Systems and methods for analysis of small RNA k-mers

Abstract

The present disclosure relates generally to the use of small RNA biomarkers in a biological sample for diagnosis of a subject based upon the detection of nucleic acid sequences in a sample, including the identification and use of molecular biomarkers for diagnosing, prognosing, or monitoring a cancer state in the subject. The disclosure also provides methods for processing biological samples to provide the foregoing information.

Description

SYSTEMS AND METHODS FOR ANALYSIS OF SMALL RNA K-MERSCROSS REFERENCE

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 510,570 filed June 27, 2023, which is incorporated herein by reference in its entirety.BACKGROUND

[0002] The widespread reprogramming of the gene expression landscape may be a hallmark of cancer development. Thus, the systematic identification of regulatory pathways that drive pathologic gene expression patterns may be a crucial step toward understanding and treating cancer. A multitude of regulatory mechanisms may be implicated in oncogenic expression of genes involved in cancer cell differentiation, survival, invasion, and spread. While transcriptional pathways that underlie oncogenesis may be studied, post-transcriptional regulatory pathways have also emerged as major regulators of this process. For example, microRNAs (small non-coding RNAs), a subclass of small RNAs that function in gene silencing, were among the first characterized post-transcriptional regulators of breast cancer progression. RNA-binding proteins (RBPs) are also critical posttranscriptional regulators of gene expression, and RBPs may be shown to affect oncogenesis and cancer progression. Further, other classes of small non-coding RNAs, such as tRNAs and tRNA fragments, may play a fundamental role in breast cancer progression.SUMMARY

[0003] Despite the diverse repertoire of regulatory mechanisms involved in cancers, a shared characteristic among them is that they co-opt and dysregulate existing pathways within the cell. In other words, cancer cells adopt myriad strategies, such as somatic mutations (e.g. KRAS), gene fusions (e.g. BCR-ABL), epigenetic modifications (e.g. promoter hypermethylation), and regulatory mechanisms disruptions (NFkB transcription factors) to overactivate oncogenic and to down-regulate tumor suppressive pathways. While these strategies rely on the pathologic modulation of regulatory programs that are already in place, there is an often-overlooked possibility that cancer cells may be capable of evolving or engineering specialized regulatory pathways that drive tumorigenesis.

[0004] The present disclosure provides systems and methods for analyzing small RNA sequences that serve as biomarkers which are indicative of cancer, tissue of origin, and cancer status of a subject, which may be used to accurately diagnose cancer even at an early stage. In some embodiments, the methods comprise use of cell-free small RNAs in a suitablesample. In some embodiments, the sample is a biological sample. In some embodiments, the biological sample is a whole blood, serum, or plasma sample.

[0005] In an aspect, the present disclosure provides a computer-implemented method of detecting a presence or an absence of cancer in a subject, (a) obtaining a plurality of sequencing reads derived from cell-free RNA of the subject; (b) processing the sequencing reads to generate a plurality of k-mer sequences, wherein the processing is performed without mapping or aligning the sequencing reads to a reference genome sequence; (c) analyzing the plurality of k-mer sequences using a trained machine learning classifier, wherein the trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject, based at least in part on presence or absence of cancer-associated k-mer sequences in the plurality of k-mer sequences; and (d) detecting the presence or absence of cancer in the subject, based at least in part on the output generated in (c). In some embodiments, the cell-free RNA comprises a small RNA. In some embodiments, the method further comprises, prior to (a), sequencing the cell-free RNA to generate the plurality of sequencing reads. In some embodiments, the method further comprises, prior to (a), extracting the cell-free RNA from a biological sample of the subject. In some embodiments, the method further comprises, prior to (a), isolating the cell-free RNA from a biological sample of the subject. In some embodiments, the method further comprises, prior to (a), obtaining a biological sample from the subject comprising the cell -free RNA.

[0006] In some embodiments, the biological sample is a serum sample or plasma sample. In some embodiments, the cancer is selected from the group consisting of brain cancer, breast cancer, colon cancer, liver cancer, lung cancer, kidney cancer, prostate cancer, ovarian cancer, spleen cancer, lymphoma, thyroid cancer, pancreatic cancer, intestinal cancer, laryngeal cancer, esophageal cancer, stomach cancer, small intestinal cancer. In some embodiments, the cancer comprises prostate cancer.

[0007] In some embodiments, analyzing the plurality of k-mer sequences comprises analyzing k-mer sequences having a length of at least about 5 nucleotides to about 50 nucleotides (e.g., at least about 5 nucleotides, at least about 10 nucleotides, at least about 15 nucleotides, at least about 20 nucleotides, at least about 25 nucleotides, at least about 30 nucleotides, at least about 35 nucleotides, at least about 40 nucleotides, at least about 45 nucleotides, or at least about 50 nucleotides). In some embodiments, analyzing the plurality of k- mer sequences comprises analyzing k-mer sequences having a length of at least about 10 nucleotides. In some embodiments, analyzing the plurality of k-mer sequences comprisesanalyzing k-mer sequences having a length of about 5 nucleotides, about 10 nucleotides, about 15 nucleotides, about 20 nucleotides, about 25 nucleotides, about 30 nucleotides, about 35 nucleotides, about 40 nucleotides, about 45 nucleotides, or about 50 nucleotides.

[0008] In some embodiments, the trained machine learning classifier is trained on a set of k-mer sequences from sequence reads from cancer samples. In some embodiments, the trained learning classifier is trained on a set of k-mer sequences from sequence reads from noncancer samples. In some embodiments, the method further comprises, prior to (c), training the trained machine learning classifier using sequencing reads of a non-cancer sample and a cancer sample. In some embodiments, the machine learning classifier comprises using multiple linear regression (MLR), partial least squares (PLS) regression and principal components regression (PCR), binary decision trees, artificial neural networks, generative Al models, back propagation networks, discriminant analyses, logistic classifiers, and support vector classifiers. In some embodiments, (c) further comprises determining a number of the sequencing reads comprising a k-mer sequence.

[0009] In some embodiments, a k-mer sequence of the plurality of k-mer sequences does not align to a wild type human reference genome. In some embodiments, the a k-mer sequence of the plurality of k-mer sequences comprises a sequence that is alignable to a splice site. In some embodiments, the k-mer sequence of the plurality of k-mer sequences comprises a sequence that is alignable to a gene fusion junction. In some embodiments, (d) further comprises identifying a cancer cell type of the cancer. In some embodiments, (d) further comprises identifying a cancer tissue of origin of the cancer. In some embodiments, (d) further comprises identifying a cancer subtype of the cancer. In some embodiments, (d) further comprises identifying a plurality of cancer types of the cancer. In some embodiments, the subject is suspected of having cancer. In some embodiments, the subject was previously diagnosed with cancer. In some embodiments, the subject is undergoing treatment for cancer. In some embodiments, the method further comprises treating the subject for the detected presence of the cancer. In some embodiments, the treating comprises administering a surgery, a chemotherapy, a radiation therapy, an immunotherapy, or a combination thereof.

[0010] In another aspect, the present disclosure provides a computer-implemented method of generating a trained machine learning classifier for cancer prediction, the method comprising: (a) providing (i) a first plurality of samples derived from non-cancer subjects, and (ii) a second plurality of samples derived from subjects with cancer; (b) assaying cell-free RNA derived from the first and second plurality of samples to generate a plurality of cancer-derivedsequencing reads and a plurality of non-cancer-derived sequencing reads; (c) processing the plurality of cancer-derived sequencing reads and the plurality of non-cancer-derived sequencing reads to generate a plurality of cancer-associated k-mer sequences, wherein the processing is performed without mapping or aligning the sequencing reads to a reference genome sequence; (d) training a machine learning classifier using the plurality of cancer-associated k-mer sequences, thereby generating the trained machine learning classifier, wherein the trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject.

[0011] In some embodiments, (c) further comprises determining a number of the plurality of cancer-derived sequencing reads or the plurality of non-cancer-derived sequencing reads that comprise a given cancer-associated k-mer sequence. In some embodiments, (c) further comprises comparing a number of the plurality of cancer-derived sequencing reads that comprise the given cancer-associated k-mer sequence to a number of the plurality of non-cancer- derived sequence reads that comprise the given cancer-associated k-mer sequence. In some embodiments, (d) further comprises determining that a cancer-associated k-mer sequence of the plurality of cancer-associated k-mer sequences being present in a larger number of the plurality of cancer-derived sequencing reads as compared to the plurality of non-cancer-derived sequencing reads as indicative of the presence of cancer. In some embodiments, (c) further comprises generating an output of a number of the plurality of cancer-derived sequencing reads or the plurality of non-cancer-derived sequencing reads that comprise a given cancer-associated k-mer sequence for each of the first and second plurality of samples. In some embodiments, the trained machine learning classifier is further programmed to provide a prediction of a cancer cell type of the cancer. In some embodiments, the trained machine learning classifier is further programmed to provide a prediction of a cancer tissue of origin of the cancer. In some embodiments, the trained machine learning classifier is further programmed to provide a prediction of a subtype of the cancer. In some embodiments, the cell-free RNA comprises a small RNA.

[0012] In some embodiments, the trained machine learning classifier predicts cancer at an accuracy of at least about 90%. In some embodiments, the trained machine learning classifier predicts cancer at an area under curve (AUC) of at least about 0.90.

[0013] In another aspect, the present disclosure provides non-transitory computer- readable storage medium comprising a set of instructions recorded thereon, which, when executed by a processor, cause the processor to implement a method for detecting a presence oran absence of cancer in a subject, the method comprising: (i) obtaining a plurality of sequencing reads derived from cell -free RNA of the subject; (ii) processing the sequencing reads to generate a plurality of k-mer sequences, wherein the processing is performed without mapping or aligning the sequencing reads to a reference genome sequence; (iii) analyzing the plurality of k- mer sequences using a trained machine learning classifier, wherein the trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject, based at least in part on presence or absence of cancer-associated k-mer sequences in the plurality of k-mer sequences; and (iv) detecting the presence or absence of cancer in the subject, based at least in part on the output generated in (iii).

[0014] In another aspect, the present disclosure provides a computer readable medium or mediums for generating a trained machine learning classifier for cancer prediction, wherein the computer readable medium or mediums comprising a set of instructions recorded thereon, wherein, when the set of instructions are executed by a processor, the following steps are implemented: (i) obtaining plurality of cancer-derived sequencing reads and a plurality of non-cancer-derived sequencing reads; (ii) processing the plurality of cancer-derived sequencing reads and the plurality of non-cancer-derived sequencing reads to generate a plurality of cancer- associated k-mer sequences, wherein the processing is performed without mapping or aligning the sequencing reads to a reference genome sequence; (iii) training a machine learning classifier using the plurality of cancer-associated k-mer sequences, thereby generating the trained machine learning classifier, wherein the trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject.

[0015] In another aspect, the present disclosure provides a system for detecting a presence or an absence of cancer in a subject, the system comprising: at least one system memory comprising (a) a plurality of sequencing reads derived from cell-free RNA of the subject, and (b) a trained machine learning classifier; and a computer memory coupled to the at least one system memory and programed to (i) process the sequencing reads to generate a plurality of k-mer sequences, wherein the processing is performed without mapping or aligning the sequencing reads to a reference genome sequence; (ii) analyze the plurality of k-mer sequences using the trained machine learning classifier, wherein the trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject, based at least in part on presence or absence of cancer-associated k-mer sequences in the plurality of k-mer sequences; and (iii) detect the presence or absence of cancer in the subject, based at least in part on the output generated in (ii).

[0016] In another aspect, the present disclosure provides a system for generating a trained machine learning classifier for cancer prediction, the system comprising: at least one system memory comprising a plurality of cancer-derived sequencing reads and a plurality of non-cancer-derived sequencing reads; and a computer memory coupled to the at least one system memory and programed to (i) process the plurality of cancer-derived sequencing reads and the plurality of non-cancer-derived sequencing reads to generate a plurality of cancer- associated k-mer sequences, wherein the processing is performed without mapping or aligning the sequencing reads to a reference genome sequence; and (ii) train a machine learning classifier using the plurality of cancer-associated k-mer sequences, thereby generating the trained machine learning classifier, wherein the trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject.

[0017] Another aspect of the present disclosure provides a non-transitory computer readable medium comprising machine executable code that, upon execution by one or more computer processors, implements any of the methods above or elsewhere herein.

[0018] Another aspect of the present disclosure provides a system comprising one or more computer processors and computer memory coupled thereto. The computer memory comprises machine executable code that, upon execution by the one or more computer processors, implements any of the methods above or elsewhere herein.

[0019] Additional aspects and advantages of the present disclosure will become readily apparent to those skilled in this art from the following detailed description, wherein only illustrative embodiments of the present disclosure are shown and described. As will be realized, the present disclosure is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the disclosure. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.INCORPORATION BY REFERENCE

[0020] 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

[0021] The novel features of the invention are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the invention are utilized, and the accompanying drawings (also “figure” and “FIG.” herein), of which:

[0022] FIG. 1 shows a receiver operating characteristic (ROC) curve showing an area under the curve (AUC) of 0.99 with true positive rate of greater than 0.95 at the specificity of 0.95 for predicting lung cancer status using the systems and methods described herein.

[0023] FIG. 2 shows a computer system that is programmed or otherwise configured to implement methods provided herein.DETAILED DESCRIPTION

[0024] While various embodiments of the invention 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 may occur to those skilled in the art without departing from the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed.

[0025] The widespread reprogramming of the gene expression landscape may be a hallmark of cancer development. Thus, the systematic identification of regulatory pathways that drive pathologic gene expression patterns may be a crucial step toward understanding and treating cancer. A multitude of regulatory mechanisms may be implicated in oncogenic expression of genes involved in cancer cell differentiation, survival, invasion, and spread. While transcriptional pathways that underlie oncogenesis may be studied, post-transcriptional regulatory pathways have also emerged as major regulators of this process. For example, microRNAs (small non-coding RNAs), a subclass of small RNAs that function in gene silencing, were among the first characterized post-transcriptional regulators of breast cancer progression. RNA-binding proteins (RBPs) are also critical posttranscriptional regulators of gene expression, and RBPs may be shown to affect oncogenesis and cancer progression. Further, other classes of small non-coding RNAs, such as tRNAs and tRNA fragments, may play a fundamental role in breast cancer progression.

[0026] Despite the diverse repertoire of regulatory mechanisms involved in cancers, a shared characteristic among them is that they co-opt and dysregulate existing pathways withinthe cell. In other words, cancer cells adopt myriad strategies, such as somatic mutations (e.g. KRAS), gene fusions (e.g. BCR-ABL), epigenetic modifications (e.g. promoter hypermethylation), and regulatory mechanisms disruptions (NFkB transcription factors) to overactivate oncogenic and to down-regulate tumor suppressive pathways. While these strategies rely on the pathologic modulation of regulatory programs that are already in place, there is an often-overlooked possibility that cancer cells may be capable of evolving or engineering specialized regulatory pathways that drive tumorigenesis.

[0027] The present disclosure provides systems and methods for analyzing small RNA sequences that serve as biomarkers which are indicative of cancer, tissue of origin, and cancer status of a subject, which may be used to accurately diagnose cancer even at an early stage. In some embodiments, the methods comprise use of cell-free small RNAs in a suitable sample. In some embodiments, the sample is a biological sample. In some embodiments, the biological sample is a whole blood, serum, or plasma sample. Small RNAs may serve as biomarkers that are indicative of any of various pathologies, including particular cancers, and which may be used to accurately diagnose a particular pathology or, when the pathology of interest is cancer, to grade cancer in a subject.

