Models for determining retrotransposon (RT) insertion burden and implications in cancer therapy

Abstract

The present disclosure provides methods, devices, and systems for determining the retrotransposon (RT) burden, RT expression and fitness of mutant p53 in a subject. The RT burden, RT RNA expression and fitness of mutant p53 may be used to determine whether a subject is at risk for cancer and will benefit from a particular anti -cancer therapy such as immune checkpoint inhibitor therapy, adoptive cell therapy, or prophylactic cancer vaccine therapy.

Description

MODELS FOR DETERMINING RETROTRANSPOSON (RT) INSERTION BURDEN AND IMPLICATIONS IN CANCER THERAPYSELECTIONCROSS-REFERENCE TO RELATED PATENT APPLICATIONS

[0001] This application claims the benefit of and priority to U.S. Application No. 63 / 464,529, filed May 5, 2023, the contents of which are incorporated herein by reference in its entirety.TECHNICAL FIELD

[0002] The present technology relates to methods, devices, and systems for determining the retrotransposon (RT) burden, RT expression and fitness of mutant p53 in a subject. The RT burden, RT RNA expression and fitness of mutant p53 may be used to determine whether a subject is at risk for cancer and will benefit from a particular anti-cancer therapy such as immune checkpoint inhibitor therapy, adoptive cell therapy, or prophylactic cancer vaccine therapy.BACKGROUND

[0003] More than half of the human genome is composed of repeat sequences1’2,3.Normally, various processes such as epigenetic repression silence these repeats4but oncogenesis is associated with disruptions to these pathways5, 6. In cancer, repeats can be reexpressed as RNA, translated in some cases into protein, and may be actively involved in genome instability and cancer immunogenicity7’8,9,10,11,12. The LINE-1 (LI) element is an especially interesting class of repeats which possesses the ability to reinsert itself via retrotransposition at new loci in the human genome13. An intact LI element is -6 kb in length, but most of the >1 million copies that comprise -17% of the human genome are 5’ truncated2. There are just over 100 Lis capable of coding for the full length ORF Ip and ORF2p proteins, the latter functioning as the reverse transcriptase and endonuclease1needed for LI activity.

[0004] Despite its frequent reintegration in cancer genomes11, measuring LI retrotransposition from short-read sequencing data to accurately quantify its movement in cancer poses technical challenges due to the many genomic copies of highly similar LI sequences, as well as different modes of the retrotransposition process (FIGs. 1A-1D) making it difficult to disambiguate the source of LI -containing sequencing reads.SUMMARY

[0005] Aspect of the present application relate to systems and methods of selecting cancer therapies based on magnitudes and frequencies of retrotransposon (RT) insertions in subject genomes. One or more processors may obtain, for a subject at risk of or diagnosed with cancer, a first dataset identifying genomic sequences derived from a first biological sample of the subject, wherein the genomic sequences are obtained via next-generation sequencing and comprise a plurality of sequencing reads. The one or more processors may identify, using the first dataset, (i) an RT burden metric corresponding to a RT indicating a level of RT activity in the genomic sequences of the subject and (ii) a plurality of p53 sequences in the genomic sequences of the subject, each of the plurality of p53 sequences including a plurality of non- silent p53 mutant sequences and a plurality of wild-type p53 sequences. The RT activity may comprise canonical insertions, inversion-containing insertions, 5’ truncated insertions, full- length insertions, and 3’ transductions. The plurality of non-silent p53 mutant sequences may comprise missense mutations, nonsense mutations, and frameshift mutations. For each p53 sequence of the plurality of p53 sequences, the one or more processors may identify , a mutated position of each p53 mutant sequence of the plurality of non-silent p53 mutant sequences. The one or more processors may generate a wild-type p53 fitness metric based on transactivation levels of at least one p53 target gene induced by a wild-type p53 polypeptide encoded by the plurality of wild-type p53 sequences, a binding affinity of the wild-type p53 polypeptide to a promoter region of the at least one p53 target gene, and expression levels of the wild-type p53 polypeptide. The one or more processors may generate a maximum p53 fitness metric based on transactivation levels of the at least one p53 target gene caused by a complete deletion of p53 gene. The one or more processors may determine a mutant p53 fitness metric based on the wild-type p53 fitness metric, the maximum p53 fitness metric, and the mutated position of eachp53 mutant sequence of the plurality of non-silent p53 mutant sequences. The one or more processors may determine, for each p53 mutant sequence in the plurality of p53 sequences, a score indicating a degree of correlation between a RT burden and mutant p53 fitness, as a function of the RT burden metric and the mutant p53 fitness metric. The one or more processors may select, from a plurality of cancer therapies, a cancer therapy for the subject based on the score. The one or more processors may store, using one or more data structures, an association between the subject and the cancer therapy.

[0006] In some embodiments, the one or more processors may obtain a second dataset identifying Ribonucleic acid (RNA) sequences from a second biological sample of the subject. In some embodiments, the one or more processors may determine, using the second dataset, an RT RNA metric indicating expressions levels of the RT in the subject. In some embodiments, the one or more processors may determine, for each p53 mutant sequence in the plurality of p53 sequences, the score indicating the degree of correlation among at least two of the RT burden, RT RNA expression, and the mutant p53 fitness, as a function of at least two of the RT burden metric, the RT RNA metric, and the mutant p53 fitness metric.

[0007] In some embodiments, the one or more processors may determine the score indicating the degree of correlation between the RT burden and the RT RNA expression, as a function of the RT burden metric and the RT RNA expression metric. In some embodiments, the one or more processors may determine the score indicating the degree of correlation between the RT RNA expression and the mutant p53 fitness, as a function of the RT RNA expression metric and the mutant p53 fitness metric. In some embodiments, the one or more processors may determine the score indicating the correlation of the RT burden with respect to the RT RNA expression and the mutant p53 fitness, as a function of the RT burden metric, the RT RNA metric, and the mutant p53 fitness metric. In some embodiments, the one or more processors may determine the score indicating the correlation of the RT RNA expression with respect to the RT burden and the mutant p53 fitness, as a function of the RT burden metric, the RT RNA metric, and the mutant p53 fitness metric. In some embodiments, the function comprises a regression model of the RT burden metric and the mutant p53 fitness metric.

[0008] In some embodiments, the one or more processors may detect, using the first dataset, the RT activity corresponding to at least one of a long terminal repeat (LTR) retrotransposon, a non-LTR retrotransposon, or deoxyribonucleic acid (DNA) retrotransposon in the genomic sequences of the subject. In some embodiments, the one or more processors may detect, using the first dataset, the RT activity corresponding to at least one of a long interspersed nuclear element-1 (LINE-1), a short interspersed nuclear element (SINE), or a SINE- Variable number of tandem repeat (VNTR)-Alus (SVA). In some embodiments, the one or more processors may identify, using the first dataset, a plurality of RT insertion sites using soft-clipped reads and / or discordant read pairs in the plurality of sequencing reads as a signal.

[0009] In some embodiments, the one or more processors may generate the RT burden metric based on the plurality of RT insertion sites identified using soft-clipped reads and / or discordant read pairs in the plurality of sequencing reads. In some embodiments, the one or more processors may identify, using the first dataset, at each insertion site of the plurality of RT insertion sites, an orientation of a corresponding RT sequence (e.g., long interspersed nuclear element-1 (LINE-1) sequence). The orientation may include at least one of a single-stranded priming integration or an inversion-containing integration. In some embodiments, the one or more processors may remove, from the first dataset, a subset corresponding to a portion of genomic sequences based on at least one of (i) lack of matched pair, (ii) duplicity, (iii) a mapping quality score not satisfying a threshold, or (iv) inconsistent base and quality score strings (i.e., of differing lengths). In some embodiments, the one or more processors may remove, from the first dataset, a subset corresponding to a portion of genomic sequences based on a chimeric read fraction exceeding a threshold. In some embodiments, the first biological sample and / or the second biological sample comprise tissue from lung, head, neck, colon, uterus, stomach, ovary, cervix, bladder, or breast.

[0010] In any of the preceding embodiments of the methods disclosed herein, the cancer is selected from among lung cancer, brain cancer, head and neck cancer, colon cancer, rectal cancer, uterine cancer, endometrial cancer, stomach cancer, ovarian cancer, cervical cancer, bladder cancer, and breast cancer.

[0011] In any and all embodiments of the methods disclosed herein, the cancer therapy comprises one or more of immune checkpoint inhibitor therapy, adoptive cell therapy, or prophylactic cancer vaccine therapy.BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIGs. 1A-1G: Schematic of insertion of a non-LTR retrotransposon. FIG.1A: Haploid copy of the genome before retrotransposition. Purple, target site for future insertion. FIG. IB: Endonuclease breaks each strand of DNA, typically at two nearby but distinct positions. The locus between the two breakpoints is the “target site.” The relative location of the two breakpoints in this figure will result in a target site duplication. FIG. 1C: LI RNA is reverse transcribed directly into the genome resulting in the synthesis of the singlestranded cDNA starting from the 3’ end of the LI transcript and extending a variable length towards the 5’ end of the LI transcript. Teal, LI RNA. Green, reverse transcribed poly(T) cDNA. Blue, reverse transcribed LI cDNA. FIG. ID: In some cases, double priming occurs resulting in the simultaneous reverse transcription of different parts of LI transcript into the two strands of the genome. Red, reverse transcribed LI cDNA on the opposite gDNA strand. FIGs. 1E-1F: Genome after synthesis of the second strand of DNA and repair. The original synthesis strands of cDNA are always complementary to the LI transcript. Components of the LI sequence are annotated with respect to the top strand of the genome. Purple, target site which is duplicated following repair. Red, newly inserted LI sequence that was synthesized on the top strand and is therefore reverse complemented with respect to LI RNA. Blue, newly inserted LI sequence that was synthesized on the bottom strand. Green, newly inserted poly(A). Paired-end reads originating from the modified genomes are shown. FIG. IE: The process shown in FIG. 1C results in a (possibly truncated at the 5’ end) “canonical” retrotransposition. FIG. IF: The double priming process shown in FIG. ID results in an “inversion-containing” retrotransposition. The resulting genomic sequence has two LI fragments in opposite orientations. FIG. 1G: Mapping of reads A-D to the unmodified (reference) genome lacking the transposon insertion (left) and the transposon sequence (right). Left, tails of reads Al, Bl, and DI that come from the novel transposon are clipped (shown asdashed lines). Right, read C2 and the clipped tails of reads Al and DI align to the transposon sequence. The clipped tail of read B 1 contains only poly(T). In the absence of inversion as in FIG. IE, captured by read Al), the alignment between the clipped sequence and the transposon sequence reflects the length of the newly inserted transposon. When inversion occurs (as in FIG. IF, captured by read DI), such an alignment will only reflect the position where the second priming occurred.

