a ratio of: (i) a first count of smRNAs from the first portion of the panel present in the cell-free sample and (ii) a second count of smRNAs from the second portion of the panel present in the cell- free sample.

26. The method of claim 25, wherein the ESRI gene expression score is a log2 of the ratio.Attorney Docket No. 61740-70960127. The method of claim 26, wherein a positive ESRI gene expression score indicates the patient is positive for ERs.

28. The method of claim 26, wherein a negative ESRI gene expression score indicates the patient is negative for ERs.

29. The method of claim 21, wherein the panel of smRNAs are positively correlated with ESRI gene expression.

30. The method of claim 29, wherein (c) comprises calculating an ESRI gene expression score, wherein the ESRI gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample.

31. The method of claim 30, wherein (c) comprises: (i) predicting the patient is positive for ERs if the ESRI gene expression score is above a threshold value and (ii) predicting the patient is negative for ERs if the ESRI gene expression score is below the threshold value.

32. The method of any of the preceding claims, wherein the biomarker comprises human epidermal growth factor receptor-2 (HER2).

33. The method of claim 32, wherein (c) comprises, based on the sequencing, determining whether the smRNAs contain one or more smRNAs from a panel of smRNAs, wherein the panel of smRNAs are associated with ERBB2 gene expression.

34. The method of claim 33, wherein the panel of smRNAs comprises any smRNAs complementary to the cDNA sequences listed in Table 3.

35. The method of claim 33, wherein the panel of smRNAs comprises any smRNA sequence for which at least 50% of the smRNA sequence is complementary to a sequence of any of the chromosomal regions listed in Table 4.

36. The method of claim 33, wherein the panel of smRNAs are positively correlated with ERBB2 gene expression.

37. The method of claim 36, wherein (c) comprises calculating an ERBB2 gene expression score, wherein the ERBB2 gene expression score is calculated based on a count of smRNAs from the panel present in the cell-free sample.

38. The method of claim 37, wherein (c) comprises: (i) predicting the patient is HER2 positive if the ERBB2 gene expression score is above a threshold value and (ii) predicting the patient is HER2 negative if the ERBB2 gene expression score is below the threshold value.

39. The method of any of the preceding claims, wherein the biomarker comprises messenger RNA (mRNA) gene expression.Attorney Docket No. 61740-70960140. The method of any of the preceding claims, wherein the biomarker comprises pathway activation.

41. The method of any of the preceding claims, wherein the biomarker comprises trophoblast cell-surface antigen 2 (TROP2).

42. The method of any of the preceding claims, wherein the biomarker comprises human epidermal growth factor receptor-3 (HER3).

43. The method of any of the preceding claims, wherein the biomarker comprises carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5).

44. The method of any of the preceding claims, wherein (a) comprises obtaining plasma from the patient, wherein the plasma comprises the cell-free sample.

45. The method of any of the preceding claims, wherein (a) comprises obtaining serum from the patient, wherein the serum comprises the cell-free sample.

46. The method of any of the preceding claims, wherein in (a), a tissue sample from the patient is not obtained.

47. The method of any of the preceding claims, wherein (a) comprises obtaining cerebrospinal fluid (CSF) from the patient, wherein the CSF comprises the cell-free sample.

48. The method of any of the preceding claims, wherein (a) comprises obtaining urine from the patient, wherein the urine comprises the cell-free sample.

49. The method of any of the preceding claims, wherein (a) comprises obtaining lymph from the patient, wherein the lymph comprises the cell-free sample.

50. The method of any of the preceding claims, wherein (a) comprises obtaining saliva from the patient, wherein the saliva comprises the cell-free sample.

51. The method of any of the preceding claims, wherein a total volume of the cell-free sample is between about 20 microliters and about 2 milliliters.

52. The method of any of the preceding claims, wherein a total volume of the cell-free sample is between about 100 microliters to about 500 microliters.

53. The method of any of the preceding claims, wherein the method further comprises (d) predicting a response of the patient to a treatment based at least in part on the predicted presence of the biomarker obtained in (c).

54. The method of claim 53, wherein the biomarker comprises estrogen receptors (ERs).

55. The method of claim 53 or 54, wherein the treatment comprises administration of a selective estrogen receptor modulator.

56. The method of claim 53 or 54, wherein the treatment comprises administration of an aromatase inhibitor.Attorney Docket No. 61740-70960157. The method of claim 53 or 54, wherein the treatment comprises administration of a selective estrogen receptor degrader.

58. The method of claim 53 or 54, wherein the treatment comprises administration of a cyclic-dependent kinase (CDK) inhibitor.

59. The method of claim 58, wherein the CDK inhibitor comprises a CDK4 inhibitor.

60. The method of claim 58, wherein the CDK inhibitor comprises a CDK6 inhibitor.

61. The method of claim 58, wherein the CDK inhibitor comprises palbociclib.

62. The method of claim 58, wherein the CDK inhibitor comprises ribociclib.

63. The method of claim 58, wherein the CDK inhibitor comprises abemaciclib.

64. The method of claim 53 or 54, wherein the treatment comprises administration of a mTOR inhibitor.

65. The method of claim 64, wherein the mTOR inhibitor comprises everolimus.

66. The method of claim 53 or 54, wherein the treatment comprises administration of a PI3K inhibitor.

67. The method of claim 66, wherein the PI3K inhibitor comprises alpelisib.

68. The method of claim 66, wherein the PI3K inhibitor comprises inavolisib.

69. The method of claim 53 or 54, wherein the treatment comprises administration of an AKT inhibitor.

70. The method of claim 69, wherein the AKT inhibitor comprises capivasertib.

71. The method of claim 53 or 54, wherein the treatment comprises administration of sacituzumab govitecan.

72. The method of claim 53 or 54, wherein the treatment comprises administration of datopotamab deruxtecan.

73. The method of claim 53 or 54, wherein the treatment comprises administration of a PARP inhibitor.

74. The method of claim 53, wherein the biomarker comprises human epidermal growth factor receptor-2 (HER2).

75. The method of claim 53 or 74, wherein the treatment comprises administration of trastuzumab.

76. The method of claim 53 or 74, wherein the treatment comprises administration of pertuzumab.

77. The method of claim 53 or 74, wherein the treatment comprises administration of margetuximab.

78. The method of claim 53 or 74, wherein the treatment comprises administration of a kinase inhibitor.Attorney Docket No. 61740-70960179. The method of claim 78, wherein the kinase inhibitor comprises neratinib.

80. The method of claim 78, wherein the kinase inhibitor comprises tucatinib.

81. The method of claim 78, wherein the kinase inhibitor comprises lapatinib.

82. The method of claim 53 or 74, wherein the treatment comprises administration of ado- trastuzumab emtansine.

83. The method of claim 53 or 74, wherein the treatment comprises administration of famtrastuzumab deruxtecan.

84. The method of any of the preceding claims, wherein the smRNAs have a length of less than 200 nucleotides.

85. The method of any of the preceding claims, wherein the smRNAs have a length from 50 to 100 nucleotides.

Images