[0028] Current methods for performing liquid biopsy for early, non-invasive cancer detection from biofluids may rely on extracting somatic and / or germline genetic variants or methylation information from circulating tumor DNA (ctDNA), exosomes, proteins, or miRNAs. These approaches may face numerous challenges, such as a need to have very high coverage of serum / blood sequencing to capture these features; reliance on mapping of the reads to the reference genome, which in many cancers may have been significantly changed; and the fact that the features and biomarkers used may be highly impacted and limited by biases and limitations of read mapping algorithms.

[0029] Recognizing the challenges above, the present disclosure provides systems and methods for using sub-sequences of small RNA-seq (smRNA-seq) reads from biofluids (e.g. serum or plasma) as potential biomarkers for predicting cancer. Such biomarkers may be referred to herein as oncKmers. The systems and methods of the present disclosure may not only provide a large number of novel biomarkers that were mostly ignored before, but also completely bypasses the mapping or alignment operations used in other liquid biopsy analysis techniques.

[0030] The systems and methods of the present disclosure may confer significant advantages over other methods, such as: obviating the need for a reference genome, which maybe significantly altered in cancer cells (e.g., aneuploidy and structural variations); obviating accurate mapping of smRNA-seq reads, which may be challenging and uncertain for short reads; the ability to discover and quantify new, cancer-specific exon-exon splice junctions and gene fusions; and the ability to assemble oncKmers into longer contigs, providing insight into cellular state of the tumor.

[0031] The systems and methods of the present disclosure may provide two key advantages over other approaches. Firstly, systems and methods of the present disclosure may consider a much larger number of potential features (oncKmers) as biomarkers for cancer. By identifying these novel sets of biomarkers that are differentially expressed in cancer, the number of features indicating cancer may be significantly increased. Secondly, cancer cells can produce multiple copies of the dysregulated small RNA, which can be secreted into the bloodstream. By leveraging these two properties, systems and methods of the present disclosure can detect signals from low-coverage sequencing of serum or plasma samples, thereby alleviating the above-mentioned challenges arising from the need to have very high coverage of sequencing to sufficiently capture useful features.

[0032] Furthermore, by considering the k-mers with altered expression in cancer (oncKmers) as the features, the need for mapping or alignment of the small reads to a reference genome may be obviated. This may not only eliminate the biases that exist due to mapping of reads to the reference genome, but also provides the ability to use signals that are non-linear in the genome space (e.g., splice junctions). This reduces the above-mentioned challenges arising from the need to map reads to a reference genome, and features and biomarkers being highly impacted and limited by biases and limitations of read mapping algorithms.

[0033] Therefore, systems and methods of the present disclosure may provide the ability to bypass mapping of short-reads to the genome for predicting cancer status, and instead directly use the sequence content (oncKmer distribution as features). This may reduce any noise that can be introduced due to ambiguities of mapping of the reads (e.g., to multiple locations in the genome). Further, systems and methods of the present disclosure may enable the consideration of potential alterations to small RNA transcripts that may not be easy to capture by mapping based approaches, for example RNA alterations.

[0034] In some embodiments, systems and methods of the present disclosure may use oncKmers counts in small RNA-seq for liquid biopsy samples for early prediction of single cancer. In some embodiments, systems and methods of the present disclosure may use oncKmers counts in small RNA-seq for liquid biopsy samples for multi -cancer early prediction.In some embodiments, systems and methods of the present disclosure may use oncKmers counts in small RNA-seq for liquid biopsy samples for cancer tissue of origin (TOO) prediction. In some embodiments, systems and methods of the present disclosure may use oncKmers counts in small RNA-seq for liquid biopsy samples for cancer subtypes prediction (e.g. CMS status in CRC cancer). In some embodiments, systems and methods of the present disclosure may use oncKmers for monitoring disease progression and return of the disease.

[0035] A k-mer may refer to a nucleotide sequence of size k. Further, an oncKmer may refer to a k-mer that is significantly enriched in sequencing reads of small RNA sequencing (smRNA-seq) reads of cancer-derived versus non-cancer-derived samples. This approach to find these oncKmers and utilize them for prediction may comprise the following operations. First, a dataset of cancer-derived and non-cancer-derived samples is used to identify k-mers in smRNA- seq data that are enriched in cancer-derived vs non-cancer-derived samples using a statistical method. These cancer-derived or non-cancer derived samples may be a tumor or tissue sample, or may be a biofluid sample, such as a plasma or serum sample derived from a healthy individual or an individual diagnosed with a cancer. Second, a statistical approach is used to select the more relevant set of k-mers denoted as oncKmers to be used as biomarkers. Third, these oncKmers are “counted” in a smRNA-seq data from serum or plasma (or another biofluid), e.g., how many sequencing reads contain each specific oncKmer for each sample. The output of this operation may be expressed as a matrix of oncKmer count per sample. Fourth, the counts are used as features in a machine learning setup for building a cancer status prediction model.

[0036] Methods and systems of the present disclosure may employ, unless otherwise indicated, techniques of molecular biology (including recombinant techniques), microbiology, cell biology, and biochemistry. Such techniques may be described by, for example, “Molecular Cloning: A Laboratory Manual”, 2ndedition (Sambrook et al., 1989); “Oligonucleotide Synthesis” (M.J. Gait, ed., 1984); “Animal Cell Culture” (R.I. Freshney, ed., 1987); “Methods in Enzymology” (Academic Press, Inc.); “Handbook of Experimental Immunology”, 4thedition (D.M. Weir & C.C. Blackwell, eds., Blackwell Science Inc., 1987); “Gene Transfer Vectors for Mammalian Cells” (J.M. Miller & M.P. Calos, eds., 1987); “Current Protocols in Molecular Biology” (F.M. Ausubel et al., eds., 1987); and “PCR: The Polymerase Chain Reaction”, (Mullis et al., eds., 1994), each of which is incorporated by reference in its entirety.

[0037] As used in the present disclosure and claims, the singular forms “a”, “an” and“the” include plural forms unless the context clearly dictates otherwise.

[0038] Whenever the term “at least,” “greater than,” or “greater than or equal to” precedes the first numerical value in a series of two or more numerical values, the term “at least” or “greater than” applies to each one of the numerical values in that series of numerical values.

[0039] Whenever the term “no more than,” “less than,” or “less than or equal to” precedes the first numerical value in a series of two or more numerical values, the term “no more than” or “less than” applies to each one of the numerical values in that series of numerical values.

[0040] It is understood that wherever embodiments are described herein with the language “comprising” otherwise analogous embodiments described in terms of “consisting of’ and / or “consisting essentially of’ are also provided. It is also understood that wherever embodiments are described herein with the language “consisting essentially of’ otherwise analogous embodiments described in terms of “consisting of’ are also provided.

[0041] The term “and / or” as used in a phrase such as “A and / or B” herein is intended to include both A and B; A or B; A (alone); and B (alone). Likewise, the term “and / or” as used in a phrase such as “A, B, and / or C” is intended to encompass each of the following embodiments: A, B, and C; A, B, or C; A or C; A or B; B or C; A and C; A and B; B and C; A (alone); B (alone); and C (alone).

[0042] The term “about” or “approximately” as used herein is meant to refer to within 5%, within 4%, within 3%, within 2%, within 1%, of a given value or range.

[0043] The term “one k-mer,” as used herein, generally refers to a k-mer of size k that is enriched (e.g., by a statistically significant difference) in small RNA sequencing (smRNA-seq) reads of cancer-derived samples as compared to non-cancer-derived samples. The length of k-mers may vary, for example, about 5, 10, 11, 12, 13, 14, 15, 20, 25, 30, 35, 40, 45, 50, 100, or 200 nucleotides in length. The k-mers may have a length of 10 nucleotides to 50 nucleotides. The k-mer can be a ribonucleotide sequence and can be determined from nucleotide sequence information of ribonucleic acid (RNA) molecules. The nucleotides in the k-mer can be exact nucleotide values or approximate nucleotide values. Disclosed herein, in some embodiments, are k-mer based statistical analyses showing presence / absence or increase / decrease in k-mers between two sample types. A statistically significant difference, as used herein, generally refers to a p value of at about 0.001, 0.005, 0.01, 0.05, or 0.1. A p-value may refer to adjusted p-values, such as multiple test corrected p-values. For example, astatistically significant difference may be determined using a multiple testing corrected p-value at a false discovery rate no higher than 10%.

[0044] The term “biomarker”, as used herein, generally refers to a biological molecule present in an individual at varying concentrations useful in predicting the cancer status of an individual. A biomarker may include but is not limited to, nucleic acids, proteins and variants and fragments thereof. A biomarker may be DNA comprising the entire or partial nucleic acid sequence encoding the biomarker, or the complement of such a sequence. Biomarker nucleic acids useful in the systems and methods of the present disclosure are considered to include both DNA and RNA comprising the entire or partial sequence of any of the nucleic acid sequences of interest.

[0045] The term “bodily fluid”, as used herein, generally refers to a biological sample in the form of a bodily fluid comprising RNA (RNA) including blood (or a fraction of blood such as plasma or serum), lymph, mucus, tears, saliva, sputum, urine, semen, stool, CSF (cerebrospinal fluid), breast milk, and, ascites fluid. In some embodiments, the bodily fluid is urine. In some embodiments, the bodily fluid is fractionated serum comprising extracellular vesicles such as exosomes.

[0046] The terms “cancer” and “cancerous”, as used herein, generally refer to or describe the physiological condition in mammals in which a population of cells are characterized by unregulated cell growth. In some embodiments, the cancer is a breast cancer. In some embodiments, the cancer is a liver cancer. In some embodiments, the cancer is a gastric cancer. In some embodiments, the cancer is a lung cancer. In some embodiments, the cancer is a colorectal cancer. In some embodiments, the cancer is a kidney cancer.

[0047] The term "correlate" or "correlating", as used herein, generally refers to a statistical association between instances of two events, where events may include numbers, data sets, and the like. For example, when the events involve numbers, a positive correlation (also referred to herein as a "direct correlation") means that as one increases, the other increases as well. A negative correlation (also referred to herein as an "inverse correlation") means that as one increases, the other decreases. The present disclosure provides small RNAs, the levels of which are correlated with a particular outcome measure, such as between the level of a small RNA and the likelihood of developing breast cancer. For example, the increased level of a small RNA may be negatively correlated with a likelihood of good clinical outcome for the patient. In this case, for example, the patient may have a decreased likelihood of long-term survival without recurrence of the cancer and / or a positive response to a chemotherapy, and the like. Such anegative correlation indicates that the patient likely has a poor prognosis or may respond poorly to a chemotherapy, and this may be demonstrated statistically in various ways, e.g., by a high hazard ratio.

[0048] The terms “identical” or “percent identity” or “homology” in the context of two or more nucleic acids, as used herein, generally refer to two or more sequences or subsequences that are the same or have a specified percentage of nucleotides or amino acid residues that are the same, when compared and aligned (introducing gaps, if necessary) for maximum correspondence, not considering any conservative amino acid substitutions as part of the sequence identity. The percent identity may be measured using sequence comparison software or algorithms or by visual inspection. Various algorithms and software that may be used to obtain alignments of amino acid or nucleotide sequences. These include, but are not limited to, BLAST, ALIGN, Megalign, BestFit, GCG Wisconsin Package, and variations thereof. In some embodiments, two nucleic acids of the present disclosure are substantially identical, meaning they have at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, and in some embodiments at least about 95%, 96%, 97%, 98%, 99% nucleotide or amino acid residue sequence identity, when compared and aligned for maximum correspondence, as measured using a sequence comparison algorithm or by visual inspection. In some embodiments, identity exists over a region of the sequences that is at least about 10, at least about 20, at least about 40-60 nucleotides, at least about 60-80 nucleotides or any integral value therebetween. In some embodiments, identity exists over a longer region than 60-80 nucleotides, such as at least about 80-100 nucleotides, and in some embodiments the sequences are substantially identical over the full length of the sequences being compared.

[0049] The term “level”, as used herein, generally refers to qualitative or quantitative determination of the number of copies of a non-coding RNA transcript. An RNA transcript exhibits an “increased level” when the level of the RNA transcript is higher in a first sample, such as in a clinically relevant subpopulation of patients (e.g., patients who have cancer), than in a second sample, such as in a related subpopulation (e.g., patients who do not have cancer). In the context of an analysis of a level of an RNA transcript in a tumor sample obtained from an individual patient, an RNA transcript exhibits “increased level” when the level of the RNA transcript in the subject trends toward, or more closely approximates, the level characteristic of a clinically relevant subpopulation of patients. The methods herein are useful in determining levels of certain small ncRNA sequences as well as the presence or absence of selected small ncRNA biomarkers and the absolute number of distinct small ncRNA sequences.

[0050] The term “metastasis”, as used herein, generally refers to the process by which a cancer spreads or transfers from the site of origin to other regions of the body with the development of a similar cancerous lesion at a new location. A “metastatic” or“ metastasizing” cell is one that loses adhesive contacts with neighboring cells and migrates (e.g., via the bloodstream or lymph) from the primary site of disease to secondary sites.

[0051] A “patient response” , as used herein, generally may be assessed using any endpoint indicating a benefit to the patient, including, without limitation, (1) inhibition, to some extent, of tumor growth, including slowing down and complete growth arrest; (2) reduction in the number of tumor cells; (3) reduction in tumor size; (4) inhibition (i.e., reduction, slowing down or complete stopping) of tumor cell infiltration into adjacent peripheral organs and / or tissues; (5) inhibition (i.e. reduction, slowing down or complete stopping) of metastasis; (6) enhancement of anti-tumor immune response, which may, but does not have to, result in the regression or rejection of the tumor; (7) relief, to some extent, of one or more symptoms associated with the cancer; (8) increase in the length of survival following treatment; and / or (9) decreased mortality at a given point of time following treatment.

[0052] The terms “polynucleotide” and “nucleic acid” and “nucleic acid molecule”, as used herein, generally refer to polymers of nucleotides of any length, and include DNA and RNA. The polynucleotides can be deoxyribonucleotides, ribonucleotides, modified nucleotides or bases, and / or their analogs, or any substrate that can be incorporated into a polymer by DNA or RNA polymerase.

[0053] The terms “polypeptide” and “peptide” and “protein”, as used herein, generally refer to polymers of amino acids of any length. The polymer may be linear or branched, it may comprise modified amino acids, and it may be interrupted by non-amino acids. The terms also encompass an amino acid polymer that has been modified naturally or by intervention; for example, disulfide bond formation, glycosylation, lipidation, acetylation, phosphorylation, or any other manipulation or modification, such as conjugation with a labeling component. Also included within the definition are, for example, polypeptides containing one or more analogs of an amino acid (including, for example, unnatural amino acids), as well as other modifications. It is understood that, because the polypeptides of the present disclosure may be based upon antibodies or fusion proteins, in certain embodiments, the polypeptides can occur as single chains or associated chains (e.g., dimers).

[0054] The term "prognosis", as used herein, generally refers to the prediction of the likelihood of cancer-attributable death or progression, including recurrence, metastatic spread, and drug resistance, of neoplastic disease, such as breast cancer.