[0013] FIGs. 2A-2D: - LI retrotransposition calls from Total ReCall throughout TCGA. N = 765 tumor normal pairs. FIG. 2A: Total number of retrotranspositions identified across 765 tumor and paired normal samples. FIG. 2B: Breakdown of total canonical or inversion-containing insertions identified within the tumor-specific somatic retrotranspositions. FIG. 2C: Estimated length of inserted LI within the canonical, tumor-specific retrotranspositions. Blue, truncated insertions. Orange, full-length insertions. FIG. 2D: Somatic tumor-specific LI retrotranspositions in each sample (“RT burden”) grouped by tumor type. Center line indicates median. Blue box indicates interquartile range. Points more than 1.5 x IQR away from the blue box are shown as individual outliers. Tumor types are sorted in descending order by median somatic RT burden.

[0014] FIGs. 3A-3C: LI RNA expression from L1EM throughout TCGA. FIG.3A: Estimated expression of intact LI RNA in each sample by tumor type. Data not shown for any tumor and sample type with fewer than 5 samples. Total N = 9,715. Blue, tumor samples. Gray, normal samples. Tumor types are sorted in descending order by median tumor-level expression. FIG. 3B: Orange, tumor types with an average of at least one retrotransposition per sample. Blue, tumor types with an average of less than one retrotransposition per sample. FIG. 3C: Expression of intact LI RNA grouped by retrotransposition count per sample. Orange, samples with any retrotransposition calls. Blue, samples with no retrotransposition calls. FIGs. 3A-3C: Center line indicates median. Blue, gray, or orange box indicates interquartile range. Points more than 1.5 x IQR away from the IQR box are shown as individual outliers. FIGs. 3B-3C: P calculated from two-sided Mann-Whitney U test.

[0015] FIGs. 4A-4B: Stratification of LI by p53 mutation. FIG. 4A: Expression of intact LI RNA in samples with mutant or wild-type p53. Expression is adjusted for intronic rate as shown in FIG. 8B. FIG. 4B: Count of somatic L I retrotranspositions in samples with mutant or wild-type p53. Somatic RT counts are adjusted for sequencing covariates as shown in FIG. 8C. FIGs. 4A-4B: Center line indicates median. Blue or orange box indicates interquartile range. Points more than 1.5 x IQR away from the IQR box are shown as individual outliers. P calculated from two-sided Mann- Whitney U test.

[0016] FIGs. 5A-5B: Mediation model relating p53 mutation, intact LI RNA expression, and somatic LI retrotransposition. FIG. 5A: p53 functional fitness score per sample by mutation category. InDei, in-frame insertion or deletion. CNA, copy number alteration. WT, wild type. Mutation categories with fewer than 10 samples are shown as individual points. Center line indicates median. Box indicates interquartile range. Points more than 1.5 x IQR away from the IQR box are shown as individual outliers. FIG. 5B: Mediation model taking p53 functional fitness as the independent variable, adjusted log2 of intact LI RNA expression as the mediating variable, and adjusted log2 of somatic LI retrotransposition count as the dependent variable. Left, schematic showing the linear regressions performed, as well as the resulting estimated weights for the mediated and unmediated pathways. 95% confidence intervals (square brackets) calculated with bootstrap resampling. Blue arrows, components of the mediated pathway. Red arrows, components of the unmediated pathway. Right, standardized fitted values for each coefficient within the mediation model and corresponding likelihood. Error bars indicate 95% confidence intervals calculated with bootstrap resampling, N= 100,000. Coefficient estimates for a, 0, T, P’, T’ estimated from ordinary least squares.

[0017] FIGs. 6A-6C: Genome in a Bottle validation of Total ReCall. FIG. 6A: Breakdown of curated LI calls identified in PacBio reads by PALMER that are perfectly recalled by Total ReCall in Illumina short-read calls. N = 193 total curated calls. Green, 137 long-read calls identified by short reads. Blue, 56 long-read calls missed by short reads. FIG. 6B: Within the long-read calls identified by short reads, breakdown of canonical and inversioncontaining insertions. FIG. 6C: Estimated length of canonical LI insertions from short-readIllumina data vs. observed insertion length in PacBio data. N = 93, R > 0.99, p < 1 x 10'10, Pearson correlation.

[0018] FIGs. 7A-7B: Adjusting tumor- and normal-specific calls per sample by sequencing depth. FIG. 7A: Left, ratio-normalizing the calls in each sample by the average read depth in that sample. Right, total of all calls per average read depth for all 765 tumor or normal samples. Y-axes in units of number of calls per read depth. N = 765 tumor and normal samples. FIG. 7B: Actual counts of calls from subset of dataset where tumor and normal sample were sequenced to comparable depth (within 20%). N = 154 tumor and normal samples. FIGs. 7A-7B: Center lines indicate median. Gray boxes indicate interquartile range. Whiskers extend to minimum and maximum values. Individual points shown for every sample. P calculated from one-sided Mann-Whitney U test. Blue, tumor samples. Orange, normal samples.

[0019] FIGs. 8A-8D: Relating LI RNA expression and LI RT. FIG. 8A: Ranking tumor types based on highest to lowest average count of somatic RTs and highest to lowest average expression of intact LI RNA. Points representing tumor types with an average of less than 1 RT per sample are not labeled, with the exception of lung adenocarcinoma (LU AD) and acute myeloid leukemia (LAML) due to their high rankings for LI RNA expression. Labels correspond to TCGA study abbreviations. N = 21 tumor types. Pearson correlation coefficient R = 0.57. FIGs. 8B-8C: Adjusted measurements based on residuals from linear regression model vs. raw measurements. FIG. 8B: Adjusted estimates of intact LI RNA expression per sample (log2 TPM) based on intronic rate of RNA-seq. N = 9,730 tumor and normal samples. FIG. 8C: Adjusted counts of somatic LI RT per sample (log2 count) based on total coverage, average base quality, read length, rate of clipped bases, and rate of chimeric alignments in both the tumor and paired normal WGS. N = 765 tumor samples. FIG. 8D: Correlation between adjusted LI RT and adjusted LI RNA per tumor sample. N = 620 tumor samples, R = 0.55, p < 1 x 10'10, Pearson correlation.

[0020] FIGs. 9A-9B: Correlation between p53 functional fitness and LI. FIG. 9A: Correlation between p53 functional fitness and expression of intact LI RNA (log2 TPM,adjusted). All tumor samples with RNA-seq and available TP53 mutation information are included; N = 8,980. R = 0.30, p < 1 x IO’10, Pearson correlation. FIG. 9B: Correlation between p53 functional fitness and count of somatic L I retrotranspositions (log2 count, adjusted). All tumor samples with WGS and available TP53 mutation information are included; N = 765. R = 0.25, p < 1 x IO’10, Pearson correlation. FIGs. 9A-9B: Dashed lines show fitted linear regression model, ordinary least squares.

[0021] FIG. 10: Correlation between LI RT and mediation model fit. N = 620, R = 0.56, p < 1 x 10'10, Pearson correlation. Dashed line, y=x.

[0022] FIG. 11: Mediation model taking TP53 alteration status as the independent variable. In contrast to FIG. 5 (which uses p53 functional fitness), adjusted log2 of intact LI RNA expression as the mediating variable, and adjusted log2 of somatic LI retrotransposition count as the dependent variable. Schematics and labels as in FIG. 5B.

[0023] FIG. 12 Estimating intact LI RNA expression from read overlaps with RepeatMasker L1HS annotations in each sample by tumor type. Sample counts, plots, and tissue type ordering are as in FIG. 3A.

[0024] FIG. 13 depicts a block diagram of a system for selecting cancer therapies based on magnitudes and frequencies of retrotransposon (RT) insertions in subject genomes in accordance with an illustrative embodiment;

[0025] FIG. 14 depicts a block diagram of a process to derive metrics to determine which cancer therapy to select in the system, in accordance with an illustrative embodiment;

[0026] FIG. 15 depict a flow diagram of a method of selecting cancer therapies based on magnitudes and frequencies of retrotransposon (RT) insertions in subject genomes in accordance with an illustrative embodiment; and

[0027] FIG. 16 depicts a block diagram of a server system and a client computer system, in accordance with one or more implementations.DETAILED DESCRIPTION

[0028] Following below are more detailed descriptions of various concepts related to, and embodiments of, systems and methods for identifying targets in x-ray projection images. It should be appreciated that various concepts introduced above and discussed in greater detail below may be implemented in any of numerous ways, as the disclosed concepts are not limited to any particular manner of implementation. Examples of specific implementations and applications are provided primarily for illustrative purposes.

[0029] Section A describes p53 regulation of LINE- 1 retrotransposition and transcription.

[0030] Section B describes systems and methods of selecting cancer therapies based on magnitude and frequencies of retrotransposon (RT) insertions in the genomes of subjects.

[0031] Section C describes a network environment and computing environment which may be useful for practicing various computing related embodiments described herein.Definitions

[0032] Unless defined otherwise, all technical and scientific terms used herein generally have the same meaning as commonly understood by one of ordinary skill in the art to which this technology belongs. As used in this specification and the appended claims, the singular forms “a”, “an” and “the” include plural referents unless the content clearly dictates otherwise.

[0033] As used herein, the term “about” in reference to a number is generally taken to include numbers that fall within a range of 1%, 5%, or 10% in either direction (greater than or less than) of the number unless otherwise stated or otherwise evident from the context (except where such number would be less than 0% or exceed 100% of a possible value).

[0034] As used herein, the “administration” of an agent or drug to a subject includes any route of introducing or delivering to a subject a compound to perform its intended function. Administration can be carried out by any suitable route, including but not limited to, orally, intranasally, parenterally (intravenously, intramuscularly, intraperitoneally, or subcutaneously),rectally, intrathecally, intratumorally or topically. Administration includes self-administration and the administration by another.