[0055] The term “reference” RNA transcript, as used herein, generally refers to an RNA transcript whose level can be used to compare the level of an RNA transcript in a test sample. In an embodiment, reference RNA transcripts include housekeeping genes, such as betaglobin, alcohol dehydrogenase, or any other RNA transcript, the level or expression of which does not vary depending on the disease status of the cell containing the RNA transcript. In another embodiment, all of the assayed RNA transcripts, or a subset thereof, may serve as reference RNA transcripts.

[0056] The term “small RNA” (smRNA), as used herein, generally refers to RNA that has a length of less than 200 nucleotides and includes transfer RNA (tRNA), ribosomal RNA (rRNA), snoRNAs, microRNA (miRNA), siRNAs, small nuclear (snRNA), Y RNA, vault RNA, antisense RNA, tiRNA (transcription initiation RNA), TSSa-RNA (transcriptional startsite associated RNA) and piwiRNA (piRNA). It is to be understood that a "non-coding RNA" may overlap, in part, with a segment of RNA that does code for a protein, although the entirety of the non-coding RNA sequence per se does not. In some embodiments, a small smRNA as used herein is between 50 and 100 nucleotides. A smRNA may be of endogenous origin (e.g., a human small non-coding RNA) or exogenous origin (e.g., virus, bacteria, parasite). In some embodiments, any of the methods disclosed herein comprise detecting any one or combination of RNAs disclosed above.

[0057] The term “subject”, as used herein, generally refers to any animal (e.g., a mammal), including, but not limited to, humans, non-human primates, canines, felines, rodents, and the like. In some embodiments, the subject is a human subject. The terms "subject," "individual," and "patient" are used interchangeably herein. The terms "subject," "individual," and "patient" thus encompass individuals having cancer (e.g., breast cancer), including those who have undergone or are candidates for resection (surgery) to remove cancerous tissue.

[0058] The term “therapeutically effective amount”, as used herein, generally refers to a quantity sufficient to achieve a desired therapeutic effect, for example, an amount which results in the prevention or amelioration of or a decrease in the symptoms associated with a disease that is being treated, e.g., disorders associated with cancer growth or a hyperproliferative disorder. The amount of compound administered to the subject will depend on the type and severity of the disease and on the characteristics of the individual, such as general health, age,sex, body weight and tolerance to drugs. It will also depend on the degree, severity and type of disease. The skilled artisan will be able to determine appropriate dosages depending on these and other factors. The regimen of administration can affect what constitutes an effective amount. Further, several divided dosages, as well as staggered dosages, can be administered daily or sequentially, or the dose can be continuously infused, or can be a bolus injection. Further, the dosages of the compound(s) of the present disclosure can be proportionally increased or decreased as indicated by the exigencies of the therapeutic or prophylactic situation. An effective amount of the compounds of the present disclosure, sufficient for achieving a therapeutic effect, may range from about 0.000001 mg per kilogram body weight per day to about 10,000 mg per kilogram body weight per day. In some embodiments, the dosage ranges are from about 0.0001 mg per kilogram body weight per day to about 100 mg per kilogram body weight per day. The compounds disclosed herein can also be administered in combination with each other, or with one or more additional therapeutic compounds.

[0059] The term “sample”, as used herein, generally refers to a biological sample obtained or derived from a source of interest, as described herein. In some embodiments, a source of interest comprises an organism, such as an animal or human. In some embodiments, a biological sample comprises biological tissue or fluid. In some embodiments, a biological sample may be or comprise bone marrow; whole blood; plasma; serum; blood cells; platelets; ascites; tissue or fine needle biopsy samples; cell-containing body fluids; free floating nucleic acids; sputum; saliva; urine; cerebrospinal fluid, peritoneal fluid; pleural fluid; pancreatic cyst fluid; feces; lymph; gynecological fluids; skin swabs; vaginal swabs; oral swabs; nasal swabs; washings or lavages such as a ductal lavages or bronchoalveolar lavages; aspirates; scrapings; bone marrow specimens; tissue biopsy specimens; surgical specimens; feces, other body fluids, secretions, and / or excretions; and / or cells therefrom, etc. In some embodiments, a biological sample is or comprises cells obtained from an individual. In some embodiments, a sample is a “primary sample” obtained directly from a source of interest by any appropriate method. For example, in some embodiments, a primary biological sample is obtained by methods selected from the group consisting of biopsy (e.g., fine needle aspiration or tissue biopsy), surgery, collection of body fluid (e.g., blood, lymph, feces etc.), etc. In some embodiments, as will be clear from context, the term “sample” refers to a preparation that is obtained by processing (e.g., by removing one or more components of and / or by adding one or more agents to) a primary sample. For example, filtering using a semi-permeable membrane. Such a “processed sample” may comprise, for example nucleic acids or proteins extracted from a sample or obtained bysubjecting a primary sample to techniques such as amplification or reverse transcription of RNA, isolation and / or purification of certain components, etc.

[0060] The terms “treating” or “treatment” or “ treat”, as used herein, generally refer to both 1) therapeutic measures that cure, slow down, lessen symptoms of, and / or halt progression of a diagnosed pathologic condition or disorder and 2) prophylactic or preventative measures that prevent or slow the development of a targeted pathologic condition or disorder. Thus those in need of treatment include those already diagnosed with the disorder; those prone to have the disorder; and those in whom the disorder is to be prevented. In some embodiments, a subject is successfully “treated” according to the methods of the present disclosure if the patient shows one or more of the following: a reduction in the number of and / or complete absence of cancer cells; a reduction in the tumor size; an inhibition of tumor growth; inhibition of and / or an absence of cancer cell infiltration into peripheral organs including the spread of cancer cells into soft tissue and bone; inhibition of and / or an absence of tumor or cancer cell metastasis; inhibition and / or an absence of cancer growth; relief of one or more symptoms associated with the specific cancer; reduced morbidity and mortality; improvement in quality of life; reduction in tumorigenicity; reduction in the number or frequency of cancer stem cells; or some combination of such effects.

[0061] The term "tumor", as used herein, generally refers to all neoplastic cell growth and proliferation, whether malignant or benign, and all pre-cancerous and cancerous cells and tissues.

[0062] The term "tumor sample", as used herein, generally refers to a sample comprising tumor material obtained from a cancer patient. The term encompasses tumor tissue samples, for example, tissue obtained by surgical resection and tissue obtained by biopsy, such as for example, a core biopsy or a fine needle biopsy. In a particular embodiment, the tumor sample is a fixed, wax-embedded tissue sample, such as a formalin-fixed, paraffin-embedded tissue sample. Additionally, the term "tumor sample" encompasses a sample comprising tumor cells obtained from sites other than the primary tumor, e.g., circulating tumor cells. The term also encompasses cells that are the progeny of the patient's tumor cells, e.g. cell culture samples derived from primary tumor cells or circulating tumor cells. The term further encompasses samples that may comprise protein or nucleic acid material shed from tumor cells in vivo, e.g., bone marrow, blood, plasma, serum, and the like. The term also encompasses samples that have been enriched for tumor cells or otherwise manipulated after their procurement and samplescomprising polynucleotides and / or polypeptides that are obtained from a patient' s tumor material.RNA Biomarkers of Cancer

[0063] The human genome encodes for a vast amount of coding and non-coding RNA in including small RNA (smRNA) transcripts. Multiple smRNA classes may be described including the highly abundant transfer RNAs (tRNAs), ribosomal RNAs (rRNAs), small nucleolar RNAs (snoRNAs), microRNAs (miRNAs), small interfering RNAs (siRNAs), small nuclear RNAs (snRNAs), and piwi-interacting RNAs (piRNAs) (Amaral et al., 2008; Martens- Uzunova et al., 2013). For example, small non-coding RNAs can act as translational repressors by binding to target mRNAs at sites with adequate sequence complementary (Ameres et al., 2007), while the highly abundant cytoplasmic Y RNAs function in RNA quality control by affecting the subcellular location of Ro proteins (Sim et al., 2009). The repressive activity of mature small non-coding RNAs on mRNA translation is shared by other classes of ncRNAs, including siRNAs and endo-siRNAs, in addition to piRNAs that silence retrotransposons at defined subcellular locations (Chuma and Pillai, 2009). Small non-coding RNA activity relies on sufficient levels of abundance in the cytoplasm, and interaction with RNA-induced silencing complexes (RISC) localized at endosomal membranes (Gibbings et al., 2009; Lee et al., 2009a), whereas low abundant small non-coding RNAs have less impact on translational repression. As a consequence, subtle alterations in the levels of certain small non-coding RNA may already influence cellular processes, while strong perturbations can cause disease. Besides abundance, interactions with (RISC) proteins but also RNA partners and correct subcellular localization are interrelated factors that control small non-coding RNA physiology (Mullokandov et al., 2012; Wee et al., 2012). RNA may be cell-free RNA. Various RNAs may be released from cells and may be present in the fluids of a subject. For example, an RNA may be shed from tumor cells and enter the blood stream of a subject.

[0064] Small RNAs can be secreted in cell-derived extracellular vesicles such as exosomes. Both mRNA and small RNA species may be found in extracellular vesicles. As such, extracellular vesicles can provide a vehicle for transfer and protection of RNA content from degradation in the extracellular environment and generate cell-free RNA, enabling a stable source for reliable detection of RNA biomarkers.

[0065] The present disclosure relates to RNA biomarkers, including small RNA biomarkers, found to be differentially present in biological samples derived from subjects having cancer, as compared with subjects who are "normal," i.e., subjects who do not have cancer. Asmall RNA biomarker or set of small RNA biomarkers is differentially present between samples if the difference between the levels of expression of the small RNA biomarker or set of small RNA biomarkers in the samples is determined to be statistically significant. Common tests for statistical significance include, but are not limited to, t-test, ANOVA, Kniskal-Wallis, Wilcoxon, Mann-Whitney, and odds ratio, small RNA biomarkers, alone or in combination, can be used to provide a measure of the relative risk that a subject has or does not have cancer.

[0066] Small RNA biomarkers of cancer may be discovered by small RNA sequencing of multiple cancer types and subtypes from various tissues of origin, and identifying previously unknown small RNAs that are specifically expressed in cancer cells. For example, over 260,000 such RNAs may be identified across various cell / tissue and corresponding cancer types. These small RNA biomarkers can now be used to determine the type of cancer and cancer status thereof of a subject, for example, a subject whose cancer status was previously unknown or who is suspected to be suffering from cancer. This may be accomplished by determining the presence or level of one or more biomarkers of RNAs in a biological sample derived from the subject. A difference in the level of one or more of these RNA biomarkers as compared to that in a biological sample derived from a normal subject is an indication that the subject has cancer of the type and tissue of origin associated with the RNA biomarkers.

[0067] Small RNA biomarkers may be identified via processing of sequencing reads of RNA samples. The processing of the sequence reads may be performed without the use of alignment or mapping of sequencing reads to a reference genome sequence. Methods that do not use mapping or alignment may have advantages over methods that do use mapping or alignment. For example, some sequencing reads or portions of sequencing reads may be difficult to map or align. Sequences comprising gene fusion junction or splice junction may be difficult to align or map. Additionally, reference genomes may be different than genomes present in cancer cells, for example, based on the presence of mutations, and may be difficult to align with sequencing reads from tumor samples. Mapping or alignment of sequences may result in a loss of sequence information in downstream processes. For example, methods that use mapping or alignment may ignore or remove sequences that do not align or map to a reference, thereby resulting in a loss of data. Small RNA sequencing reads may be difficult to align based on the length of the molecule and sequencing reads. As such, methods using mapping or aligning may be more difficult to perform on sequencing data derived from small RNA.

[0068] The processing of sequencing reads may comprise generating a plurality of k- mer sequences derived from the sequencing reads. K-mer sequences of the sequence reads maycomprise subsequences of the sequencing reads of the length k. A set of k-mer sequences can comprises all subsequence of the sequencing reads that comprises a length k. A set of k-mer sequences can comprises a plurality of subsequences of the sequencing reads that comprises a length k. k-mer sequences can be at least 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, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53,54, 55, 56, 57, 58, 59, 60, 61, 62, 63,64, 65, 66, 67, 68, 69, 70, 71, 72, 73,74, 75, 76, 77, 78, 79, 80, 81, 82, 83,84, 85, 86, 87, 88, 89, 90, 91, 92, 93,94, 95, 96, 97, 98, 99, 100, or more nucleotides long, k-mer sequences can be about 2, 3, 4, 5, 6, 7, 8, 9, 10 ,11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53,54, 55, 56, 57, 58, 59, 60, 61,62, 63,64, 65, 66, 67, 68, 69, 70, 71, 72, 73,74, 75, 76, 77, 78, 79, 80, 81, 82, 83,84, 85, 86, 87,88, 89, 90, 91, 92, 93,94, 95, 96, 97, 98, 99, 100 or more nucleotides, k-mer sequences can be no more than about 2, 3, 4, 5, 6, 7, 8, 9, 10 ,11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,52, 53,54, 55, 56, 57, 58, 59, 60, 61, 62, 63,64, 65, 66, 67, 68, 69, 70, 71, 72, 73,74, 75, 76, 77,78, 79, 80, 81, 82, 83,84, 85, 86, 87, 88, 89, 90, 91, 92, 93,94, 95, 96, 97, 98, 99, 100 , or less nucleotides. For example, k-mer sequences may be about 10 nucleotides to 50 nucleotides long. Generating k-mer sequences may comprise generating all possible k-mer sequence of length k for a sequencing read. Generating k-mer sequences may comprise generating a portion of all possible k-mer sequence of length k for a sequencing read. Generating k-mer sequences may comprise generating a portion of all possible k-mer sequence of length k for a sample.

[0069] The k-mer sequences may analyzed using a trained classifier (e.g., machine learning classifier) to classify a sample. The trained classifier may be trained on a k-mer sequences that are associated with a cancer sample. The trained classifier may be trained on a k- mer sequences that are associated with a non-cancer sample. The trained classifier may be trained using machine learning to identify k-mer sequences associated with cancer and / or k-mer sequence associated with non-cancer. Based on the presence of cancer-associated k-mer sequence in a sample, the trained classifier may classify a sample. A small RNA biomarker may be a k-mer.

[0070] A subject having a difference in the level of one or more RNA biomarkers (e.g., one or more k-mers) as compared to a normal subject may have cancer, including early- stage, moderate or mid-stage, or severe or late-stage cancer. In one embodiment, the level of one or more RNA biomarkers (e.g., one or more k-mers) may be used to diagnose cancer, in a subject having symptoms characteristic of early-stage cancer.

[0071] In one embodiment, the level, presence or absence, and / or number of RNA biomarkers (e.g., one or more k-mers) correlated with cancer may be used to monitor the course of cancer progression in a subject. The cancer status of a subject can change over time. For example, the cancer may worsen or improve over time. With such worsening or improvement, the level, presence or absence, and / or number of small RNA biomarkers may change in a statistically significant fashion, as detected in samples derived from the subject. For example, the level, presence or absence, and / or number of small RNA biomarkers may increase over time with the development of the cancer, and the ability to detect a particular small RNA biomarker may increase with time as well. Thus, the course of the cancer progression, in a subject can be monitored by determining the level, presence or absence, and / or number of small RNA biomarkers in a first sample derived from a subject, and determining the level, presence or absence, and / or number of one or more small RNA biomarkers in a second sample derived from a subject, where the second sample is obtained after the first sample. The levels, presence or absence, and / or number of small RNA biomarkers in the second sample relative to those in the first sample are indicative of disease progression.

[0072] Whether or not the level, presence or absence, or number of small RNA biomarkers in a biological sample derived from a test subject is different from the level, presence or absence, or number of the small RNA biomarkers present in a normal subject may be ascertained by comparing the level, presence or absence, or number of the small RNA biomarkers in the sample from the test subject with a control. For example, a control may be a biological sample derived from a known subject, e.g., a subject known to be a normal subject that does not have cancer. If a control is obtained from a normal subject, a statistically significant difference in the level of a small RNA biomarker in a test subject relative to the suitable control is indicative that the subject has cancer. In one embodiment, the difference in the level of a small RNA biomarker is an increase. A control may also be a reference standard. A reference standard may serve as a reference level for comparison, such that test samples can be compared to the reference standard in order to infer the cancer status of a subject. A reference standard may be representative of the level of one or more small RNA biomarkers in a known subject, e.g., a subject known to be a normal subject, or a subject known to a particular cancer or cancer status. Likewise, a reference standard may be representative of the level of one or more small RNA biomarkers in a population of known subjects, e.g., a population of subjects known to be normal subjects, or a population of subjects known to have the cancer or a particular cancer status. The reference standard may be obtained, for example, by pooling samples from a plurality of individuals and determining the level of a small RNA biomarker in the pooledsamples, to thereby produce a standard over an averaged population. Such a reference standard represents an average level of a small RNA biomarker among a population of individuals. A reference standard may also be obtained, for example, by averaging the level of a small RNA biomarker determined to be present in individual samples obtained from a plurality of individuals. Such a standard is also representative of an average level of a small RNA biomarker among a population of individuals. A reference standard may also be a collection of values each representing the level of a small RNA biomarker in a known subject in a population of individuals. In certain embodiments, test samples may be compared against such a collection of values in order to infer the cancer status of a subject. In certain embodiments, the reference standard is an absolute value. In such embodiments, test samples may be compared against the absolute value in order to infer the cancer, tissue or origin, or cancer status of a subject. In a one embodiment, a comparison between the level of one or more small RNA biomarkers in a sample relative to a control is made by executing a software classification algorithm. For example, the comparing of a level, presence or absence, or number of the small RNA biomarkers may be performed via analysis with a trained classifier. The trained classifier may be trained via an input of control samples that have a level, presence or absence, or number of the small RNA biomarkers. By analyzing a level, presence or absence, or number of the small RNA biomarkers of the test subject sample, the trained classifier can generate a classification for the sample.Biological Samples

[0073] The level or presence of one or more small RNA biomarkers may be determined in a biological sample derived from a subject. A sample derived from a subject is one that originates from a subject. Such a sample may be further processed after it is obtained from the subject. For example, RNA may be isolated from a sample. In this example, the RNA isolated from the sample is also a sample derived from a subject. A biological sample useful for determining the level of one or more small RNA biomarkers may be obtained from essentially any source, including cells, tissues, and fluids throughout the body.

[0074] In some embodiments, the biological sample used for determining the level of one or more small RNA biomarkers is a sample containing cell-free RNA. Cell-free RNAs can freely circulate in a wide range of biological material, including bodily fluids, such as fluids from the circulatory system, e.g., a blood sample or a lymph sample, or from another bodily fluid such as urine or saliva. Accordingly, in some embodiments, the biological sample used for determining the level of one or more small RNA biomarkers is a bodily fluid, for example, blood, fractions thereof, serum, plasma, urine, saliva, tears, sweat, semen, vaginal secretions, lymph, bronchial secretions, CSF, whole blood, stool, interstitial fluid, synovial fluid, gastricacid, sebum, mucus, bile, etc. In some embodiments, the sample is a sample that is obtained non-invasively, such as a stool sample. In some embodiments, the sample is a serum sample from a human. In some embodiments, the sample is a plasma sample from a human.

[0075] The substance may be solid, for example, a biological tissue. The substance may comprise normal healthy tissues. The tissues may be associated with various types of organs. Non-limiting examples of organs may include brain, breast, liver, lung, kidney, prostate, ovary, spleen, lymph node (including tonsil), thyroid, pancreas, heart, skeletal muscle, intestine, larynx, esophagus, stomach, or combinations thereof.

[0076] The substance may comprise a tumor. Tumors may be benign (non-cancer), pre-malignant, or malignant (cancer), or any metastases thereof.

[0077] The substances may comprise a mix of normal healthy tissues or tumor tissues. The tissues may be associated with various types of organs. Non-limiting examples of organs may include brain, breast, liver, lung, kidney, prostate, ovary, spleen, lymph node (including tonsil), thyroid, pancreas, heart, skeletal muscle, intestine, larynx, esophagus, stomach, or combinations thereof.

[0078] In some embodiments, the substance may comprise a variety of cells, including: eukaryotic cells, prokaryotic cells, fungi cells, heart cells, lung cells, kidney cells, liver cells, pancreas cells, reproductive cells, stem cells, induced pluripotent stem cells, gastrointestinal cells, blood cells, cancer cells, bacterial cells, bacterial cells isolated from a human microbiome sample, and circulating cells in the human blood. In some embodiments, the substance may comprise contents of a cell, such as, for example, the contents of a single cell or the contents of multiple cells.

[0079] In some embodiments, any of the methods disclosed herein comprise using a small volume sample. In some embodiments, the methods disclosed comprise isolating total RNA or small RNA, e.g., small RNA, and / or amplifying total or small RNA in a sample of no more than about 20 microliters of sample, 40 microliters of sample, 80 microliters of sample, 100 microliters of sample, 200 microliters of sample, 300 microliters of sample, 400 microliters of sample, 500 microliters of sample, 600 microliters of sample, 700 microliters of sample, 800 microliters of sample, 900 microliters of sample, 1 milliliter of sample, 1.1 milliliters of sample, 1.2 milliliters of sample, 1.3 milliliters of sample, 1.4 milliliters of sample, 1.5 milliliters of sample, 1.6 milliliters of sample, 1.7 milliliters of sample, 1.8 milliliters of sample, 1.9 milliliters of sample, 2.0 milliliters of sample. In some embodiments, the sample size is fromabout 25 microliters to about 2 milliliters of liquid sample in the form of subject plasma, whole blood or serum.

[0080] In some embodiments, the methods disclosed comprise isolating total RNA and / or amplifying non-coding RNA in a sample of no more than about 20 microliters of serum, 40 microliters of serum, 80 microliters of serum, 100 microliters of serum, 200 microliters of serum, 300 microliters of serum, 400 microliters of serum, 500 microliters of serum, 600 microliters of serum, 700 microliters of serum, 800 microliters of serum, 900 microliters of serum, 1 milliliter of serum, 1.1 milliliters of serum, 1.2 milliliters of serum, 1.3 milliliters of serum, 1.4 milliliters of serum, 1.5 milliliters of serum, 1.6 milliliters of serum, 1.7 milliliters of serum, 1.8 milliliters of serum, 1.9 milliliters of serum, 2.0 milliliters of serum.

[0081] Circulating small RNAs include small RNAs in cells, extracellular small RNAs in microvesicles, in exosomes and extracellular small RNAs that are not associated with cells or microvesicles (extracellular, non-vesicular small RNA). In some embodiments, the biological sample used for determining the level of one or more small RNA biomarkers (e.g., a sample containing circulating small RNA) may contain cells. In other embodiments, the biological sample may be free or substantially free of cells (e.g., a serum sample). In some embodiments, a sample containing circulating small RNAs, e.g., extracellular small RNAs, is a blood-derived sample. Blood-derived sample types may include, e.g., a plasma sample, a serum sample, a blood sample, etc. In other embodiments, a sample containing circulating small RNAs is a lymph sample. Circulating small RNAs are also found in urine and saliva, and biological samples derived from these sources are likewise suitable for determining the level of one or more small RNA biomarkers.

[0082] In some embodiments, any of the methods of the disclosure comprise the operation of isolating total RNA or small RNA from a sample or cell or extracellular vesicle. Methods of isolating RNA for expression analysis from blood, plasma and / or serum (see for example, Tsui NB et al. (2002) Clin. Chem. 48,1647-53, incorporated by reference in its entirety herein) and from urine (see for example, Boom R et al. (1990) J Clin Microbiol. 28, 495-503, incorporated by reference in its entirety herein) have been described.Processing of Biological Samples

[0083] Biological samples may be subjected to one or more processing operations as part of methods and systems as described herein. The one or more processing operations may be to isolate a substance from the biological sample, to purify the biological sample, to separate thebiological sample into one or more fractions for further use or processing, to quantitate the amount of a substance, to the detect the presence or absence or one or more substances, to transform or modify a substance for further downstream processing or analysis, or any combination thereof. For example, enzymes may be added to digest protein and remove contamination, or to inactivate nucleases that might otherwise degrade nucleic acids (RNA or DNA) during purification. The one or more processing operations may immediately follow sample collection, may be immediately prior to another processing or assaying operation, or may be carried out contemporaneously with another processing or assaying operation. The processing operation(s) may also be carried out on a sample or an intermediate in the process that has been appropriately stored for a designated amount of time. Any number of suitable processing operations may be performed on the biological sample or part thereof.

[0084] Processing operations as described herein may include, but are not limited to, immunoassays, enzyme-linked immunosorbent assays (ELISA), radioimmunoassays (RIA), ligand binding assays, functional assays, enzymatic assays, enzymatic treatments (e.g., with kinases, phosphatases, ligases, transcriptases, reverse transcriptases), enzymatic digestions (e.g., nucleases), spectroscopic assays (e.g., UV-vis spectroscopy, Fourier transform infrared spectroscopy, circular dichroism spectroscopy) spectrophotometric assays (e.g., ultraviolet- visible light spectrophotometry), immunoprecipitations (IP), sequencing reactions, electrophoresis, chromatography, enrichments, pull-downs, and mass spectrometry (MS). In some embodiments, a method as described herein comprises not performing one or more processing operations.

[0085] In some embodiments, a processing operation as described herein may be performed on a sample containing one or more nucleic acid molecules. In some embodiments, the nucleic acid molecules comprise ribonucleic acid (RNA) molecules or derivatives thereof. In some embodiments, the RNA molecules comprise small RNAs (smRNAs). In some embodiments, the RNA molecules comprise small RNAs (sncRNAs).

[0086] One or more processing operations may be performed on a sample or portion thereof. The methods and systems disclosed herein may comprise performing 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more processing operations on a sample or portion thereof. The one or more processing operations may be performed sequentially or simultaneously. The one or more processing may be performed on the same sample or portions (e.g., aliquots, fractions) thereof or they may be performed on different samples.

[0087] In some embodiments, biological samples containing or suspected of containing RNAs are subjected to one or more processing operations to facilitate downstream processing operations (e.g., isolation). In some embodiments, a sample containing or suspected of containing RNAs may be treated with one or more enzymes, cofactors, and / or other reagents to bring about an end-repair process. The sample containing or suspected of containing one or more RNAs is subjected to treatment with a polynucleotide kinase (PNK). PNK enzymes may phosphorylate or dephosphorylate nucleic acid molecules in a sample to permit the nucleic acid molecules to participate in subsequent reactions, such as ligation reactions. For example, an RNA which may or may not be dephosphorylated by treatment with PNK may be subsequently polyadenylated and then incorporated into downstream processing steps.

[0088] In some embodiments, biological samples containing or suspected of containing one or more RNAs are subjected to one or more processing operations to remove or change RNA modifications that may inhibit further downstream processing (e.g., subsequent reverse transcription). Alternatively or additionally, chemical modifications may be removed or retained to determine the presence or absence of a relationship between chemical modifications and a disease state or any pathological state, e.g., a cancer state of a subject or population of subjects. Chemically modified RNA bases may include, without limitation, N6-methyladenosine (m6A), inosine (I), 5-methylcytosine (m5C), pseudouridine ( ), 5-hydroxymethylcytosine (hm5C), Nkmethyladenosine (mxA), or 7-methylguanosine (m7G). For example, biological samples which contain or are suspected of contained methylated RNAs (e.g, comprising m6A, m5C, hm5C, mxA, or m7G) may be treated with one or more demethylation enzymes (e.g, AlkB) to remove alkyl groups which may interfere with downstream processing (e.g., by reverse transcriptases).

[0089] In some embodiments, the biological sample is subjected to one or more immunoprecipitation (IP) reactions, such as an in vitro or in vivo crosslinking and IP reaction (CLIP). The one or more IP reactions may enrich or pull-down one or more substances of interest. In some embodiments, an IP processing operation includes a cross-linking operation to covalently link two or more different substances (e.g., to link protein and DNA or protein and RNA). Immunoprecipitation of the cross-linked substances may provide an indication of biological substances that are associated with one another and / or may be used to enrich for specific substances that are known or suspected to interact with one another. In an example, one or more RNAs of interest are cross-linked to one or more corresponding proteins. IP of the proteins cross-linked with the RNAs allows for subsequent isolation and downstream processing of the RNAs of interest. In another example, an IP reaction may be carried out using antibodiesspecific for an RNA modification of interest, e.g., an adenosine modification (such as m6A, mxA, alternative polyadenylation, or adenosine-to-inosine RNA editing), a uridine modification (such as conversion to pseudouridine), or other RNA modifications as alluded to above.

[0090] In some embodiments, the sample is subjected to one or more isolation operations. Isolation operations may target a general class of molecules (e.g., nucleic acids, such as RNAs) or a specific molecule (e.g., a specific annotated RNA molecule).

[0091] In some embodiments, RNA is converted to DNA (cDNA) prior to analysis. cDNA can be generated by reverse transcription of isolated RNA using conventional techniques. In some embodiments, the small RNA is amplified prior to additional assays. Many amplification-based methods exist for amplification of small RNA nucleic acid sequences, including, but not limited to, PCR, RT-PCR, qPCR, and rolling circle amplification. Other amplification-based techniques include, for example, ligase chain reaction, multiplex ligatable probe amplification, in vitro transcription (IVT), strand displacement amplification, transcription-mediated amplification, RNA (Eberwine) amplification, and other methods.

[0092] A PCR reaction may include multiple operations, or cycles, that selectively amplify target nucleic acid species: a denaturing operation, in which a target nucleic acid is denatured; an annealing operation, in which a set of PCR primers (i.e., forward and reverse primers) anneal to complementary DNA strands, and an elongation operation, in which a thermostable DNA polymerase elongates the primers. By repeating these operations multiple times, a DNA fragment is amplified to produce an amplicon, corresponding to the target sequence. PCR reactions may include five or more cycles of denaturation, annealing, and elongation, including at least 10 such cycles, at least 20 such cycles, and the like. In many cases, the annealing and elongation operations can be performed concurrently, in which case the cycle contains only two operations. A reverse transcription reaction (which produces a cDNA sequence having complementarity to an RNA) may be performed prior to PCR amplification. Reverse transcription reactions include the use of, e.g., a RNA-based DNA polymerase (reverse transcriptase) and a primer.

[0093] In some embodiments, two or more small RNAs are amplified in a single reaction volume. For example, multiplex PCR can enable simultaneous amplification of at least two small RNAs of interest in one reaction volume by using more than one pair of primers. The primer pairs can comprise at least one amplification primer that specifically binds each small RNA.