[0035] The terms “cancer” or “tumor” are used interchangeably and refer to the presence of cells possessing characteristics typical of cancer-causing cells, such as uncontrolled proliferation, immortality, metastatic potential, rapid growth and proliferation rate, and certain characteristic morphological features. Cancer cells are often in the form of a tumor, but such cells can exist alone within an animal, or can be a non-tumorigenic cancer cell. As used herein, the term “cancer” includes premalignant, as well as malignant cancers. In some embodiments, the cancer is colorectal cancer, lung cancer, breast cancer, ovarian cancer, uterine cancer, or thyroid cancer.

[0036] The terms “complementary” or “complementarity” as used herein with reference to polynucleotides (z.e., a sequence of nucleotides such as an oligonucleotide or a target nucleic acid) refer to the base-pairing rules. The complement of a nucleic acid sequence as used herein refers to an oligonucleotide which, when aligned with the nucleic acid sequence such that the 5' end of one sequence is paired with the 3’ end of the other, is in “antiparallel association.” For example, the sequence “5'-A-G-T-3”’ is complementary to the sequence “3’-T-C-A-5.” Certain bases not commonly found in naturally-occurring nucleic acids may be included in the nucleic acids described herein. These include, for example, inosine, 7-deazaguanine, Locked Nucleic Acids (LNA), and Peptide Nucleic Acids (PNA). Complementarity need not be perfect; stable duplexes may contain mismatched base pairs, degenerative, or unmatched bases. Those skilled in the art of nucleic acid technology can determine duplex stability empirically considering a number of variables including, for example, the length of the oligonucleotide, base composition and sequence of the oligonucleotide, ionic strength and incidence of mismatched base pairs. A complement sequence can also be an RNA sequence complementary to the DNA sequence or its complement sequence, and can also be a cDNA.

[0037] As used herein, a "control" is an alternative sample used in an experiment for comparison purpose. A control can be "positive" or "negative." A “control nucleic acid sample” or “reference nucleic acid sample” as used herein, refers to nucleic acid moleculesfrom a control or reference sample. In certain embodiments, the reference or control nucleic acid sample is a wild type or a non-mutated DNA or RNA sequence. In certain embodiments, the reference nucleic acid sample is purified or isolated (e.g., it is removed from its natural state). In other embodiments, the reference nucleic acid sample is from a non-tumor sample, e.g., a normal adjacent tumor (NAT), or any other non-cancerous sample from the same or a different subject.

[0038] As used herein, “discordant read pairs” refer to paired-end reads that arise from fragments spanning the inserted RT sequence that have one read map near one RT insertion site, but the other read maps to one of the many other RT sequences elsewhere in the genome (e.g., a different chromosome or a distant site on the same chromosome).

[0039] As used herein, the term “effective amount” refers to a quantity sufficient to achieve a desired therapeutic and / or prophylactic effect, e.g., an amount which results in the prevention of, or a decrease in a disease or condition described herein or one or more signs or symptoms associated with a disease or condition described herein. In the context of therapeutic or prophylactic applications, the amount of a composition administered to the subject will vary depending on the composition, the degree, 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. The skilled artisan will be able to determine appropriate dosages depending on these and other factors. The compositions can also be administered in combination with one or more additional therapeutic compounds. In the methods described herein, the therapeutic compositions may be administered to a subject having one or more signs or symptoms of a disease or condition described herein. As used herein, a "therapeutically effective amount" of a composition refers to composition levels in which the physiological effects of a disease or condition are ameliorated or eliminated. A therapeutically effective amount can be given in one or more administrations.

[0040] As used herein, “fitness” of a p53 mutation refers to the probability or propensity of a p53 mutation to be naturally selected and propagated during tumor evolution.

[0041] As used herein, “expression” includes one or more of the following: transcription of the gene into precursor mRNA; splicing and other processing of the precursor mRNA to produce mature mRNA; mRNA stability; translation of the mature mRNA into protein (including codon usage and tRNA availability); and glycosylation and / or other modifications of the translation product, if required for proper expression and function.

[0042] “Gene” as used herein refers to a DNA sequence that comprises regulatory and coding sequences necessary for the production of an RNA, which may have a non-coding function (e.g., a ribosomal or transfer RNA) or which may include a polypeptide or a polypeptide precursor. The RNA or polypeptide may be encoded by a full length coding sequence or by any portion of the coding sequence so long as the desired activity or function is retained. Although a sequence of the nucleic acids may be shown in the form of DNA, a person of ordinary skill in the art recognizes that the corresponding RNA sequence will have a similar sequence with the thymine being replaced by uracil, i.e., "T" is replaced with "U."

[0043] As used herein, the terms “individual”, “patient”, or “subject” are used interchangeably and refer to an individual organism, a vertebrate, a mammal, or a human. In a preferred embodiment, the individual, patient or subject is a human.

[0044] As used herein, a “mutation” of a gene refers to the presence of a variation within the gene or gene product that affects the expression and / or activity of the gene or gene product as compared to the normal or wild-type gene or gene product. The genetic mutation can result in changes in the quantity, structure, and / or activity of the gene or gene product in a cancer tissue or cancer cell, as compared to its quantity, structure, and / or activity, in a normal or healthy tissue or cell (e.g., a control). For example, a mutation can have an altered nucleotide sequence (e.g., a mutation), amino acid sequence, expression level, protein level, protein activity, in a cancer tissue or cancer cell, as compared to a normal, healthy tissue or cell. Exemplary mutations include, but are not limited to, point mutations (e.g., silent, missense, or nonsense), deletions, insertions, inversions, linking mutations, duplications, translocations, inter- and intra-chromosomal rearrangements. Mutations can be present in the coding or non-coding region of the gene. In certain embodiments, the mutations are associatedwith a phenotype, e.g., a cancerous phenotype (e.g., one or more of cancer risk, oncogenesis, or responsiveness to treatment). In one embodiment, the mutation is associated with one or more of: a genetic risk factor for cancer, a positive treatment response predictor, a negative treatment response predictor, a positive prognostic factor, a negative prognostic factor, or a diagnostic factor. As used herein, a “missense mutation” refers to a mutation in which a single nucleotide substitution alters the genetic code in a way that produces an amino acid that is different from the usual amino acid at that position. In some embodiments, missense mutations alter one or more functions or physical -chemi cal properties of the encoded protein.

[0045] As used herein, “silent mutations” are mutations, or changes, to a gene’s DNA sequence that have no effect on the amino acid sequence encoded by the gene. In contrast, “non-silent mutations” alter the resulting amino acid sequence encoded by the gene and include missense mutations which change an amino acid to another amino acid, nonsense mutations which change an amino acid to a stop codon, mutations in splice sites and insertions or deletions (indels). The indels can also be divided into two types namely, inframe indels and frameshift indels according to whether it changes the reading frame.

[0046] The term “promoter region” of a target gene as used herein refers to a segment of the target gene to which RNA polymerase can bind to and initiate transcription of the target gene. In some embodiments, the promoter region may include the first 250 nucleotides (nt), first 300 nt, first 350 nt, first 400 nt, first 450 nt, first 500 nt, first 1 kb, first 5 kb, first 10 kb, first 15 kb, first 20 kb, first 21 kb or first 22 kb of genomic sequence directly upstream of the translation start site of target gene.

[0047] As used herein, a “Retrotransposon” or “RT” refers to an evolutionarily widespread genetic element that replicates through reverse transcription of an RNA copy and integrates the product DNA into new sites in a host genome. Retrotransposons comprise significant fractions of metazoan genomes. RTs differ from DNA transposons. DNA transposons move using a cut-and-paste mechanism. In contrast, retrotransposons move in a copy-and-paste fashion by duplicating the element into a new genomic location via an RNAintermediate. Thus, retrotransposons increase their copy number more rapidly than DNA transposons.

[0048] As used herein, “RT burden” refers to somatic tumor-specific retrotranspositions of a transposable repeat element (e.g., LI) in a biological sample. In some embodiments, RT burden varies by tumor type.

[0049] As used herein, “RT activity” refers to the somatic retrotransposition activity (e.g., insertions) of transposable mobile elements (known as retrotransposons) within the genome of a host subject. In some embodiments, RT activity comprises “canonical” retrotransposition insertions, where a single stretch of a RT starting from its 3’ end is inserted in the genome of a host subject. In certain embodiments, RT activity comprises “inversion-containing” retrotransposition insertions, where a genomic insertion that contains part of an RT sequence is adjacent to an inverted part of the RT sequence. In other embodiments, RT activity comprises “3’ transductions” where during RT RNA expression, a RNA polymerase overruns the poly(A) signal at the 3’ end of the RT element resulting in transcription and retrotransposition of some host subject genomic sequence downstream of the 3’ end of the RT in addition to the RT itself.

[0050] As used herein, a “sample” refers to a substance that is being assayed for the presence of a mutation in a nucleic acid of interest. Processing methods to release or otherwise make available a nucleic acid for detection may include steps of nucleic acid manipulation. A biological sample may be a body fluid or a tissue sample. In some cases, a biological sample may consist of or comprise blood, plasma, sera, urine, feces, epidermal sample, vaginal sample, skin sample, cheek swab, sperm, amniotic fluid, cultured cells, bone marrow sample, tumor biopsies, aspirate and / or chorionic villi, cultured cells, and the like. Fresh, fixed or frozen tissues may also be used. In one embodiment, the sample is preserved as a frozen sample or as formaldehyde- or paraformaldehyde-fixed paraffin-embedded (FFPE) tissue preparation. For example, the sample can be embedded in a matrix, e.g., an FFPE block or a frozen sample. Whole blood samples of about 0.5 to 5 ml collected with EDTA, ACD or heparin as anticoagulant are suitable.

[0051] As used herein, “soft-clipping” refers to a sequence alignment technique where paired-end sequences are aligned to a reference genome using BWA-MEM v0.7.15-rl 14O40with options “-K 100000000 -p -v 3 -t 16 -Y -T 0”, where the “-Y” option retains the clipped read sequences for supplementary alignments. As described herein, reads that span the RT insertion site breakpoint result in chimeric reads that contain portions of both the insertion site sequence from the human genome as well as the inserted RT sequence, and lead to “soft clipped” alignments. Supplementary alignments are relevant when the longer part of the read spanning the breakpoint contains the RT sequence (and whose primary mapping is thus to RT genomic sequence elsewhere in the host genome). The use of such supplementary alignments is useful because they provide a longer clipped sequence, which can in turn be aligned to the RT sequence more specifically.

[0052] As used herein, the terms “target gene”, “target sequence” and “target nucleic acid sequence” refer to a specific nucleic acid sequence to be detected and / or quantified in the sample to be analyzed.