[0094] In some embodiments, a processing operation may comprise adding (e.g., spiking in) one or more substances. The one or more spike-in substances may be for any suitable purpose, including but not limited to, quality control, enrichment of target species, the act of depleting non-target species, or any combination thereof. In some embodiments, a spike-in substance may comprise a synthetic biomolecule (e.g., nucleic acid, such as an RNA or a modified RNA, i.e., an RNA containing base modifications as alluded to above) corresponding to a target biomolecule. In some embodiments, the spike-in substance may comprise an endogenous or exogenous biomolecule. The spike-in molecule may be selected based on any appropriate property such as abundance or relative abundance, origin, sequence or part thereof, global or local structure (such as secondary or tertiary structure), or any combination thereof. Quantification of a spike-in substance following one or more downstream processing operations may be used for quality control as well as quantity control.

[0095] In some embodiments, a processing operation may comprises associating a target biomolecule or set of target biomolecules with one or more unique molecular identifiers (UMIs). UMIs may be used to associate biomolecules and indicate them as being derived from the same sample or part thereof. In an example, UMIs (e.g., nucleic acid barcodes) are assigned to or associated with individual samples or parts thereof. Alternatively or additionally, the UMIs are assigned to or associated with an individual subject. UMIs associated with individual biological samples or subjects may allow for pooling of samples during downstream processing.

[0096] Aside from associating biomolecules with particular samples or individuals, UMIs may also allow for downstream process control and absolute quantitation or biomolecules (e.g., nucleic acids, such as small RNAs). In such cases, a single UMI may correspond to one or substantially one target biomolecule. In an example, target biomolecules derived from a biological sample are tagged with individual UMIs corresponding to individual biomolecules. The UMIs allow quantitation of the corresponding target biomolecules and enable controlling for sequencing artifacts (e.g., PCR duplication).

[0097] In some embodiments, the UMIs comprise nucleic acid barcodes which are associated with nucleic acids derived from a biological sample as described herein. The nucleic acid barcodes are attached or otherwise associated with the sample-derived nucleic acids to give a set of tagged nucleic acid constructs. In some embodiments, the target nucleic acids are associated with nucleic acid barcodes corresponding to an individual sample. In some embodiments, the target nucleic acids are associated with nucleic acid barcodes corresponding to individual molecules.

[0098] One subset of processing operations may serve a different function than another subset of processing operations. For example, one processing operation may be performed to determine the presence of a substance in a sample and a second assay may be performed to isolate the substance from the sample. Processing operations may be performed in any appropriate order. For example, a sample may first be processed to determine the presence of a target substance. A second processing operation may then be used to isolate the target substance from the sample. Optionally, a third processing operation may be performed to purify the isolated substance. In another example, a sample is first processed to isolate a substance or plurality of substances. A second processing operation is then performed on the isolate to determine the presence or absence of a target substance or substances in the isolate. Optionally, a third processing operation is performed between the first and second processing operations to purify the one or more target substances. In another example, a processing operation is performed to enrich for the presence or one or more RNAs in a sample, for instance by enriching DNA made from the RNA, wherein the enrichment may be carried out using a solid support (e.g., streptavidin beads) that specifically binds to the DNAs of interest (which may, for instance, be modified so as to have a covalently bound biotin group or bound to a complementary nucleic acid sequence that is biotinylated). Before or after the enrichment operation, another processing operation is optionally performed to determine the presence of the one or more RNAs. Still other combinations and sequences of processing operations on a sample are contemplated herein.

[0099] In some embodiments, a sample is processed or analyzed by a third party. For example, one party may conduct a purification or enrichment processing operation on a sample. The purified or enriched sample may then be subjected to a subsequent processing operation (e.g., a quantitation) by a different party.Determining the Level, Presence or Absence, or Number of Small RNA Biomarkers in a Sample

[0100] Sequencing methods can be used to generate sequencing reads derived small RNA. For example, small RNAs can be detected using Next Generation Sequencing (NGS), e.g., Sequencing-By-Synthesis using, for example, the NovaSeq, Next Seq, HiSeq, HiScan, GenomeAnalyzer, or MiSeq systems (Illumina, Inc., San Diego CA) or the Ultima Genomics Sequencer (Ultima Genomics, Newark CA). Small RNAs can also be detected using Ion Torrent Sequencing (Ion Torrent Systems, Inc., Gulliford, Conn.) or other suitable methods of semiconductor sequencing.

[0101] Determining a level of small RNA biomarkers may comprising counting a number of sequencing reads comprising a small RNA biomarker. The number of sequencing reads that comprise a given biomarker may correspond to an amount of a small RNA biomarker in a sample.Quality Control Methods and Systems

[0102] Further disclosed herein are methods and systems for quality control of a determination of biomarker (e.g., small RNA expression) levels in a sample. Quality control may be carried out on biological samples to correct for biases, normalize a sample, ensure the presence of target or absence of non-target species, ensure biological samples are free of contamination, or any other appropriate purpose. In an example, biomarkers of interest may be in or suspected of being in a plasma or serum sample from a subject. Biomarkers which may be present in white blood cells (WBC) or red blood cells that are exposed to the plasma / serum upon hemolysis may confound a determination of the levels of the same or one or more other biomarkers in the sample. Thus, a quality control inspection may ensure that a plasma / sample is not contaminated as a result of hemolysis or by WBCs.

[0103] Classification and machine learning models as described elsewhere herein may be used for quality control of quantification of biomarker levels in the sample. For example, based on the comparison of the quantified level of one or more target biomarkers with one or more control biomarkers, a sample may be determined as either passing or failing a quality control inspection.

[0104] Control biomarkers may comprise any suitable control. For example, a control biomarker may be an exogenous or endogenous biomarker (e.g., small RNA) which is optionally spiked-in to the sample or portion or derivative thereof. The one or more control biomarkers may be subjected to the same quantitation procedure (e.g., qPCR, microarray hybridization, sequencing) as the target biomarker(s). Based on a comparison of the measured amounts of the one or more targets and the one or more control biomarkers, a particular sample or groups of samples may either be determined as passing or failing a quality control inspection or operation.

[0105] In an example, a set of processed biological samples containing or suspected of containing one or more small RNAs of interest are amplified (e.g., by qPCR), optionally in parallel. The biological samples optionally have an exogenous RNA spiked-in as a control. A classification model as described elsewhere herein which takes in data corresponding to theamplification, such as cycle threshold (Ct) data, or other data derived from the amplified sample, such as electrophoretic data obtained and optionally processed by, e.g., a TapeStation system (Agilent Technologies, Inc.). The classification model may make a determination of which biological sample or samples to accept or reject for further processing based on the input data. In addition to evaluating individual samples using the foregoing methodology, batch averages are determined and the acceptance or rejection of samples can be performed on the batch level in addition to the sample level.

[0106] A quality control operation or inspection may be performed before or after any suitable operation of the methods described herein. For example, quality control may be performed before isolation of one or more nucleic acids (e.g., small RNAs), after isolation of one or more nucleic acids, before determining the level of one or more biomarkers, or after determining the level or one or more biomarkers in a sample. Alternatively, a quality control operation or process may be carried out simultaneously with an isolation or quantitation operation.

[0107] In some embodiments, a quality control process may increase sequencing efficiency or depth. A quality control operation may be used to deplete highly present species without cancer-specific information. In some embodiments, Depletion of Abundant Sequences by Hybridization (DASH), or hybridization methods generally, may be used to target regions or biomarkers that do not contain or are presumed not to contain cancer-specific information (Gu et al, Genome Biology, 17:41 (2016)). In some embodiments, a species that contains a high amount of cancer-specific information or is suspect of containing a high amount of cancerspecific information may be enriched.Generation of Classification Algorithms for Qualifying Cancer Status

[0108] In some embodiments, data that are generated using samples such as “known samples” can then be used to “train” a classification model. A “known sample” is a sample that has been pre-classified, e.g., classified as being derived from a normal subject, or from a subject having a certain cancer or subtype of cancer. The data that are derived from the spectra and are used to form the classification model can be referred to as a “training data set.” Once trained, the classification model can recognize patterns in data derived from spectra generated using unknown samples. The classification model can then be used to classify the unknown samples into classes. This can be useful, for example, in predicting whether or not a particular biological sample is associated with a certain biological condition (e.g., diseased versus non-diseased).

[0109] In some embodiments, data for the training data set that is used to form the classification model can via sequencing reads derived from samples. The sample may be derived from subjects with cancer and non-cancer subjects. The sequencing reads may be sequencing reads generated by sequencing small RNA (e.g., cell-free small RNA). A plurality of k-mer sequences may be derived from the sequencing reads. K-mer sequences of the sequence reads may comprise subsequences of the sequencing reads of the length k. A set of k-mer sequences can comprises all subsequences of the sequencing reads that comprises a length k. A set of k- mer sequences can comprises a plurality of subsequences of the sequencing reads that comprises a length k. k-mer sequences can be at least 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, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53,54, 55, 56, 57, 58, 59, 60, 61, 62, 63,64, 65, 66, 67, 68, 69, 70, 71, 72, 73,74, 75, 76, 77, 78, 79, 80, 81, 82, 83,84, 85, 86, 87, 88, 89, 90, 91, 92, 93,94, 95, 96, 97, 98, 99, 100, or more nucleotides long, k-mer sequences can be about 2, 3, 4, 5, 6, 7, 8, 9, 10 ,11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53,54, 55, 56, 57, 58, 59, 60, 61,62, 63,64, 65, 66, 67, 68, 69, 70, 71, 72, 73,74, 75, 76, 77, 78, 79, 80, 81, 82, 83,84, 85, 86, 87,88, 89, 90, 91, 92, 93,94, 95, 96, 97, 98, 99, 100 or more nucleotides, k-mer sequences can be no more than about 2, 3, 4, 5, 6, 7, 8, 9, 10 ,11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,52, 53,54, 55, 56, 57, 58, 59, 60, 61, 62, 63,64, 65, 66, 67, 68, 69, 70, 71, 72, 73,74, 75, 76, 77,78, 79, 80, 81, 82, 83,84, 85, 86, 87, 88, 89, 90, 91, 92, 93,94, 95, 96, 97, 98, 99, 100 , or less nucleotides. For example, k-mer sequences may be about 10 nucleotides to 50 nucleotides long. Generating k-mer sequences may comprise generating all possible k-mer sequence of length k for a sequencing read. Generating k-mer sequences may comprise generating a portion of all possible k-mer sequence of length k for a sequencing read. Generating k-mer sequences may comprise generating a portion of all possible k-mer sequence of length k for a sample.

[0110] The presence of specific k-mer sequences in a sequencing read (e.g., a small RNA sequencing read) of cancer-associated samples can be identified. Similarly, the presence of specific k-mer sequences in sequencing read (e.g., a small RNA sequencing read) of non-cancer samples can be identified. Enrichment of k-mer sequences (e.g., an increase in the number of occurrence of a k-mer) in cancer associated samples can be compared to non-cancers samples can be used to identify k-mer sequences that are associated with cancer. These k-mer sequences can be used to train the classifier to classify a sample. Similarly, k-mer sequences enriched in certain cancer types, subtypes, tissues of origin, can be used to train a classifier.

[0111] In some embodiments, data for the training data set that is used to form the classification model may comprise other health data from a subject or set of subjects. For example, the data may comprise laboratory data at one or more time points (e.g., level of a metabolite or other biomarker in a sample from the subject), qualitative measures of cell pathology (e.g., as determined by histopathological analysis), structured or unstructured health data (e.g., biographic data), medical imaging data (e.g., results of an x-ray, computed tomography (CT) scan, magnetic resonance imaging (MRI) scan, positron-emission tomography (PET) scan, or ultrasound), previous or current cancer status (e.g., remission, metastasis), current or previous therapeutic interventions, or long-term outcome. The health data may alternatively or additionally comprise the contents of electronic health records (EHRs) corresponding to a subject or set of subjects. Such EHRs may comprise an electronic version of a subject’s or set of subjects’ medical history(ies) over time, including demographics, clinician notes, medications, vital signs, medical history, immunizations, laboratory data, and medical imaging data.

[0112] Data may be taken at one or more timepoints. In some embodiments, data are taken at an initial timepoint and a later timepoint. The initial timepoint and the later timepoint may be spaced by any appropriate amount of time, such as 1 hour, 1 day, 1 week, 2 weeks, 3 weeks, 4 weeks, 6 weeks, 12 weeks, 4 months, 5 months, 6 months, 7 months, 8 months, 9 months, 10 months, 11 months, 1 years, 2 years, 3 years, 4 years, 5 years, 6 years, 7 years, 8 years, 9 years, 10 years, or more. In some embodiments, the data are from more than two timepoints. In some embodiments, the data are from 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more timepoints.

[0113] Classification models can be formed using any suitable statistical classification (or “learning”) method that attempts to segregate bodies of data into classes based on objective parameters present in the data. Classification methods may be either supervised or unsupervised. Examples of supervised and unsupervised classification processes are described in Jain, "Statistical Pattern Recognition: A Review", IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, No. 1, January 2000, the contents of which are incorporated by reference herein.

[0114] In supervised classification, training data containing examples of known categories are presented to a learning mechanism, which learns one or more sets of relationships that define each of the known classes. New data may then be applied to the learning mechanism, which then classifies the new data using the learned relationships. Examples of supervised classification processes include linear regression processes (e.g., multiple linear regression(MLR), partial least squares (PLS) regression and principal components regression (PCR)), binary decision trees (e.g., recursive partitioning processes such as CART— classification and regression trees), artificial neural networks such as back propagation networks, discriminant analyses (e.g., Bayesian classifier or Fischer analysis), logistic classifiers, and support vector classifiers (support vector machines).

[0115] In other embodiments, the classification models that are created can be formed using unsupervised learning methods. Unsupervised classification attempts to learn classifications based on similarities in the training data set, without pre-classifying the spectra from which the training data set was derived. Unsupervised learning methods include cluster analyses. A cluster analysis attempts to divide the data into ’’clusters” or groups that ideally should have members that are very similar to each other, and very dissimilar to members of other clusters. Similarity is then measured using some distance metric, which measures the distance between data items, and clusters together data items that are closer to each other. Clustering techniques include the MacQueen's K-means algorithm and the Kohonen's SelfOrganizing Map algorithm. Learning algorithms asserted for use in classifying biological information are described, for example, in PCT International Publication No. WO 01 / 31580 (Barnhill et al., "Methods and devices for identifying patterns in biological systems and methods of use thereof'), U.S. patent application No. 2002 0193950 Al (Gavin et al, "Method or analyzing mass spectra"), U.S. patent application No. 2003 0004402 Al (Hitt et al., "Process for discriminating between biological states based on hidden patterns from biological data"), and U.S. patent application No. 2003 0055615 Al (Zhang and Zhang, “Systems and methods for processing biological expression data”). The contents of the foregoing patent applications are incorporated herein by reference in their entirety.

[0116] Systems and methods as described herein may use more than one classification model to determine an output (e.g., cancer status of a subject). Systems and methods may comprise 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more classification models. A classification model of the plurality of classification models may be trained on a particular type of data (e.g., small RNA expression data or other health data). Alternatively, a classification model may be trained on more than one type of data. The inputs of one classification model may comprise the outputs of one or more other classification models. Additionally, a classification model may receive as its input the output of one or more classification models. The classification models can be formed on and used on any suitable digital computer. Suitable digital computers include micro, mini, or large computers using anystandard or specialized operating system, such as a Unix, WINDOWS or LINUX based operating system.