[0053] “Treating” or “treatment” as used herein covers the treatment of a disease or disorder described herein, in a subject, such as a human, and includes: (i) inhibiting a disease or disorder, i.e., arresting its development; (ii) relieving a disease or disorder, i.e., causing regression of the disorder; (iii) slowing progression of the disorder; and / or (iv) inhibiting, relieving, or slowing progression of one or more symptoms of the disease or disorder. In some embodiments, treatment means that the symptoms associated with the disease are, e.g., alleviated, reduced, cured, or placed in a state of remission.

[0054] It is also to be appreciated that the various modes of treatment of disorders as described herein are intended to mean “substantial,” which includes total but also less than total treatment, and wherein some biologically or medically relevant result is achieved. The treatment may be a continuous prolonged treatment for a chronic disease or a single, or few time administrations for the treatment of an acute condition.A. p53 regulation of LINE 1 retrotransposition and transcription

[0055] Somatic activity of LINE-1 (LI) mobile elements has been implicated in cancer etiology, which may be related to the loss of p53-mediated regulation as a result of TP53 mutations. Quantifying the mechanisms of L I regulation in cancer has been challenging. Here, a statistical model of LI regulation was built by simultaneously quantifying LI retrotransposition, LI expression, and the fitness costs of mutated TP53 with precision. Total ReCall was first developed, an algorithm specifically tailored to the mechanisms of LI reintegration, to detect LI insertions from short-read whole-genome sequencing. Applying Total ReCall to high-quality data consisting of >750 paired tumor and normal samples from The Cancer Genome Atlas (TCGA) shows high LI insertion heterogeneity among tumor types, with increased retrotransposition burden in lung squamous cell carcinoma, head and neck, and colon cancers. Next, the active RNA expression of intact LI in >9,000 TCGA tumor samples was assessed, establishing, for the first time, a clear correlation between LI expression and retrotransposition. Finally, the number of LI insertions, LI expression and a mathematical model of TP53 fitness were integrated into a multi-modal model of p53- mediated mechanisms of LI regulation. It shows that TP53 mutations enable retrotransposition both by disinhibiting LI expression and enabling its reintegration and quantify the relative weights of this dual regulatory role. It is then demonstrated how mechanism-based multi-modal modeling applied at scale can statistically disentangle the complex interplay between canonical driver events in tumor evolution and retrotransposon activity.

[0056] While methods have been developed for detecting somatic retrotransposition events14,11,15a first principles approach was set out to build accounts for the specific retrotransposition cycle of LI (FIG. 1). The resulting “Total ReCall” method relies on two key signals that arise when aligning sequencing reads to the reference genome (FIG. 1G): (i) reads that span the retrotransposon insertion site breakpoint will result in chimeric reads which contain portions of both the insertion site sequence from the human genome as well as the inserted LI sequence, and lead to “soft clipped” alignments, and (ii) paired-end reads that arise from fragments spanning the inserted LI sequence will often have one read map near the insertion site, but the other read map to one of the many LI sequences elsewhere in thegenome, even to a different chromosome or a distant site on the same chromosome, resulting in “discordant read pairs”.

[0057] Total ReCall was applied to assess somatic LI retrotransposition prevalence in a pan-cancer cohort of 765 paired tumor-normal samples across 22 tissue types from The Cancer Genome Atlas (TCGA), assessing heterogeneity across types. To achieve a clear picture of the LI life cycle in cancer, LI RNA was aimed to accurately quantify expression simultaneously with retrotransposition, since RNA not only encodes the protein machinery for retrotransposition but also acts as the substrate for new genomic copies of LI. To do so, The L1EM mechanistic model of Ll-driven transcription (“active expression”) was used at intact LI copies in the genome16. Using this, the heterogeneity of active expression of LI RNA was assessed in 9,011 tumor samples across 32 tissue types, and similar patterns were found between the RNA levels of L I and the abundance of its retrotransposition. TP53, the most frequently mutated gene in cancers17, encodes a protein that functions to maintain genome stability, including suppression of LI retrotransposition18. Studies involving genetic manipulation of TP53 in human cell lines and animal models have demonstrated that loss of p53 function leads to derepression of retroelements as seen with common TP53 cancer mutations19’20. Combined LI insertional and transcriptional data was integrated with TP 53 mutational status and fitness21, which allowed us to model this critical regulatory mechanism at large scale. P53 was found to play a dual role in restraining retrotransposition, both by repression of LI transcription and by regulation of its integration into the genome.

[0058] The study substantiates with quantitative statistical evidence that a regulatory relationship exists between p53 and LI throughout human cancers. A theoretical fitness model was employed to evaluate the extent to which mutation has altered p53 function in each tumor context, but the significance of the statistical mediation model is robust to the method for quantifying mutational status. Both the mediated and unmediated modulation of LI retrotransposition by p53 remain significant when the model is evaluated using a binary mutated category (i.e., in every sample, p53 is either altered (1) or not (0)) as defined by cBioPortal (FIG. 11)25,26. While there are limitations when using observational data, such as cancer types examined, causal assumptions of the directionality of mediation, and survivorshipbias of cancer cells with mutated p53 under LI overexpression28, the large pan-cancer study is well-powered and is consistent with previous studies using human cell lines as model systems19,29. The analysis suggests that p53 transrepression of L I through direct binding of the genomic LI 5’ UTR is likely to be somatically active across tissues and significantly mutagenically disrupted in cancers.

[0059] Many studies have noted that the role of p53 as a general regulator of cell cycle control, apoptosis, and senescence is insufficient to explain the extent of tumor suppression by p5330,31,32. In a 2018 review, Tiwari et al. hypothesized that p53 may also be suppressing tumor formation through its role in directly restricting expression of transposons27. The theory here is that various cellular stress response pathways may be activated by p53 in response to retrotransposition, and in addition p53 regulates LI via different mechanisms, likely by acting on the LI promoter that controls LI RNA expression. Using the mechanistic models of stimulus-dependent regulation and tonic regulation27as a framework for interpretation, the results suggest that LI modulation in cancer by p53 uses both modes of regulation and occurs at both the transcriptional and retrotransposition levels. Previous studies have demonstrated mechanisms by which p53 can directly transrepress LI, by binding its 5’ UTR RNA polymerase II promoter and facilitating the deposition of histone marks19,20. The finding of a significant correlation between p53 mutational fitness and LI RNA expression across a dataset of clinical tumor samples is consistent with this transrepression being somatically active and relevant to cancers. Repression of the LI lifecycle at the transcriptional stage will logically preclude retrotransposition, as is represented by the RNA-mediated path of the statistical model. The non-RNA-mediated effect of p53 mutation modulating LI retrotransposition may be imposed by negative selection against reintegrations28or their consequences for gene expression33,34,35following stimulus-dependent regulation by wild-type p53 (e.g., cell cycle arrest), or via a still undescribed mechanism.

[0060] The present disclosure provides the first pan-cancer analysis of the expression levels of active LI mRNA. Of note, a previous study observed relationships between LI expression and DNA damage and replication stress, suggesting that the oncogenicity of LI derepression is in excess of what can be attributed to completed cycles of retrotransposition thatdisrupt gene expression36, and therefore LI RNA may itself be oncogenic through some mechanisms. For example, abundance of LI transcripts can play a role in heterochromatin erosion37. Additionally, the presence of cytosolic L I RNA / cDNA hybrids resulting from reverse transcription could lead to an inflammation response that can alter the tumor microenvironment by activating the cGAS innate immune receptor, which in turn activates the interferon pathway38. This analysis uncovers tissue- and tumor-type dependence of LI expression at the RNA level and informs understanding of LI behavior and activity in these different contexts. For instance, a previous study investigating LI ORFlp binding in prostate cancer suggested the possibility that cytoplasmic ORFlp may affect RNA processing after finding many non-Ll transcripts associated with ORFlp, including in particular RNAs that were also enriched at p-bodies38. Although that study focused on prostate cancer, the analysis did not find significant activity of LI retrotransposition in prostate tumors (only one somatic insertion was identified out of 15 tumor samples profiled), nor did intact LI RNA expression in prostate tissues vary significantly between tumor and normal samples.

[0061] In an earlier study (on a smaller cancer dataset), LI ORFlp positively correlated with copy number alteration in breast, ovarian, and endometrial tumors, especially when p53 was mutated36. The observations provide additional evidence into the diverse regulatory roles of p53 that influence LI RT burden in cancer. Using state-of-the-art multi-modal modeling of LI retrotranspositions (Total ReCall), LI RNA expression (L1EM), as well as p53 mutational fitness, we are able to leverage statistical modeling of causality to uncover associations that were previously masked by experimental and biological heterogeneity. Having robust tools to measure LI RNA expression, LI retrotransposition, and p53 mutational fitness is useful in cancer risk stratification and advance personalized cancer medicine.B. Systems and Methods for Selecting Cancer Therapies Based on Magnitudes and Frequencies of Retrotransposon (RT) Insertions in Subject Genomes

[0062] Referring now to FIG. 13, depicted is a block diagram of a system 100 for selecting cancer therapies based on magnitudes and frequencies of retrotransposon (RT) insertions in subject genomes. In brief overview, the system 100 may include at least one dataprocessing system 105, at least one sequencing device (e.g., a next-generation sequencer) 110, and at least one administrative device 115, communicatively coupled with one another via at least one network 120. The data processing system 105 may include at least one dataset aggregator 125, at least one genome analyzer 130, at least one fitness analyzer 135, at least one ribonucleic acid (RNA) analyzer 140, at least one function evaluator 145, at least one output handler 150, at least one correlation function 155 (sometimes herein referred to as a correlation model or mediation model), and at least one database 160, among others. Each of the components in the system 100 as detailed herein may be implemented using hardware (e.g., one or more processors coupled with memory), or a combination of hardware and software as detailed herein in Section C. Each of the components in the system 100 may implement or execute the functionalities detailed herein, such as those described in Section A.