[0117] The output of classification models as described herein may comprise one or more output values which comprises one of a fixed number of possible values (e.g., a linear classifier, a logistic regression classifier, etc.) indicating a classification of the biological sample and / or the subject by the classifier. The classification model may comprise a binary classifier, such that each of the one or more output values comprises one of two values (e.g., {0, 1 }, {positive, negative}, or {high-risk, low-risk}) indicating a classification of the biological sample and / or subject by the classifier. The classification model may be another type of classifier, such that each of the one or more output values comprises one of more than two values (e.g., {0, 1, 2}, {positive, negative, or indeterminate}, or {high-risk, intermediate-risk, or low-risk}) indicating a classification of the biological sample and / or subject by the classifier. The output values may comprise descriptive labels, numerical values, or a combination thereof. Some of the output values may comprise descriptive labels. Such descriptive labels may provide an identification or indication of the disease or disorder state of the subject, and may comprise, for example, positive, negative, high-risk, intermediate-risk, low-risk, or indeterminate. Such descriptive labels may provide an identification of a treatment for the subject’s cancer state or status, and may comprise, for example, a therapeutic intervention, a duration of the therapeutic intervention, and / or a dosage of the therapeutic intervention suitable to treat a subject classified in a particular cancer-related category.

[0118] Some of the output values may comprise numerical values, such as binary, integer, or continuous values. Such binary output values may comprise, for example, {0, 1 }, {positive, negative}, or {high-risk, low-risk}. Such integer output values may comprise, for example, {0, 1, 2}. Such continuous output values may comprise, for example, a probability value of at least 0 and no more than 1. Such continuous output values may comprise, for example, an un-normalized probability value of at least 0. Such continuous output values may indicate a prognosis of the cancer-related category of the subject. Some numerical values may be mapped to descriptive labels, for example, by mapping 1 to “positive” and 0 to “negative.”

[0119] Some of the output values may be assigned based on one or more cutoff values. For example, a binary classification of samples may assign an output value of “positive” or 1 if the sample indicates that the subject has at least a 50% probability of having a cancer status (e.g., type, subtype, or stage of cancer). For example, a binary classification of samples may assign an output value of “negative” or 0 if the sample indicates that the subject has lessthan a 50% probability of having a cancer status. In this case, a single cutoff value of 50% is used to classify samples into one of the two possible binary output values. Examples of single cutoff values may include about 1%, about 2%, about 5%, about 10%, about 15%, about 20%, about 25%, about 30%, about 35%, about 40%, about 45%, about 50%, about 55%, about 60%, about 65%, about 70%, about 75%, about 80%, about 85%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98%, and about 99%.

[0120] As another example, a classification of samples may assign an output value of “positive” or 1 if the sample indicates that the subject belongs to a cancer status (e.g., cancer diagnosis or prognosis) of at least about 50%, at least about 55%, at least about 60%, at least about 65%, at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, at least about 91%, at least about 92%, at least about 93%, at least about 94%, at least about 95%, at least about 96%, at least about 97%, at least about 98%, at least about 99%, or more. The classification of samples may assign an output value of “positive” or 1 if the sample indicates that the subject has a cancer status (e.g., predicted long-term outcome) of more than about 50%, more than about 55%, more than about 60%, more than about 65%, more than about 70%, more than about 75%, more than about 80%, more than about 85%, more than about 90%, more than about 91%, more than about 92%, more than about 93%, more than about 94%, more than about 95%, more than about 96%, more than about 97%, more than about 98%, or more than about 99%.

[0121] The classification of samples may assign an output value of “negative” or 0 if the sample indicates that the subject has a probability of having a cancer status (e.g., positive for breast cancer) of less than about 50%, less than about 45%, less than about 40%, less than about 35%, less than about 30%, less than about 25%, less than about 20%, less than about 15%, less than about 10%, less than about 9%, less than about 8%, less than about 7%, less than about 6%, less than about 5%, less than about 4%, less than about 3%, less than about 2%, or less than about 1%. The classification of samples may assign an output value of “negative” or 0 if the sample indicates that the subject has a probability of having a cancer status (e.g., for breast cancer) of no more than about 50%, no more than about 45%, no more than about 40%, no more than about 35%, no more than about 30%, no more than about 25%, no more than about 20%, no more than about 15%, no more than about 10%, no more than about 9%, no more than about 8%, no more than about 7%, no more than about 6%, no more than about 5%, no more than about 4%, no more than about 3%, no more than about 2%, or no more than about 1%.

[0122] The classification of samples may assign an output value of “indeterminate” or 2 if the sample is not classified as “positive”, “negative”, 1, or 0. In this case, a set of two cutoff values is used to classify samples into one of the three possible output values. Examples of sets of cutoff values may include { 1%, 99%}, {2%, 98%}, {5%, 95%}, { 10%, 90%}, { 15%, 85%}, {20%, 80%}, {25%, 75%}, {30%, 70%}, {35%, 65%}, {40%, 60%}, and {45%, 55%}. Similarly, sets of n cutoff values may be used to classify samples into one of n+ possible output values, where n is any positive integer.

[0123] The training data set(s) and the classification models can be embodied by computer code that is executed or used by a digital computer. The computer code can be stored on any suitable computer readable media including optical or magnetic disks, sticks, tapes, etc., and can be written in any suitable computer programming language including C, C++, visual basic, etc.

[0124] The learning algorithms described above can be used for developing classification algorithms for small RNA biomarkers for cancer status. The classification algorithms can, in turn, be used in diagnostic tests by providing diagnostic values (e.g., cut-off points) for biomarkers used singly or in combination.

[0125] The small RNA biomarkers described herein can be used to assess the cancer state or status of a subject. Cancer status includes the presence or absence of a particular cancer (e.g., breast, lung, liver, kidney, gastric, or colorectal cancer), a particular stage of cancer, a tissue of origin of a cancer, a particular subtype of cancer, a predicted response to a therapeutic intervention for a cancer, a subject having a risk level (e.g., high risk, low risk) for a particular cancer or subtype of cancer, and a predicted long-term outcome of a cancer (e.g., distant metastasis, biochemical recurrence, partial response, complete response, overall survival, cancer-specific survival, progression free survival, five year survival, disease free survival, death), or any combination thereof. Cancer status may also include monitoring the course of a cancer, for example, monitoring disease progression, subtype (including subtype selection, switching or conversion) or determining the level of residual tumor burden. Based on the cancer status of a subject, additional procedures may be indicated, including, for example, additional diagnostic tests or therapeutic procedures.

[0126] Non-limiting examples of tumors and associated cancers that may be detected, diagnosed, prognosed, and the like by methods and systems as described herein include acoustic neuroma, acute lymphoblastic leukemia, acute myeloid leukemia, adenocarcinoma, adrenocortical carcinoma, AIDS-related cancers, AIDS-related lymphoma,anal cancer, angiosarcoma, appendix cancer, astrocytoma, basal cell carcinoma, bile duct cancer, bladder cancer, bone cancers, brain tumors, such as cerebellar astrocytoma, cerebral astrocytoma / malignant glioma, ependymoma, medulloblastoma, supratentorial primitive neuroectodermal tumors, visual pathway and hypothalamic glioma, breast cancer, bronchial adenomas, Burkitt lymphoma, carcinoma of unknown primary origin, central nervous system lymphoma, bronchogenic carcinoma, cerebellar astrocytoma, cervical cancer, childhood cancers, chondrosarcoma, chordoma, choriocarcinoma, chronic lymphocytic leukemia, chronic myelogenous leukemia, chronic myeloproliferative disorders, colon cancer, colon carcinoma, craniopharyngioma, cutaneous T-cell lymphoma, cystadenocarcinoma, desmoplastic small round cell tumor, embryonal carcinoma, endocrine system carcinomas, endometrial cancer, endotheliosarcoma, ependymoma, epithelial carcinoma, esophageal cancer, Ewing’s sarcoma, fibrosarcoma, germ cell tumors, gallbladder cancer, gastric cancer, gastrointestinal carcinoid tumor, gastrointestinal stromal tumor, gastrointestinal system carcinomas, genitourinary system carcinomas, gliomas, hairy cell leukemia, head and neck cancer, heart cancer, hemangioblastoma, hepatocellular (liver) cancer, Hodgkin lymphoma, Hypopharyngeal cancer, intraocular melanoma, islet cell carcinoma, Kaposi sarcoma, kidney cancer, laryngeal cancer, leiomyosarcoma, lip and oral cavity cancer, liposarcoma, liver cancer, lung cancers, such as non-small cell and small cell lung cancer, lung carcinoma, lymphangiosarcoma, lymphangioendotheliosarcoma, lymphomas, leukemias, macroglobulinemia, malignant fibrous histiocytoma of bone / osteosarcoma, medulloblastoma, medullary carcinoma, melanomas, meningioma, mesothelioma, metastatic squamous neck cancer with occult primary, mouth cancer, multiple endocrine neoplasia syndrome, myelodysplastic syndromes, myeloid leukemia, myxosarcoma, nasal cavity and paranasal sinus cancer, nasopharyngeal carcinoma, neuroblastoma, non-Hodgkin lymphoma, non-small cell lung cancer, oligodendroma, oral cancer, oropharyngeal cancer, osteosarcoma / malignant fibrous histiocytoma of bone, ovarian cancer, ovarian epithelial cancer, ovarian germ cell tumor, pancreatic cancer, pancreatic cancer islet cell, papillary adenocarcinoma, papillary carcinoma, paranasal sinus and nasal cavity cancer, parathyroid cancer, penile cancer, pharyngeal cancer, pheochromocytoma, pineal astrocytoma, pineal germinoma, pituitary adenoma, pleuropulmonary blastoma, plasma cell neoplasia, primary central nervous system lymphoma, prostate cancer, rectal cancer, renal cell carcinoma, renal pelvis and ureter transitional cell cancer, retinoblastoma, rhabdomyosarcoma, salivary gland cancer, sarcomas, sebaceous gland carcinoma, seminoma, skin cancers, skin carcinoma, merkel cell carcinoma, small intestine cancer, soft tissue sarcoma, squamous cell carcinoma, gastric cancer, sweat gland carcinoma, synovioma, T-cell lymphoma, testicular tumor, throat cancer, thymoma, thymic carcinoma, thyroid cancer, trophoblastic tumor(gestational), cancers of unknown primary site, urethral cancer, uterine sarcoma, vaginal cancer, vulvar cancer, Waldenstrom macroglobulinemia, Wilms tumor, or combinations thereof. The tumors or cancers may be associated with various types of organs. Non-limiting examples of organs may include brain, breast, colon, liver, lung, kidney, prostate, ovary, spleen, lymph node (including tonsil), thyroid, pancreas, heart, skeletal muscle, intestine, larynx, esophagus, stomach, small intestine, or combinations thereof.

[0127] The power of the classifier to correctly predict disease status can be measured in terms of the accuracy of the assay, the sensitivity of the assay, the specificity of the assay, the positive predictive value (PPV) of the assay, the negative predictive value (NPV) of the assay, or the "Area Under a Curve" (AUC), for example, the area under a Receiver Operating Characteristic (ROC) curve. As used herein, accuracy is a measure of the fraction of misclassified samples. Accuracy may be calculated as the total number of correctly classified samples divided by the total number of samples, e.g., in a test population. Sensitivity is a measure of the "true positives" that are predicted by a test to be positive, and may be calculated as the number of correctly identified cancer samples divided by the total number of cancer samples. Specificity is a measure of the "true negatives" that are predicted by a test to be negative, and may be calculated as the number of correctly identified normal samples divided by the total number of normal samples. AUC is a measure of the area under a Receiver Operating Characteristic curve, which is a plot of sensitivity vs. the false positive rate (1 -specificity). The greater the AUC, the more powerful the predictive value of the test. In some embodiments, the trained machine learning classifier predicts cancer at an area under curve (AUC) of at least about 0.9. In some embodiments, the trained machine learning classifier predicts cancer at an area under curve (AUC) of at least about 0.91, 0.92, 0.93, 0.94, 0.95, 0.96, 0.97, 0.98, 0.99, or higher. In some embodiments, the trained machine learning classifier predicts cancer at an area under curve (AUC) of at least about 0.50, at least about 0.55, at least about 0.60, at least about 0.65, at least about 0.70, at least about 0.75, at least about 0.80, at least about 0.85, at least about 0.90, at least about 0.95, or more. Other useful measures of the utility of a test include the "positive predictive value," which is the percentage of actual positives who test as positives, and the "negative predictive value," which is the percentage of actual negatives who test as negatives. In some embodiments, the trained machine learning classifier predicts cancer at a positive predictive value of at least about 50%, at least about 55%, at least about 60%, at least about 65%, at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, at least about 95%, or more In some embodiments, the trained machine learning classifier predicts cancer at a negative predictive value of at least about 50%, at least about 55%,at least about 60%, at least about 65%, at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, at least about 95%, or more. In some embodiments, the level of one or more RNA biomarkers (e.g,. one or more k-mers) in samples derived from subjects having different cancer statuses show a statistically significant difference of at least p = 0.05, e.g., p = 0.05, p = 0.01, p = 0.005, p = 0.001, etc. relative to normal subjects, as determined relative to a suitable control. In some embodiments, diagnostic tests that use RNA biomarkers described herein individually or in combination show an accuracy of at least about 75%, e.g., an accuracy of at least about 75%, about 80%, about 85%, about 90%, about 95%, about 97%, about 99% or about 100%. For example, the trained machine learning classifier may predict cancer at an accuracy of at least about 50%, at least about 55%, at least about 60%, at least about 65%, at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, at least about 95%, or more. In other embodiments, diagnostic tests that use RNA biomarkers described herein individually or in combination show a specificity of at least about 75%, e.g., a specificity of at least about 75%, about 80%, about 85%, about 90%, about 95%, about 97%, about 99% or about 100%. For example, the trained machine learning classifier may predict cancer at a specificity of at least about 50%, at least about 55%, at least about 60%, at least about 65%, at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, at least about 95%, or more. In some embodiments, diagnostic tests that use RNA biomarkers described herein individually or in combination show a sensitivity of at least about 75%, e.g., a sensitivity of at least about 75%, about 80%, about 85%, about 90%, about 95%, about 97%, about 99% or about 100%. For example, the trained machine learning classifier may predict cancer at a sensitivity of at least about 50%, at least about 55%, at least about 60%, at least about 65%, at least about 70%, at least about 75%, at least about 80%, at least about 85%, at least about 90%, at least about 95%, or more. In other embodiments, diagnostic tests that use RNA biomarkers described herein individually or in combination show a specificity and sensitivity of at least about 75% each, e.g., a specificity and sensitivity of at least about 75%, about 80%, about 85%, about 90%, about 95%, about 97%, about 99% or about 100% (for example, a specificity of at least about 80% and sensitivity of at least about 80%, or for example, a specificity of at least about 80% and sensitivity of at least about 95%).

[0128] In another example, biomarkers (e.g., one or more k-mers) may be identified as being associated with certain type(s) of cancer (e.g., kidney, liver, lung, breast, colorectal, gastric). In some embodiments, the biomarker(s) may be identified as being associated with more than one type of cancer. In some embodiments, one particular small RNA biomarker may be associated with only one type of cancer.