[0063] In further detail, the data processing system 105 may (sometimes herein generally referred to as a computing system or a server) be any computing device including one or more processors coupled with memory and software and capable of performing the various processes and tasks described herein. The data processing system 105 can be in communication with the sequencing device 110, the administrative device 115, the database 160, and other devices, via the network 120. The data processing system 105 may be situated, located, or otherwise associated with at least one server group. The server group may correspond to a data center, a branch office, or a site at which one or more servers corresponding to the data processing system 105 is situated. The data processing system 105 can perform or implement any of the functionalities detailed herein in conjunction with Section A.|0064] In the data processing system 105, the dataset aggregator 125 may acquire or obtain genome sequencing datasets or RNA sequencing datasets from the sequencing device 110. The genome analyzer 130 may process the genome sequencing datasets to detect RT insertion sites and determine RT burden metrics. The fitness analyzer 135 may process the genome sequencing datasets to identify p53 genes and to determine p53 mutant fitness metrics. The RNA analyzer 140 may process the RNA sequencing datasets to determine RT RNA metrics. The function evaluator 145 may use the correlation function 155 to generate a scoreindicating degree of correlation between at least two of the RT burden, p53 mutant, and RT RNA, among others. The correlation function 155 may include any type of function or model to determine the score, such as a regression model (e.g., a linear, non-linear, or logarithmic regression model), a correlation function (e.g., Pearson correlation coefficient, Spearman’s rank correlation coefficient, or Kendall’s Tau), a component analysis model (e.g., principal component analysis (PCA), directional component analysis (DCA), or factor analysis), or statistical model (e.g., analysis of variance (ANOVA)), among others. The output handler 150 may generate instructions to select cancer therapy based on the score indicating the degree of correlation.

[0065] The sequencing device 110 may be any device to perform genome sequencing or RNA sequencing on samples taken from subjects. The sequencing device 110 may be communicatively coupled with the data processing system 105 and the administrative device 115 via the network 120. The sequencing device 110 may be associated with an entity examining subjects or facilitating the derivation of genome and RNA sequencing from such subject. The sequencing device 110 may perform genetic sequencing on a deoxyribonucleic acid (DNA) sample taken from a subject to generate genome sequencing datasets. The genetic sequencing carried out may be a high throughput, massively parallel sequencing technique (sometimes herein referred to as next -gen eration sequencing), such as short-read or long-read whole genome sequencing (WGS), pyrosequencing, Reversible dye-terminator sequencing, SOLiD sequencing, Ion semiconductor sequencing, Helioscope single molecule sequencing, among others. Thus, the genomic sequences and RNA sequences obtained via next-generation sequencing comprise a plurality of sequencing reads.|0066] In addition, the sequencing device 110 may perform RNA sequencing on an RNA sample to generate RNA sequencing datasets. The RNA sequencing performed may include, for example, Poly(A) Selection RNA-seq, Total RNA-seq, strand-specific RNA-seq, single-cell RNA-seq, Long-read RNA-seq, small RNA-seq, or RNA capture sequencing, among others. The genetic sequencing may be targeted to find retrotransposon insertion sites in the genomes of subjects. The genome sequencing dataset may include long reads (e.g., greater than 1000 base pairs) or short reads (e.g., between 50 to 300 base pairs). The RNA sequencingdataset may include long reads (e.g., greater than 1000 base pairs) or short reads (e.g., between 50 to 300 base pairs). The gene sequencing and RNA sequence may be maintained using one or more files according to a format (e.g., FASTQ, BCL, or VCF formats).

[0067] The administrative device 115 (sometimes herein referred to as an end user computing device) may be any computing device comprising one or more processors coupled with memory and software and capable of performing the various processes and tasks described herein. The administrative device 115 may be in communication with the data processing system 105 and the sequencing device 110 via the network 120. The administrative device 115 may have at least one display. The administrative device 115 may be associated with an entity (e.g., a clinician) examining the subject or genomic or RNA data from the subject. The display may present information about the subject provided by the data processing system 105.

[0068] The database 160 may store and maintain various resources and data associated with the data processing system 105, the sequencing device 110, and the administrative device 115, among others. The database 160 may include a database management system (DBMS) to arrange and organize the data maintained thereon. The database 160 may be in communication with the data processing system 105, the sequencing device 110, and the administrative device 115, via the network 120. While running various operations, the data processing system 105, the sequencing device 110, and the administrative device 115 may access the database 160 to retrieve identified data therefrom. The data processing system 105, the sequencing device 110, and the administrative device 115 may also write data onto the database 160 from running such operations.

[0069] Referring now to FIG. 14, depicted is a block diagram of a process 200 to derive metrics to determine which cancer therapy to select in the system 100. The process 200 may include or correspond to operations performed in the system 100 to obtain genome and RNA data from a subject to derive metrics with which to select a cancer therapy for the subject. Under the process 200, the sequencing device 110 may execute, carry out, or otherwise perform genome sequencing and RNA sequencing on one or more biological samples obtained from at least one subject 205. The subject 205 may at risk of or diagnosed with cancer. The biologicalsamples may be obtained from any tissue or organ of the subject 205, such as lung, head, neck, colon, rectum, uterus, endometrium, stomach, ovary, cervix, bladder, or breast, among others. The cancer may include, for example, lung cancer, brain cancer, head and neck cancer, colon cancer, rectal cancer, uterine cancer, endometrial cancer, stomach cancer, ovarian cancer, cervical cancer, bladder cancer, or breast cancer, among others. In some embodiments, the one or more biological samples comprise the same tissue organ type.

[0070] The sequencing device 110 may perform genome sequencing on at least one biological sample of the subject 205 to output, produce, or otherwise generate at least one genome sequence dataset 210. The biological sample may be obtained from the organ affected by the cancer. The genome sequence dataset 210 may include or identify a set of genomic sequences from the biological sample of the subject 205. One or more portions of the genomic sequences may include or identify reads corresponding to RT insertion sites. One or more portions of the genomic sequences may include or identify p53 genes. The genome sequence dataset 210 may include or identify metadata associated with the set of genomic sequences, such as an identifier (e.g., an anonymized identifier) for the subject 205. With the generation, the sequencing device 110 may provide, send, or otherwise transmit the genome sequence dataset 210 to the data processing system 105.

[0071] In addition, the sequencing device 110 may perform RNA sequencing on at least one biological sample of the subject 205 to output, produce, or otherwise generate at least one RNA sequence dataset 215. The RNA sequence dataset 215 may include a set of RNA sequences from the biological sample of the subject 205. One or more portions of the set of RNA sequences may include or identify transcriptions of RT expressions. In some embodiments, the biological sample for the RNA sequencing may differ from the biological sample for the genome sequencing. In some embodiments, the biological sample for the RNA may be from the same organ as the organ of the biological sample for genome sequencing. For instance, two biological samples may be taken from the organ affected by the cancer. One biological sample may be used for the genome sequencing and the other biological sample may be used for the RNA sequencing. The RNA sequence dataset 215 may include or identify metadata associated with the set of RNA sequences, such as an identifier (e.g., an anonymizedidentifier) for the subject 205. With the generation, the sequencing device 110 may provide, send, or otherwise transmit the RNA sequence dataset 215 to the data processing system 105.

[0072] The dataset aggregator 125 executing on the data processing system 105 may receive, identify, or otherwise obtain the genome sequence dataset 210 from the sequencing device 110. The genome sequence dataset 210 may identify or include the set of genomic sequences of the biological sample from the subject 205. With the obtaining of the genome sequence dataset 210, the dataset aggregator 125 may perform filtering (e.g., for quality control) on the genomic sequences identified in the genome sequence dataset 210. The filers may be for quality control data filtering as detailed herein in Section A. In filtering, the dataset aggregator 125 may calculate, determine, or otherwise generate a quality score for the reads of the genome sequence dataset 205 based on one or more of: a total number of reads, average base quality, average read length, or fractions of reads with split or discordant alignments, among others. The dataset aggregator 125 may filter out at least a subset of the genome sequence dataset 210 based on the quality score of the corresponding portion of genomic sequences.

[0073] In some embodiments, the dataset aggregator 125 may delete, erase, or otherwise remove at least a subset of the genome sequence dataset 210 corresponding to a portion of the genomic sequences based on one or more of: (i) a lack of a matched pair (e.g., reads lacking a matched pair), (ii) duplicity (e.g., reads marked as duplicates), (iii) a quality score not satisfying a threshold (e.g., a mapping quality of reads), or (iv) inconsistent base and quality score strings (i.e., of differing lengths), among others. In some embodiments, the dataset aggregator 125 may delete, erase, or otherwise remove at least a subset of the genome sequence dataset 210 corresponding to a portion of the genomic sequences based on a chimeric read fraction of the portion exceeding a threshold. The fraction (or percentage) may correspond to the portion of the genomic sequences with chimeric or split reads (e.g., multiple sections of reads aligned with different positions in a reference genome). The filtered genome sequence dataset 210 may be used for additional processing by the data processing system 105.

[0074] In addition, the dataset aggregator 125 may receive, identify, or otherwise obtain at least one RNA sequence dataset 215 from the sequencing device 110. The RNA sequence dataset 215 may identify or include the set of RNA sequences of the biological sample from the subject 205. With the obtaining of the RNA sequence dataset 215, the dataset aggregator 125 may perform filtering (e g., for quality control) on the genomic sequences identified in the RNA sequence dataset 215. The filers may be for quality control data filtering as detailed herein in Section A. In some embodiments, the dataset aggregator 125 may delete, erase, or otherwise remove at least a subset of the RNA sequence dataset 215 corresponding to a portion of the RNA sequences with single-end reads. In some embodiments, the dataset aggregator 125 may remove at least a subset of the RNA sequence dataset 215 corresponding to a portion of the RNA sequences based on metastatic, recurrent, or new primary tumor samples, among others. The filtered RNA sequence dataset 215 may be used for additional processing by the data processing system 105.

[0075] The genome analyzer 130 executing on the data processing system 105 may calculate, determine, or otherwise identify at least one RT burden 220 metric using the genome sequence dataset 210. The RT burden metric may identify or correspond to a level of RT activity in the set of genomic sequences in the biological sample from the subject 205. The RT activity may comprise canonical insertions, inversion-containing insertions, 5’ truncated insertions, full-length insertions, and 3’ transductions of an RT. The RT activity may include or correspond to, for example, one or more of a long interspersed nuclear element- 1 (LINE-1), a short interspersed nuclear element (SINE), or a SINE- Variable number of tandem repeat (VNTR)-Alus (SVA) within the genomic sequences of the host subject. To identify the RT burden metric, the genome analyzer 130 may process or parse the set of genomic sequences identified by the genome sequence dataset 210 (e.g., the filtered genome sequence set 210). From parsing, the genome analyzer 130 may detect or identify a plurality of RT insertion sites using soft-clipped reads and / or discordant read pairs in the plurality of sequencing reads as a signal. The insertion sites may correspond to the RT activity (RT burden) within the genomic sequences of the subject.