[0129] Accordingly, methods are provided for diagnosing a cancer in a subject, by determining the level of at least one small RNA in a sample containing circulating small RNA from the subject, wherein a difference in the level of the at least one small RNA versus that in a control subject is indicative of the cancer in the test subject. For example, the present disclosure provides a method of determining the level of at least one small RNA in a sample containing circulating small RNA derived from the subject, wherein an increase in the level of the at least one small RNA relative to a control is indicative of breast cancer in the subject. In another example, the present disclosure provides a method of determining the level of at least one small RNA in a sample containing circulating small RNA derived from the subject, wherein an increase in the level of the at least one small RNA relative to a control is indicative of lung cancer in a subject. In another example, the present disclosure provides a method of determining the level of at least one small RNA in a sample containing circulating small RNA derived from the subject, wherein an increase in the level of the at least one small RNA relative to a control is indicative of liver cancer in a subject. In another example, the present disclosure provides a method of determining the level of at least one small RNA in a sample containing circulating small RNA derived from the subject, wherein an increase in the level of the at least one small RNA relative to a control is indicative of kidney cancer in a subject. In another example, the present disclosure provides a method of determining the level of at least one small RNA in a sample containing circulating small RNA derived from the subject, wherein an increase in the level of the at least one small RNA relative to a control is indicative of gastric cancer in a subject. In another example, the present disclosure provides a method of determining the level of at least one small RNA in a sample containing circulating small RNA derived from the subject, wherein an increase in the level of the at least one small RNA relative to a control is indicative of colorectal cancer in a subject.

[0130] Optionally, the method may further comprise providing a diagnosis that the subject has or does not have a cancer based on the level of at least one small RNA biomarker (e.g., one or more k-mer) in the sample. In addition or alternatively, the method may further comprise correlating a difference in the level or levels of at least one small RNA relative to a suitable control with a diagnosis of cancer in the subject, or a difference with respect to the presence or absence of one or more small RNA biomarker (e.g., one or more k-mer) previously identified as differentiating cancer versus non-cancer samples (in which case a control is not necessary). In some embodiments, such a diagnosis may be provided directly to the subject, or it may be provided to another party involved in the subject's care.

[0131] While individual small RNA biomarkers are useful in diagnostic applications for cancer, as shown herein, a combination of small RNA biomarkers may provide greater predictive value of cancer status than the small RNA biomarkers when used alone. Specifically, the detection of a plurality of small RNA biomarkers can increase the accuracy, sensitivity, and / or specificity of a diagnostic test. The detection of a plurality of small RNA biomarkers may also assist in narrowing down the cancer type, subtype, tissue of origin, or status of the subject. Multiple small RNA biomarkers (e.g. one or more k-mers) can be used to train the classifiers as described herein.

[0132] Accordingly, methods are provided for diagnosing cancer in a subject, by determining the level, or presence or absence of small RNA biomarkers (e.g., one or more k- mers) in a sample containing cell-free RNA from the subject, wherein a difference in the level, or presence or absence of small RNA biomarkers versus that in a normal subject (as determined relative to a suitable control) is indicative of cancer in the subject.

[0133] Also provided is a method of diagnosing cancer in a subject by determining the levels of two or more small RNA biomarkers in a sample containing cell-free small RNA from the subject, comparing the levels of the two or more small RNAs in the sample to a set of data representing levels of the same small RNAs present in normal subjects and subjects having a cancer, and diagnosing the subject as having or not having the cancer based on the comparison.

[0134] Comparison of the biological sample from the subject with the set of data may be assisted by a classification algorithm, which computes whether or not a statistically significant difference exists between the collective levels of the two or more small RNAs in the sample, and the levels of the same small RNAs present in a control subject having a certain cancer or cancer state.Methods of Treatment

[0135] In some embodiments, where a subject is diagnosed with cancer by the methods described herein, the present disclosure further provides methods of treating such subjects identified to have cancer. In some embodiment, the methods further comprise administering a therapeutically effective amount of a cancer therapeutic to the test subject. In another embodiment, the present disclosure relates to a method of treating a subject having cancer, comprising identifying a subject having cancer as classified by the trained machine learning classifier and administering a therapeutically effective amount of a cancer therapeuticto the subject. In some embodiment, the methods further comprise performing a surgery on a subject. In some embodiments, the method further comprises administering an immunotherapy. In some embodiments, the method further comprises administering a chemotherapy. In some embodiments, the method further comprises administering a radiation therapy. Treatment can comprise a combinations of different therapies or therapeutic interventions.

[0136] For example, a cancer therapeutic can comprise substances approved by the U.S. Food and Drug Administration for the treatment of breast cancer. Drugs approved to treat breast cancer include, but are not limited to, Abemaciclib, Abitrexate (Methotrexate), Abraxane (Paclitaxel Albumin-stabilized Nanoparticle Formulation), Ado-Trastuzumab Emtansine, Afinitor (Everolimus), Anastrozole, Aredia (Pamidronate Disodium), Arimidex (Anastrozole), Aromasin (Exemestane), Capecitabine, Clafen (Cyclophosphamide), Cyclophosphamide, Cytoxan (Cyclophosphamide), Docetaxel, Doxorubicin Hydrochloride, Ellence (Epirubicin Hydrochloride), Epirubicin Hydrochloride, Eribulin Mesylate, Everolimus, Exemestane, 5-FU (Fluorouracil Injection), Fareston (Toremifene), Faslodex (Fulvestrant), Femara (Letrozole), Fluorouracil Injection, Fol ex (Methotrexate), Fol ex PFS (Methotrexate), Fulvestrant, Gemcitabine Hydrochloride, Gemzar (Gemcitabine Hydrochloride), Goserelin Acetate, Halaven (Eribulin Mesylate), Herceptin (Trastuzumab), Ibrance (Palbociclib), Ixabepilone, Ixempra (Ixabepilone), Kadcyla (Ado-Trastuzumab Emtansine), Kisqali (Ribociclib), Lapatinib, Ditosylate, Letrozole, Megestrol Acetate, Methotrexate, Methotrexate LPF (Methotrexate), Mexate (Methotrexate), Mexate-AQ (Methotrexate), Neosar (Cyclophosphamide), Neratinib Maleate, Nerlynx (Neratinib Maleate), Nolvadex (Tamoxifen Citrate), Paclitaxel, Paclitaxel Albumin-stabilized Nanoparticle Formulation, Palbociclib, Pamidronate Disodium, Peijeta (Pertuzumab), Pertuzumab, Ribociclib, Tamoxifen Citrate, Taxol (Paclitaxel), Taxotere (Docetaxel), Thiotepa, Toremifene, Trastuzumab, Tykerb (Lapatinib Ditosylate), Velban (Vinblastine Sulfate), Velsar (Vinblastine Sulfate), Verzenio (Abemaciclib), Vinblastine Sulfate, Xeloda (Capecitabine), and Zoladex (Goserelin Acetate).

[0137] The cancer therapeutics may be administered to a subject using a pharmaceutical composition. Suitable pharmaceutical compositions comprise a pharmaceutically effective amount of a breast cancer therapeutic (or a pharmaceutically acceptable salt or ester thereof), and optionally comprise a pharmaceutically acceptable carrier). In certain embodiments, these compositions optionally further comprise one or more additional therapeutic agents.

[0138] As used herein, the term "pharmaceutically acceptable salt" refers to those salts which are, within the scope of sound medical judgment, suitable for use in contact with the tissues of humans and lower animals without undue toxicity, irritation, allergic response and the like, and are commensurate with a reasonable benefit / risk ratio. Pharmaceutically acceptable salts of amines, carboxylic acids, and other types of compounds, may be used. For example, S. M. Berge, et al. describe pharmaceutically acceptable salts in detail in J. Pharmaceutical Sciences, 66: 1-19 (1977), incorporated herein by reference. The salts can be prepared in situ during the final isolation and purification of the compounds of the present disclosure, or separately by reacting a free base or free acid function with a suitable reagent. For example, a free base function can be reacted with a suitable acid. Furthermore, where the compounds carry an acidic moiety, suitable pharmaceutically acceptable salts thereof may, include metal salts such as alkali metal salts, e.g. sodium or potassium salts; and alkaline earth metal salts, e.g. calcium or magnesium salts.

[0139] The term "pharmaceutically acceptable ester", as used herein, refers to esters that hydrolyze in vivo and include those that break down readily in the human body to leave the parent compound or a salt thereof. Suitable ester groups include, for example, those derived from pharmaceutically acceptable aliphatic carboxylic acids, particularly alkanoic, alkenoic, cycloalkanoic and alkanedioic acids, in which each alkyl or alkenyl moiety advantageously has not more than 6 carbon atoms.

[0140] As described above, the pharmaceutical compositions may additionally comprise a pharmaceutically acceptable carrier. The term carrier includes any and all solvents, diluents, or other liquid vehicle, dispersion or suspension aids, surface active agents, isotonic agents, thickening or emulsifying agents, preservatives, solid binders, lubricants and the like, suitable for preparing the particular dosage form desired. Remington's Pharmaceutical Sciences, Sixteenth Edition, E. W. Martin (Mack Publishing Co., Easton, Pa., 1980) discloses various carriers used in formulating pharmaceutical compositions and known techniques for the preparation thereof. Some examples of materials which can serve as pharmaceutically acceptable carriers include, but are not limited to, sugars such as lactose, glucose and sucrose; starches such as corn starch and potato starch; cellulose and its derivatives such as sodium carboxymethyl cellulose, ethyl cellulose and cellulose acetate; powdered tragacanth; malt; gelatin; talc; excipients such as cocoa butter and suppository waxes; oils such as peanut oil, cottonseed oil; safflower oil, sesame oil; olive oil; corn oil and soybean oil; glycols; such as propylene glycol; esters such as ethyl oleate and ethyl laurate; agar; buffering agents such as magnesium hydroxide and aluminum hydroxide; alginic acid; pyrogen-free water; isotonicsaline; Ringer's solution; ethyl alcohol, and phosphate buffer solutions, as well as other nontoxic compatible lubricants such as sodium lauryl sulfate and magnesium stearate, as well as coloring agents, releasing agents, coating agents, sweetening, flavoring and perfuming agents, preservatives and antioxidants can also be present in the composition, according to the judgment of the formulator.

[0141] Compositions for use in the present disclosure may be formulated to have any concentration of the cancer therapeutic desired. In some embodiments, the composition is formulated such that it comprises a therapeutically effective amount of the cancer therapeutic.Computer Systems

[0142] The present disclosure provides computer systems that are programmed to implement methods of the disclosure. FIG. 2 shows a computer system 101 that is programmed or otherwise configured to, for example, implement a trained machine learning classifier to determine the cancer status a subject from small RNA sequencing read derived from a sample from the subject, plan a course of treatment for a subject based on small RNA sequencing reads derived from a sample from the subject, monitor the cancer status of a subject over time (e.g., before and after treatment) based on small RNA sequencing reads derived from a sample from the subject, measure the efficacy or tolerability of a therapeutic intervention for an individual based on on small RNA sequencing reads derived from a sample from the subject, or process a biological sample as described herein.

[0143] The computer system 101 can regulate various aspects of analysis, calculation, and generation of the present disclosure, such as, for example, implement a trained machine learning classifier to determine the cancer status a subject on small RNA sequencing reads derived from a sample from the subject, planning a course of treatment for a subject based on small RNA sequencing reads derived from a sample from the subject, monitoring the cancer status of a subject over time (e.g., before and after treatment) based on small RNA sequencing reads derived from a sample from the subject, measuring the efficacy or tolerability of a therapeutic intervention for an individual based on small RNA sequencing reads derived from a sample from the subject, or processing a biological sample as described herein. The computer system 101 can be an electronic device of a user or a computer system that is remotely located with respect to the electronic device. The electronic device can be a mobile electronic device.

[0144] The computer system 101 includes a central processing unit (CPU, also “processor” and “computer processor” herein) 105, which can be a single core or multi coreprocessor, or a plurality of processors for parallel processing. The computer system 101 also includes memory or memory location 110 (e.g., random-access memory, read-only memory, flash memory), electronic storage unit 115 (e.g., hard disk), communication interface 120 (e.g., network adapter) for communicating with one or more other systems, and peripheral devices 125, such as cache, other memory, data storage and / or electronic display adapters. The memory 110, storage unit 115, interface 120 and peripheral devices 125 are in communication with the CPU 105 through a communication bus (solid lines), such as a motherboard. The storage unit 115 can be a data storage unit (or data repository) for storing data. The computer system 101 can be operatively coupled to a computer network (“network”) 130 with the aid of the communication interface 120. The network 130 can be the Internet, an internet and / or extranet, or an intranet and / or extranet that is in communication with the Internet.

[0145] The network 1310 In some embodiments is a telecommunication and / or data network. The network 130 can include one or more computer servers, which can enable distributed computing, such as cloud computing. For example, one or more computer servers may enable cloud computing over the network 130 (“the cloud”) to perform various aspects of analysis, calculation, and generation of the present disclosure, such as, for example, implementing a trained machine learning classifier to determine the cancer status a subject on small RNA sequencing reads derived from a sample from the subject, planning a course of treatment for a subject based on small RNA sequencing reads derived from a sample from the subject, monitoring the cancer status of a subject over time (e.g., before and after treatment) based on small RNA sequencing reads derived from a sample from the subject, measuring the efficacy or tolerability of a therapeutic intervention for an individual based on small RNA sequencing reads derived from a sample from the subject, or processing a biological sample as described herein. Such cloud computing may be provided by cloud computing platforms such as, for example, Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform, and IBM cloud. The network 130, In some embodiments with the aid of the computer system 101, can implement a peer-to-peer network, which may enable devices coupled to the computer system 101 to behave as a client or a server.

[0146] The CPU 105 may comprise one or more computer processors and / or one or more graphics processing units (GPUs). The CPU 105 can execute a sequence of machine- readable instructions, which can be embodied in a program or software. The instructions may be stored in a memory location, such as the memory 110. The instructions can be directed to the CPU 105, which can subsequently program or otherwise configure the CPU 105 to implementmethods of the present disclosure. Examples of operations performed by the CPU 105 can include fetch, decode, execute, and writeback.

[0147] The CPU 105 can be part of a circuit, such as an integrated circuit. One or more other components of the system 101 can be included in the circuit. In some embodiments, the circuit is an application specific integrated circuit (ASIC).

[0148] The storage unit 115 can store files, such as drivers, libraries and saved programs. The storage unit 115 can store user data, e.g., user preferences and user programs. The computer system 101 in some embodiments can include one or more additional data storage units that are external to the computer system 101, such as located on a remote server that is in communication with the computer system 101 through an intranet or the Internet.

[0149] The computer system 101 can communicate with one or more remote computer systems through the network 130. For instance, the computer system 101 can communicate with a remote computer system of a user. Examples of remote computer systems include personal computers (e.g., portable PC), slate or tablet PC’s (e.g., Apple® iPad, Samsung® Galaxy Tab), telephones, Smart phones (e.g., Apple® iPhone, Android-enabled device, Blackberry®), or personal digital assistants. The user can access the computer system 101 via the network 130.

[0150] Methods as described herein can be implemented by way of machine (e.g., computer processor) executable code stored on an electronic storage location of the computer system 101, such as, for example, on the memory 110 or electronic storage unit 115. The machine executable or machine readable code can be provided in the form of software. During use, the code can be executed by the processor 105. In some embodiments, the code can be retrieved from the storage unit 115 and stored on the memory 110 for ready access by the processor 105. In some situations, the electronic storage unit 115 can be precluded, and machine-executable instructions are stored on memory 110.