[0076] Using the soft-clipped reads or discordant reads, the genome analyzer 130 may identify or otherwise detect at least one of a long terminal repeat (LTR) retrotransposon (also referred herein as class I transposable LTR element), a non-LTR retrotransposon (e.g., also referred herein as class I non-LTR element), or DNA retrotransposon (e.g., also referred herein as class II element) in the set of genomic sequences of the subject. The occurrences of the LTR retrotransposon, non-LTR retrotransposon, or DNA retrotransposon may correspond to RT activity. In some embodiments, the genome analyzer 130 may identify or detect insertions of at least one of LINE- 1, SINE, or SVA as the RT activity in the set of genomic sequences of the subject. In some embodiments, the genome analyzer 130 may determine or identify an orientation of the RT (e.g., LINE-1) sequence for each insertion site detected in the genomic sequences of the subject. The orientation may include or correspond to at least one of singlestranded priming or an inversion-containing integration, among others.

[0077] Based on the RT activity detected in the genomic sequences of the subject, the genome analyzer 130 may calculate, determine, or otherwise generate the RT burden metric 220. The RT burden metric 220 may be a numerical value indicating the level of RT activity within the genomic sequences of the biological sample. In general, the higher the RT activity, the higher the RT burden metric 220 may be; conversely, the lower the RT activity, the lower the RT burden metric 220 may be. In some embodiments, the genome analyzer 130 may generate the RT burden metric 220 based on the number of insertion sites detected in the set of genomic sequences from the genome sequence dataset 210. In some embodiments, the genome analyzer 130 may determine the RT metric 220 based on a number of detected occurrences of the LTR retrotransposon, non-LTR retrotransposon, or DNA retrotransposon in the genomic sequences of the subject. In some embodiments, the genome analyzer 130 may determine the RT metric 220 based on a number of detected insertions of LINE- 1, SINE, or SVA in the genomic sequences of the subject.

[0078] The fitness analyzer 135 executing on the data processing system 105 may select, detect, or otherwise identify a set of p53 sequences in the genomic sequences of the subject using the genome sequence dataset 210. Each p53 sequence may correspond to or include at least one of a set of non-silent p53 mutant sequences or a set of wild-type p53sequences. The non-silent p53 mutant sequences may correspond to p53 somatic mutations, missense mutations, nonsense mutations, frameshift mutations copy number variations, structural variations (e.g., indels), translocations, or any other alterations among others. The wild-type p53 sequences may correspond to p53 genes without any such mutations, silent p53 mutant sequences or p53 mutant sequences in non-coding regions. The fitness analyzer 135 may parse or process the genomic sequences of the genome sequence dataset 210 to identify the p53 sequences. From the set of p53 sequences, the fitness analyzer 135 may categorize, classify, or otherwise select a subset of the sequences as part of the set of non-silent p53 mutant sequences. The fitness analyzer 135 may select another subset of sequences as part of the wildtype p53 sequences. In some embodiments, silent p53 mutant sequences or p53 mutant sequences in non-coding regions are selected as part of the set of wild-type p53 sequences.

[0079] For each identified p53 sequence, the fitness analyzer 135 may calculate, generate, or otherwise determine at least one mutant p53 metric 225. To determine, the fitness analyzer 135 may determine or identify a mutated position of each p53 mutant sequence within the p53 sequence using the genome sequence dataset 210. The mutation may include, for example, a nonsense or frameshift mutation, among others, of the p53 gene. The position may be a numeric value corresponding to a location within the p53 gene sequence with the mutation. For example, the position may be an integer between 1 to 393, inclusive, as the p53 protein is 393 amino acids long, corresponding to the location of the first altered residue along the protein. In addition, the fitness analyzer 135 may calculate, generate, or otherwise determine a maximum p53 fitness metric. The maximum p53 fitness metric may correspond to transaction levels of at least one p53 target gene (e.g., WAF1, MDM2, BAX, hl433s, AIP1, GADD45, NOXA, and P53R) caused by a complete deletion of a p53 gene from the p53 gene sequence.

[0080] Furthermore, the fitness analyzer 135 may calculate, generate, or otherwise determine a wild-type p53 fitness metric. The wild-type p53 fitness metric may identify or correspond to a level of wild-type p53 concentrations and wild-type p53 affinities to DNA promoter regions of one or more p53 target genes. The determination of the wild-type p53 fitness metric may be based on one or more of transactivation levels of at least one p53 targetgene (e.g., WAF1, MDM2, BAX, hl433s, AIP1, GADD45, NOXA, and P53R) induced by a wild-type p53 polypeptide encoded by the set of wild-type p53 sequences, a binding affinity of the wild-type p53 polypeptide to a promoter region of the at least one p53 target gene, and expression levels of the wild-type p53 polypeptide, among others. The binding affinity may correspond to strength of interaction of the wild-type p53 polypeptide to the promoter region of the p53 target gene. The expression level may correspond to a quantity of the wild-type p53 polypeptide in the p53 sequence.

[0081] Based on the mutated position of each p53 mutant sequence, the wild-type p53 fitness metric, and the maximum p53 metric, the fitness analyzer 135 may calculate, generate, or otherwise determine the mutant p53 metric 225. The combination of the mutated position of each p53 mutant sequence, the wild-type p53 fitness metric, and the maximum p53 metric to determine the mutant p53 metric 225 may be in accordance with the following:where FMTis the mutant p53 metric 225, FWTis the fitness metric of wild-type p53, Fmaxis the maximum p53 fitness metric, and p is the position of the p53 mutant sequence (e.g., an integer value between 1 and 393, inclusive). The mutant p53 metric 225 may identify or correspond to an oncogenic advantage of having a p53 mutation by the probability of mutant p53 not binding to transcriptional targets, factoring in the presence of wild-type p53 in the p53 sequences. In general, the higher the p53 metric 225, the more likely the mutant p53 is to not bind to target sites and by extension a higher loss of DNA-binding abilities. Conversely, the lower the p53 metric 22, the less likely the mutant p53 is to not bind to target sites and by extension a lower loss of DNA-binding abilities. The fitness analyzer 135 may repeat the operations detailed herein with respect to each identified p53 sequence of the set of genomic sequences in the genome sequence dataset 210.

[0082] The RNA analyzer 140 executing on the data processing system 105 may calculate, identify, or otherwise determine at least one RT RNA metric 230 using the RNA sequencing dataset 215. The RT RNA metric 230 may correspond to or indicate expressionlevels of the RT in the biological sample from the subject t205 for the RNA sequence dataset 215. To determine, the RNA analyzer 140 may process or parse the RNA sequences of the RNA sequence dataset 215. From parsing, the RNA analyzer 140 may identify or detect expression levels of an RT in the RNA sequences. The RT expression levels may correspond to the transcription of LINE- 1, SINE, or SVA elements, among others, in the RNA sequences. In some embodiments, the RNA analyzer 140 may identify or select unique instances of the RT expressions in the RNA sequences of the RNA sequencing dataset 215. Based on the number of detected occurrences of RT expressions, the RNA analyzer 140 may determine the RT RNA metric 230.

[0083] The function evaluator 145 executing on the data processing system 105 may calculate, generate, or otherwise determine at least one correlation score 235 among at least two of the RT burden, the mutant p53 fitness, or the RT RNA expression, in accordance with the correlation function 155. The correlation score 235 may indicate a degree of correlation, mediation, or effect of one metric to at least one other metric (e.g., the RT burden metric 220, the mutant p53 fitness metric 225, or the RT RNA metric 230). The correlation function 155 may include any function, algorithm, or model (e.g., regression model, correlation function, component analysis, or partition of variance) used to determine the correlation score 235. For instance, the correlation function 155 may be a linear regression model (e.g., as described in Section A) in which at least one metric is an input (or independent variable) and at least one other metric is an output (or dependent variable). The linear regression model may also include a coefficient defining the relationship between the input and output, and the coefficient may be determined as the correlation score 235. In some embodiment, the function evaluator 145 may determine the correlation score 235 to indicate the degree of correlation among at least two of the RT burden, the mutant p53 fitness, or the RT RNA, using the correlation function 155 with at least two of the RT burden metric 220, the mutant p53 fitness metric 225, or the RT RNA metric 230.

[0084] In some embodiments, the function evaluator 145 may determine the correlation score 235 to indicate the degree of correlation between the RT burden and the mutant p53 fitness in accordance with the correlation function 155. The correlation score 235 may bedetermined by the function evaluator 145 for each p53 mutant sequence in the set of p53 sequences from the genomic sequences of the genome sequencing dataset 210. The correlation function 155 may include the mutant p53 fitness metric 225 as input and the RT burden metric 220 as the output. The function evaluator 145 may determine the correlation score 235 defining the relationship between the RT burden metric 220 and the mutant p53 fitness metric 225 using the correlation function 155. The correlation score 235 may indicate the level of mediation or effect of the mutant p53 fitness metric 225 on the RT burden metric 220, or vice- versa.

[0085] In some embodiments, the function evaluator 145 may determine the correlation score 235 to indicate the degree of correlation between the RT burden and the RT RNA expression in accordance with the correlation function 155. The correlation score 235 may be determined by the function evaluator 145 for each genome sequence identified from genome sequencing dataset 210. The correlation function 155 may include the RT burden metric 220 as the output and the RT RNA metric 230 as the input. The function evaluator 145 may determine the correlation score 235 defining the relationship between the RT burden metric 220 and the RT RNA metric 230 using the correlation function 155. The correlation score 235 may indicate the level of correlation of the RT RNA metric 230 on the RT burden metric 220, or vice-versa.

[0086] In some embodiments, the function evaluator 145 may determine the correlation score 235 to indicate the degree of correlation between the RT RNA expression and the mutant p53 fitness in accordance with the correlation function 155. The correlation score 235 may be determined by the function evaluator 145 for each p53 sequence identified from genome sequencing dataset 210. The correlation function 155 may include the RT RNA metric 230 as the output and the mutant p53 fitness metric 220 as the input. The function evaluator 145 may determine the correlation score 235 defining the relationship between the RT RNA metric 230 and the mutant p53 fitness metric 220 using the correlation function 155. The correlation score 235 may indicate the level of correlation of the p53 mutant fitness metric 230 on the RT RNA metric 225, or vice-versa.