[0151] The code can be pre-compiled and configured for use with a machine having a processer adapted to execute the code or can be compiled during runtime. The code can be supplied in a programming language that can be selected to enable the code to execute in a precompiled or as-compiled fashion.

[0152] Embodiments of the systems and methods provided herein, such as the computer system 101, can be embodied in programming. Various aspects of the technology maybe thought of as “products” or “articles of manufacture” typically in the form of machine (or processor) executable code and / or associated data that is carried on or embodied in a type of machine readable medium. Machine-executable code can be stored on an electronic storage unit, such as memory (e.g., read-only memory, random-access memory, flash memory) or a hard disk. “Storage” type media can include any or all of the tangible memory of the computers, processors or the like, or associated modules thereof, such as various semiconductor memories, tape drives, or disk drives, which may provide non-transitory storage at any time for the software programming. All or portions of the software may at times be communicated through the Internet or various other telecommunication networks. Such communications, for example, may enable loading of the software from one computer or processor into another, for example, from a management server or host computer into the computer platform of an application server. Thus, another type of media that may bear the software elements includes optical, electrical and electromagnetic waves, such as used across physical interfaces between local devices, through wired and optical landline networks and over various air-links. The physical elements that carry such waves, such as wired or wireless links, optical links or the like, also may be considered as media bearing the software. As used herein, unless restricted to non-transitory, tangible “storage” media, terms such as computer or machine “readable medium” refer to any medium that participates in providing instructions to a processor for execution.

[0153] Hence, a machine readable medium, such as computer-executable code, may take many forms, including a tangible storage medium, a carrier wave medium or physical transmission medium. Non-volatile storage media include, for example, optical or magnetic disks, such as any of the storage devices in any computer(s) or the like, such as may be used to implement the databases, etc. shown in the drawings. Volatile storage media include dynamic memory, such as main memory of such a computer platform. Tangible transmission media include coaxial cables; copper wire and fiber optics, including the wires that comprise a bus within a computer system. Carrier-wave transmission media may take the form of electric or electromagnetic signals, or acoustic or light waves such as those generated during radio frequency (RF) and infrared (IR) data communications. Common forms of computer-readable media therefore include for example: a floppy disk, a flexible disk, hard disk, magnetic tape, any other magnetic medium, a CD-ROM, DVD or DVD-ROM, any other optical medium, punch cards paper tape, any other physical storage medium with patterns of holes, a RAM, a ROM, a PROM and EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave transporting data or instructions, cables or links transporting such a carrier wave, or any other medium from which a computer may read programming code and / or data. Many of these formsof computer readable media may be involved in carrying one or more sequences of one or more instructions to a processor for execution.

[0154] The computer system 101 can include or be in communication with an electronic display 135 that comprises a user interface (UI) 140 for providing, for example, a visual display indicative of the cancer status a subject based on small RNA sequencing reads derived from a sample from the subject, a visual display indicative of a planned course of treatment for a subject based on small RNA sequencing reads derived from a sample from the subject, a visual display indicative of monitoring the cancer status of a subject over time (e.g., before and after treatment) based on small RNA sequencing reads derived from a sample from the subject, a visual display indicated of a measurement the efficacy or tolerability of a therapeutic intervention for an individual based on small RNA sequencing reads derived from a sample from the subject, or a visual display indicative of processing a biological sample as described herein. Examples of UIs include, without limitation, a graphical user interface (GUI) and web-based user interface.

[0155] Methods and systems of the present disclosure can be implemented by way of one or more algorithms. An algorithm can be implemented by way of software upon execution by the central processing unit 105. The algorithm can, for example, determine the cancer status a subject from small RNA expression data derived from a sample from the subject, plan a course of treatment for a subject based on small RNA sequencing reads derived from a sample from the subject, monitor the cancer status of a subject over time (e.g., before and after treatment) based on small RNA sequencing reads derived from a sample from the subject, measure the efficacy or tolerability of a therapeutic intervention for an individual based on small RNA sequencing reads derived from a sample from the subject, or process a biological sample as described herein.EXAMPLESExample 1

[0156] smRNA-seq data was obtained from samples of prostate cancer tumors and normal tissues (n=499 and 679, respectively) (via publicly available data from the The Cancer Genome Atlas (TCGA)). smRNA sequencing data was processed via a k-mer analysis on the sequence reads to identify a set of k-mers that were enriched in the prostate cancer tumor samples compared to the normal sample, known as “oncKmers”. A set of 1,133,966 statistically significant oncKmers was identified. The abundance of these oncKmers was counted as features in a dataset of 96 prostate cancer cases and 96 healthy controls with no history of cancer. A 5- fold cross validation approach was then used to train a logistic regression model with elastic netregularization for predicting prostate cancer status using these features. The prediction AUC was greater than 0.9, with a true positive rate of 0.53 at the specificity of 0.95 for held-out prostate cancer samples.Example 2

[0157] smRNA-seq data was obtained from samples of lung cancer tumors and normal tissues (via publicly available data from the The Cancer Genome Atlas (TCGA)). smRNA sequencing data was processed via a k-mer analysis on the sequence reads to identify a set of k-mers that were enriched in the lung cancer tumor samples compared to the normal sample. The set of selected oncKmers were assembled to produce longer sequences to be set of oncKmers and a set of 191,353 statistically significant oncKmers was identified. The abundance of these oncKmers was counted as features in a dataset of lung cancer cases and healthy controls with no history of cancer. A 5-fold cross validation approach was then used to train a deeplearning based generative Al model for predicting lung cancer status using these features. As shown in FIG. 1., the prediction AUC was greater than 0.99, with a true positive rate of 0.95 at the specificity of 0.95 for held-out lung cancer samples.

[0158] While preferred embodiments of the present invention 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. It is not intended that the invention be limited by the specific examples provided within the specification. While the invention has been described with reference to the aforementioned specification, the descriptions and illustrations of the embodiments herein are not meant to be construed in a limiting sense. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the invention. Furthermore, it shall be understood that all aspects of the invention are not limited to the specific depictions, configurations or relative proportions set forth herein which depend upon a variety of conditions and variables. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in practicing the invention. It is therefore contemplated that the invention shall also cover any such alternatives, modifications, variations or equivalents. It is intended that the following claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A computer-implemented method of detecting a presence or an absence of cancer in a subject, comprising:(a) obtaining a plurality of sequencing reads derived from cell-free RNA of said subject;(b) processing said sequencing reads to generate a plurality of k-mer sequences, wherein said processing is performed without mapping or aligning said sequencing reads to a reference genome sequence;(c) analyzing said plurality of k-mer sequences using a trained machine learning classifier, wherein said trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject, based at least in part on presence or absence of cancer-associated k-mer sequences in said plurality of k-mer sequences; and(d) detecting said presence or absence of cancer in said subject, based at least in part on said output generated in (c).

2. The method of claim 1, wherein said cell-free RNA comprises a small RNA.

3. The method of claim 1, further comprising, prior to (a), sequencing said cell-free RNA to generate said plurality of sequencing reads.

4. The method of claim 1, further comprising, prior to (a), extracting said cell-free RNA from a biological sample of said subject.

5. The method of claim 1, further comprising, prior to (a), isolating said cell-free RNA from a biological sample of said subject.

6. The method of claim 1, further comprising, prior to (a), obtaining a biological sample from said subject comprising said cell-free RNA.

7. The method of claim 6, wherein said biological sample is a serum sample or plasma sample.

8. The method of claim 1, wherein said cancer is selected from the group consisting of brain cancer, breast cancer, colon cancer, liver cancer, lung cancer, kidney cancer, prostate cancer, ovarian cancer, spleen cancer, lymphoma, thyroid cancer, pancreatic cancer, intestinal cancer, laryngeal cancer, esophageal cancer, stomach cancer, small intestinal cancer.

9. The method of claim 8, wherein the cancer comprises prostate cancer.

10. The method of claim 1, wherein analyzing said plurality of k-mer sequences comprises analyzing k-mer sequences having a length of at least about 10 nucleotides to about 50 nucleotides.

11. The method of claim 1, wherein analyzing said plurality of k-mer sequences comprises analyzing k-mer sequences having a length of at least about 10 nucleotides.

12. The method of claim 11, wherein analyzing said plurality of k-mer sequences comprises analyzing k-mer sequences having a length of about 30 nucleotides13. The method of claim 1, wherein the trained machine learning classifier is trained on a set of k-mer sequences from sequence reads from cancer samples.

14. The method of claim 1, wherein the trained learning classifier is trained on a set of k-mer sequences from sequence reads from non-cancer samples.

15. The method of claim 1, further comprising, prior to (c), training said trained machine learning classifier using sequencing reads of a non-cancer sample and a cancer sample.

16. The method of claim 1, wherein said machine learning classifier comprises using multiple linear regression (MLR), partial least squares (PLS) regression and principal components regression (PCR), binary decision trees, artificial neural networks, back propagation networks, discriminant analyses, logistic classifiers, and support vector classifiers.

17. The method of claim 1, wherein (c) further comprises determining a number of said sequencing reads comprising a k-mer sequence.

18. The method of claim 1, wherein a k-mer sequence of the plurality of k-mer sequences does not align to a wild type human reference genome.

19. The method of claim 1, wherein a k-mer sequence of the plurality of k-mer sequences comprises a sequence that is alignable to a splice site.

20. The method of claim 1, wherein a k-mer sequence of the plurality of k-mer sequences comprises a sequence that is alignable to a gene fusion junction.

21. The method of claim 1, wherein (d) further comprises identifying a cancer cell type of said cancer.

22. The method of claim 1, wherein (d) further comprises identifying a cancer tissue of origin of said cancer.

23. The method of claim 1, wherein (d) further comprises identifying a cancer subtype of said cancer.

24. The method of claim 1, wherein (d) further comprises identifying a plurality of cancer types of said cancer.

25. The method of claim 1, wherein the subject is suspected of having cancer.

26. The method of claim 1, wherein the subject was previously diagnosed with cancer.

27. The method of claim 1, wherein the subject is undergoing treatment for cancer.

28. The method of claim 1, further comprising treating said subject for said detected presence of said cancer.

29. The method of claim 28, wherein said treating comprises administering a surgery, a chemotherapy, a radiation therapy, an immunotherapy, or a combination thereof.

30. A computer-implemented method of generating a trained machine learning classifier for cancer prediction, the method comprising:(a) providing (i) a first plurality of samples derived from non-cancer subjects, and (ii) a second plurality of samples derived from subjects with cancer;(b) assaying cell-free RNA derived from said first and second plurality of samples to generate a plurality of cancer-derived sequencing reads and a plurality of non- cancer-derived sequencing reads;(c) processing said plurality of cancer-derived sequencing reads and said plurality of non-cancer-derived sequencing reads to generate a plurality of cancer-associated k- mer sequences, wherein said processing is performed without mapping or aligning said sequencing reads to a reference genome sequence;(d) training a machine learning classifier using said plurality of cancer-associated k-mer sequences, thereby generating said trained machine learning classifier, wherein said trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject.

31. The method of claim 30, wherein (c) further comprises determining a number of said plurality of cancer-derived sequencing reads or said plurality of non-cancer-derived sequencing reads that comprise a given cancer-associated k-mer sequence.

32. The method of claim 31, wherein (c) further comprises comparing a number of said plurality of cancer-derived sequencing reads that comprise said given cancer-associated k-mer sequence to a number of said plurality of non-cancer-derived sequence reads that comprise said given cancer-associated k-mer sequence.

33. The method of claim 32, wherein (d) further comprises determining that a cancer- associated k-mer sequence of said plurality of cancer-associated k-mer sequences being present in a larger number of said plurality of cancer-derived sequencing reads as compared to said plurality of non-cancer-derived sequencing reads as indicative of the presence of cancer.

34. The method of claim 31, wherein (c) further comprises generating an output of a number of said plurality of cancer-derived sequencing reads or said plurality of non-cancer-derived sequencing reads that comprise a given cancer-associated k-mer sequence for each of said first and second plurality of samples.

35. The method of claim 30, wherein said trained machine learning classifier is further programmed to provide a prediction of a cancer cell type of said cancer.

36. The method of claim 30, wherein said trained machine learning classifier is further programmed to provide a prediction of a cancer tissue of origin of said cancer.

37. The method of claim 30, wherein said trained machine learning classifier is further programmed to provide a prediction of a cancer subtype of said cancer.

38. The method of claim 30, wherein said trained machine learning classifier predicts cancer at an accuracy of at least about 90%.

39. The method of claim 30, wherein said trained machine learning classifier predicts cancer at an area under curve (AUC) of at least about 0.90.

40. The method of claim 30, wherein said cell-free RNA comprises a small RNA.

41. A non-transitory computer-readable storage medium comprising a set of instructions recorded thereon, which, when executed by a processor, cause the processor to implement a method for detecting a presence or an absence of cancer in a subject, said method comprising:(i) obtaining a plurality of sequencing reads derived from cell-free RNA of said subject;(ii) processing said sequencing reads to generate a plurality of k-mer sequences, wherein said processing is performed without mapping or aligning said sequencing reads to a reference genome sequence;(iii) analyzing said plurality of k-mer sequences using a trained machine learning classifier, wherein said trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject, based at least in part on presence or absence of cancer-associated k-mer sequences in said plurality of k-mer sequences; and(iv) detecting said presence or absence of cancer in said subject, based at least in part on said output generated in (iii).

42. A non-transitory computer-readable storage medium comprising a set of instructions recorded thereon, which, when executed by a processor, cause the processor to implement a method for generating a trained machine learning classifier for cancer prediction, said method comprising:(i) obtaining plurality of cancer-derived sequencing reads and a plurality of non- cancer-derived sequencing reads;(ii) processing said plurality of cancer-derived sequencing reads and said plurality of non-cancer-derived sequencing reads to generate a plurality of cancer- associated k-mer sequences, wherein said processing is performed without mapping or aligning said sequencing reads to a reference genome sequence;(iii) training a machine learning classifier using said plurality of cancer- associated k-mer sequences, thereby generating said trained machine learning classifier, wherein said trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject.

43. A system for detecting a presence or an absence of cancer in a subject, said system comprising: at least one system memory comprising (a) a plurality of sequencing reads derived from cell-free RNA of said subject, and (b) a trained machine learning classifier; and a computer memory coupled to said at least one system memory and programed to (i) process said sequencing reads to generate a plurality of k-mer sequences, wherein said processing is performed without mapping or aligning said sequencing reads to a reference genome sequence;(ii) analyze said plurality of k-mer sequences using said trained machine learning classifier, wherein said trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject, based at least in part on presence or absence of cancer-associated k-mer sequences in said plurality of k-mer sequences; and(iii) detect said presence or absence of cancer in said subject, based at least in part on said output generated in (ii).

44. A system for generating a trained machine learning classifier for cancer prediction, the system comprising: at least one system memory comprising a plurality of cancer-derived sequencing reads and a plurality of non-cancer-derived sequencing reads; and a computer memory coupled to said at least one system memory and programmed to:(i) process said plurality of cancer-derived sequencing reads and said plurality of non-cancer-derived sequencing reads to generate a plurality of cancer-associated k-mer sequences, wherein said processing is performed without mapping or aligning said sequencing reads to a reference genome sequence; and(ii) train a machine learning classifier using said plurality of cancer-associated k- mer sequences, thereby generating said trained machine learning classifier, wherein said trained machine learning classifier is programmed to generate an output indicative of a presence or an absence of cancer in a subject.

Images