[0087] In some embodiments, the function evaluator 145 may determine the correlation score 235 to indicate the degree of correlation of the RT burden with respect to the RT RNA expression and the mutant p53 fitness. The correlation score 235 may be determined by the function evaluator 145 for each p53 sequence identified from genome sequencing dataset 210 or each genome sequence of the genome sequencing dataset 210. The correlation function 155 may include the RT burden metric 225 as output, and the mutant p53 fitness metric 230 and the RT RNA metric 235 as the inputs. The function evaluator 145 may determine the correlation score 235 defining the relationship of the RT burden with respect to the RT RNA expression and the mutant p53 fitness using the correlation function 155. The correlation score 235 may indicate the magnitude of mediation or effect of the RT RNA metric 230 and the mutant p53 fitness metric 225 on the RT burden metric 220.

[0088] In some embodiments, the function evaluator 145 may determine the correlation score 235 to indicate the degree of correlation of the RT RNA expression with respect to the RT burden and the mutant p53 fitness. The correlation score 235 may be determined by the function evaluator 145 for each p53 sequence identified from genome sequencing dataset 210 or each genome sequence of the genome sequencing dataset 210. The correlation function 155 may include the RT RNA metric 235 as output, and the mutant p53 fitness metric 230 and the RT burden metric 225 as the inputs. The function evaluator 145 may determine the correlation score 235 defining the relationship of the RT RNA expression with respect to the RT burden and the mutant p53 fitness using the correlation function 155. The correlation score 235 may indicate the magnitude of mediation or effect of the RT burden metric 220 and the mutant p53 fitness metric 225 on the RT RNA metric 230.|0089] The output handler 150 executing on the data processing system 105 may identify, determine, or otherwise select at least one cancer therapy from a set of candidate cancer therapies based on the correlation score 235. The correlation score 235 may be used to determine whether the subject 205 will benefit from a particular cancer therapy. The set of candidate cancer therapies may include, for example, immune checkpoint inhibitor therapy, adoptive cell therapy, or prophylactic cancer vaccine therapy (e.g., neoantigen vaccine therapy). The immune checkpoint inhibitor therapy may include, for example, an anti- PD-1antibody, an anti-PD-Ll antibody, an anti-PD-L2 antibody, an anti-CTLA-4 antibody, an anti- TIM3 antibody, an anti-TIGIT antibody, an anti-VISTA antibody, an anti-B7-H3 antibody, an anti- BTLA antibody, an anti-CD73 antibody, or an anti-LAG-3 antibody, among others. The adoptive cell therapy may include, for example, chimeric antigen receptor (CAR) T-cell therapy, tumor-infiltrating lymphocyte (TIL) therapy, engineered T cell receptor (TCR) therapy, or Natural killer (NK) cell therapy, among others. The neoantigen vaccine therapy may include, for example, an RNA neoantigen vaccine, a synthetic long peptide neoantigen vaccine, or a dendritic cell (DC)- based neoantigen vaccine, among others.

[0090] Each candidate cancer therapy may correspond to a respective range of values for the correlation score 235. The output handler 150 may select the cancer therapy from the set of candidate cancer therapies by comparing the correlation score 235 with the range of values for each candidate cancer therapy. When the correlation score 235 is within the range of values for the immune checkpoint inhibitor therapy, the output handler 150 may select the immune checkpoint inhibitor therapy. When the correlation score 235 is within the range of values for the adoptive cell therapy, the output handler 150 may select the adoptive cell therapy. When the correlation score 235 is within the range of values for the neoantigen vaccine therapy, the output handler 150 may select the neoantigen vaccine therapy. When the correlation score 235 is within the range of values for the prophylactic cancer vaccine therapy, the output handler 150 may select the prophylactic cancer vaccine therapy.

[0091] With the selection, the output handler 150 may store and maintain an association between the subject 205 and the selected cancer therapy using one or more data structures on the database 160. The data structures may include, for example, a linked list, a tree, an array, a table, a matrix, a stack, a queue, or a heap, among others. In some embodiments, the output handler 150 may store an association among two or more of the subject 205, the RT burden metric 220, the mutant p53 fitness metric 225, the RT RNA metric 230, the correlation score 235, or the selected cancer therapy. The association may be stored and maintained as the one or more data structures on the database 160.

[0092] In some embodiments, the output handler 150 may provide, send, or otherwise transmit at least one instruction 240 to the administrative device 115. The output handler 150 may generate the instruction 240 to include information based on the association. For instance, the information may include an identifier of the subject 205 (e.g., anonymized identifier), one or more metrics (e.g., the RT burden metric 220, the mutant p53 fitness metric 225, the RT RNA metric 230), the correlation score 235, and an identification of the selected cancer therapy, among others. The instruction 240 may also include a script defining the presentation of the information on the administrative device 115. With the generation, the output handler 150 may send the instruction 240 to the administrative device 115. Upon receipt of the instruction 240, the administrative device 115 may display, render, or otherwise present the information. For instance, the administrative device 115 may display the information via a graphical user interface of an application that is interfacing with the data processing system 105. Using the information, a clinician examining the subject 205 may make further determination as to which actions to take or which cancer therapy to administer to the subject 205 to alleviate or treat the cancer.

[0093] In this manner, the data processing system 105 may more accurately identify and quantify RT activity from the genome sequence dataset 210 and RT RNA expressions from the RNA sequence dataset 215. By filtering and de-duplicating gene and RNA sequences, the data processing system 105 may significantly cut the consumption of amount of computer resources (e.g., processor and memory) in processing such sequences. In addition, the data processing system 105 may be able to better disambiguate such sequences to pinpoint and identify RT (e.g., LI retrotransposition) from sequencing reads, relative to other techniques. The ability to better disambiguate may also decrease the amount of computing resources (e.g., processor and memory) and time in processing the reads to disambiguate, compared to other techniques without such abilities to disambiguate. Furthermore, using the RT burden metric 220, the mutant p53 fitness metric 225, and the RT RNA metric 230, the data processing system 105 may determine the correlation scores 235 indicating degrees of correlation or effects among the RT burden, RT RNA expressions, and the mutant p53 fitness. With the correlation score 235, the data processing system 105 may be able to select more appropriatecancer therapies, with the expectation that such cancer therapies will better treat the cancer in the subject 205.

[0094] Referring now to FIG. 15, depicted is a flow diagram of a method 300 of selecting cancer therapies based on magnitudes and frequencies of retrotransposon (RT) insertions in subject genomes. The method 300 can be implemented by any components detailed herein, such as the system 100. Under the method 300, a computing system may obtain genomic and RNA sequences data for a subject (305). The computing system may identify an RT burden metric (310). The computing system may determine a p53 mutant fitness metric (315). The computing system may generate a RT RNA metric (320). The computing system may determine a correlation score (325). The computing system may select a cancer therapy (330). The computing system may store an association between the selected therapy and the subject (335).C. Computing and Network Environment

[0095] Various operations described herein can be implemented on computer systems. FIG. 16 shows a simplified block diagram of a representative server system 400, client computing system 414, and network 426 usable to implement certain embodiments of the present disclosure. In various embodiments, server system 400 or similar systems can implement services or servers described herein or portions thereof. Client computing system 414 or similar systems can implement clients described herein. The system 400 described herein can be similar to the server system 400. Server system 400 can have a modular design that incorporates a number of modules 402 (e.g., blades in a blade server embodiment); while two modules 402 are shown, any number can be provided. Each module 402 can include processing unit(s) 404 and local storage 406.

[0096] Processing unit(s) 404 can include a single processor, which can have one or more cores, or multiple processors. In some embodiments, processing unit(s) 404 can include a general -purpose primary processor as well as one or more special-purpose co-processors such as graphics processors, digital signal processors, or the like. In some embodiments, some or all processing units 404 can be implemented using customized circuits, such as applicationspecific integrated circuits (ASICs) or field programmable gate arrays (FPGAs). In some embodiments, such integrated circuits execute instructions that are stored on the circuit itself. In other embodiments, processing unit(s) 404 can execute instructions stored in local storage 406. Any type of processors in any combination can be included in processing unit(s) 404.

[0097] Local storage 406 can include volatile storage media (e.g., DRAM, SRAM, SDRAM, or the like) and / or non-volatile storage media (e.g., magnetic or optical disk, flash memory, or the like). Storage media incorporated in local storage 406 can be fixed, removable or upgradeable as desired. Local storage 406 can be physically or logically divided into various subunits such as a system memory, a read-only memory (ROM), and a permanent storage device. The system memory can be a read-and-write memory device or a volatile read-and- write memory, such as dynamic random-access memory. The system memory can store some or all of the instructions and data that processing unit(s) 404 need at runtime. The ROM can store static data and instructions that are needed by processing unit(s) 404. The permanent storage device can be a non-volatile read-and-write memory device that can store instructions and data even when module 402 is powered down. The term “storage medium” as used herein includes any medium in which data can be stored indefinitely (subject to overwriting, electrical disturbance, power loss, or the like) and does not include carrier waves and transitory electronic signals propagating wirelessly or over wired connections.

[0098] In some embodiments, local storage 406 can store one or more software programs to be executed by processing unit(s) 404, such as an operating system and / or programs implementing various server functions such as functions of the system 500 of FIG. 5 or any other system described herein, or any other server(s) associated with system 500 or any other system described herein.

[0099] “Software” refers generally to sequences of instructions that, when executed by processing unit(s) 404 cause server system 400 (or portions thereof) to perform various operations, thus defining one or more specific machine embodiments that execute and perform the operations of the software programs. The instructions can be stored as firmware residing in read-only memory and / or program code stored in non-volatile storage media that can be readinto volatile working memory for execution by processing unit(s) 404. Software can be implemented as a single program or a collection of separate programs or program modules that interact as desired. From local storage 406 (or non-local storage described below), processing unit(s) 404 can retrieve program instructions to execute and data to process in order to execute various operations described above.

[0100] In some server systems 400, multiple modules 402 can be interconnected via a bus or other interconnect 408, forming a local area network that supports communication between modules 402 and other components of server system 400. Interconnect 408 can be implemented using various technologies including server racks, hubs, routers, etc.

[0101] A wide area network (WAN) interface 410 can provide data communication capability between the local area network (interconnect 408) and the network 426, such as the Internet. Technologies can be used, including wired (e.g., Ethernet, IEEE 402.3 standards) and / or wireless technologies (e.g., Wi-Fi, IEEE 402.24 standards).

[0102] In some embodiments, local storage 406 is intended to provide working memory for processing unit(s) 404, providing fast access to programs and / or data to be processed while reducing traffic on interconnect 408. Storage for larger quantities of data can be provided on the local area network by one or more mass storage subsystems 412 that can be connected to interconnect 408. Mass storage subsystem 412 can be based on magnetic, optical, semiconductor, or other data storage media. Direct attached storage, storage area networks, network-attached storage, and the like can be used. Any data stores or other collections of data described herein as being produced, consumed, or maintained by a service or server can be stored in mass storage subsystem 412. In some embodiments, additional data storage resources may be accessible via WAN interface 410 (potentially with increased latency).

[0103] Server system 400 can operate in response to requests received via WAN interface 410. For example, one of modules 402 can implement a supervisory function and assign discrete tasks to other modules 402 in response to received requests. Work allocation techniques can be used. As requests are processed, results can be returned to the requester via WAN interface 410. Such operation can generally be automated. Further, in someembodiments, WAN interface 410 can connect multiple server systems 400 to each other, providing scalable systems capable of managing high volumes of activity. Other techniques for managing server systems and server farms (collections of server systems that cooperate) can be used, including dynamic resource allocation and reallocation.

[0104] Server system 400 can interact with various user-owned or user-operated devices via a wide-area network such as the Internet. An example of a user-operated device is shown in FIG. 4 as client computing system 414. Client computing system 414 can be implemented, for example, as a consumer device such as a smartphone, other mobile phone, tablet computer, wearable computing device (e.g., smart watch, eyeglasses), desktop computer, laptop computer, and so on.

[0105] For example, client computing system 414 can communicate via WAN interface 410. Client computing system 414 can include computer components such as processing unit(s) 416, storage device 418, network interface 420, user input device 422, and user output device 424. Client computing system 414 can be a computing device implemented in a variety of form factors, such as a desktop computer, laptop computer, tablet computer, smartphone, other mobile computing device, wearable computing device, or the like.

[0106] Processing unit(s) 416 and storage device 418 can be similar to processing unit(s) 404 and local storage 406 described above. Suitable devices can be selected based on the demands to be placed on client computing system 414; for example, client computing system 414 can be implemented as a “thin” client with limited processing capability or as a high-powered computing device. Client computing system 414 can be provisioned with program code executable by processing unit(s) 416 to enable various interactions with server system 400.

[0107] Network interface 420 can provide a connection to the network 426, such as a wide area network (e.g., the Internet) to which WAN interface 410 of server system 400 is also connected. In various embodiments, network interface 420 can include a wired interface (e.g., Ethernet) and / or a wireless interface implementing various RF data communication standards such as Wi-Fi, Bluetooth, or cellular data network standards (e.g., 3G, 4G, LTE, etc.).

[0108] User input device 422 can include any device (or devices) via which a user can provide signals to client computing system 414; client computing system 414 can interpret the signals as indicative of particular user requests or information. In various embodiments, user input device 422 can include any or all of a keyboard, touch pad, touch screen, mouse or other pointing device, scroll wheel, click wheel, dial, button, switch, keypad, microphone, and so on.

[0109] User output device 424 can include any device via which client computing system 414 can provide information to a user. For example, user output device 424 can include a display to present images generated by or delivered to client computing system 414. The display can incorporate various image generation technologies, e.g., a liquid crystal display (LCD), light-emitting diode (LED) including organic light-emitting diodes (OLED), projection system, cathode ray tube (CRT), or the like, together with supporting electronics (e.g., digital- to-analog or analog-to-digital converters, signal processors, or the like). Some embodiments can include a device such as a touchscreen that function as both input and output device. In some embodiments, other user output devices 424 can be provided in addition to or instead of a display. Examples include indicator lights, speakers, tactile “display” devices, printers, and so on.

[0110] Some embodiments include electronic components, such as microprocessors, storage and memory that store computer program instructions in a computer-readable storage medium. Many of the features described in this specification can be implemented as processes that are specified as a set of program instructions encoded on a computer-readable storage medium. When these program instructions are executed by one or more processing units, they cause the processing unit(s) to perform various operation indicated in the program instructions. Examples of program instructions or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter. Through suitable programming, processing unit(s) 404 and 416 can provide various functionality for server system 400 and client computing system 414, including any of the functionality described herein as being performed by a server or client, or other functionality.

[0111] It will be appreciated that server system 400 and client computing system 414 are illustrative and that variations and modifications are possible. Computer systems used in connection with embodiments of the present disclosure can have other capabilities not specifically described here. Further, while server system 400 and client computing system 414 are described with reference to particular blocks, it is to be understood that these blocks are defined for convenience of description and are not intended to imply a particular physical arrangement of component parts. For instance, different blocks can be but need not be located in the same facility, in the same server rack, or on the same motherboard. Further, the blocks need not correspond to physically distinct components. Blocks can be configured to perform various operations, e.g., by programming a processor or providing appropriate control circuitry, and various blocks might or might not be reconfigurable depending on how the initial configuration is obtained. Embodiments of the present disclosure can be realized in a variety of apparatus including electronic devices implemented using any combination of circuitry and software.

[0112] While the disclosure has been described with respect to specific embodiments, one skilled in the art will recognize that numerous modifications are possible. Embodiments of the disclosure can be realized using a variety of computer systems and communication technologies including but not limited to the specific examples described herein. Embodiments of the present disclosure can be realized using any combination of dedicated components and / or programmable processors and / or other programmable devices. The various processes described herein can be implemented on the same processor or different processors in any combination. Where components are described as being configured to perform certain operations, such configuration can be accomplished, e.g., by designing electronic circuits to perform the operation, by programming programmable electronic circuits (such as microprocessors) to perform the operation, or any combination thereof. Further, while the embodiments described above may make reference to specific hardware and software components, those skilled in the art will appreciate that different combinations of hardware and / or software components may also be used and that particular operations described as being implemented in hardware might also be implemented in software or vice versa.

[0113] Computer programs incorporating various features of the present disclosure may be encoded and stored on various computer-readable storage media; suitable media include magnetic disk or tape, optical storage media such as compact disk (CD) or DVD (digital versatile disk), flash memory, and other non-transitory media. Computer-readable media encoded with the program code may be packaged with a compatible electronic device, or the program code may be provided separately from electronic devices (e.g., via Internet download or as a separately packaged computer-readable storage medium).

[0114] Thus, although the disclosure has been described with respect to specific embodiments, it will be appreciated that the disclosure is intended to cover all modifications and equivalents within the scope of the following claims.EXAMPLES

[0115] The present technology is further illustrated by the following Examples, which should not be construed as limiting in any way.Example 1: Materials and MethodsPan-cancer datasetsTCGA

[0116] Whole genome sequencing data were downloaded from the Genomics Data Commons (GDC) cloud storage, using the GA4GH standard Data Repository Service (DRS) for URI resolution. DRS links to the hgl9-aligned BAM files were collated from pre-existing Terra workspaces (https: / / app.terra.bio) for TCGA (vl.O of ControlledAccess data). RNA-Seq data were downloaded from the GDC directly (https: / / portal.gdc.cancer.gov / repository) as hg38- aligned BAM files using the GDC Data Transfer Tool (https: / / docs.gdc.cancer.gov / Data_Transfer_Tool). All subsequently described processing of these data was performed within the Terra.bio cloud data platform (https: / / terra.bio / ).Genome in a Bottle

[0117] Alignments to hg38 and indices (BAM and BAI fdes, respectively) for PacBio and Illumina reads for the Genome in a Bottle project22were downloaded from the NCBI archive. A complete list of access links can be found in Supplementary Table 1.Validation against long-read sequencing of LINE- 1 retrotransposition

[0118] Non-reference LI elements were called in the long read data using PALMER v 2.0.039, and high confidence autosomal calls were selected. Then the calls were manually curated by visually reviewing the PacBio reads at the loci of PALMER calls in IGV and identifying likely false negatives (non-reference LI elements not called by PALMER in some of the members of the trio) and manually adding the calls in additional trio members according to the manual assessment (based on reads with insertions and / or clipped reads). LI elements that were present in the genome of all the members of the son / father / mother trio were removed from consideration. 70 LI elements were left present in the genomes of only one or two members of the trio of Ashkenazi Jewish ancestry (HG002 / HG003 / HG004) and 123 LI elements present in the genomes of only one or two members of the trio of Han Chinese ancestry (HG005 / HG006 / HG007). As a sanity check of this curated call set, it was verified that there were no LI elements present in the genome of a child (HG002 or HG005) and absent from the genomes of both parents. This curated set of LI elements was used for comparison with the LI calls using Total ReCall on the short reads.

[0119] Total ReCall (the Illumina short read-based LI insertion caller described below) was ran using different members of each trio as “case” and “control” (resulting in all 6 pairwise comparisons for each trio) and compared the results to the curated long-read calls as “ground truth”. For a “perfect” agreement between a PacBio and an Illumina call, it was required that the call is present in all pairwise comparisons where it should be present with the correct filtering status (e. g., if a given LI element is present only in the maternal genome in PacBio calls, it should be called in the mother vs. father and the mother vs. child and not called in any other of the 6 pairwise comparisons; if a given LI element is present in the paternal and the child genomes in PacBio calls, it should be called in the father vs. mother and the child vs. mother, but also called in the father vs. child and the child vs. father yet filtered-out due to itspresence in the “case” sample). In addition, it was also assessed if the strandedness and the presence or absence of a LI inversion agreed between the PacBIo and Illumina calls.

[0120] Overall, 137 out of 193 long-read calls were also called properly by Total ReCall, including 93 canonical retrotranspositions and 44 inversion-containing retrotranspositions. The transposon length was inferred for each of these 93 calls using PALMER calls as the distance between the middle points of the inferred confidence intervals for the start and end points in the LI (thus excluding possible 3’ transductions) and compared it to the length inferred using Total ReCall. Agreement was very good (Pearson correlation R > 0.99, p < 1 x io10, t-test, FIG. 6).Short-read whole-genome sequencing data from TCGAReprocessing of alignment files

[0121] Whole-genome sequencing reads data from The Cancer Genome Atlas (TCGA) were reverted to unaligned BAM format using the GATK v4.1.8.1 tool RevertSam with options “-SORT ORDER queryname -VALIDATION STRINGENCY SILENT.” Any reads that were not paired-end were dropped, and the remaining reads were converted to FASTQ format with GATK v4.1.8.1 tool SamToFastq with options “ INTERLEA VE=trueINCLUDE NON PF READS=true”. The paired- end FASTQ files were then aligned to the hg38 human reference using BWA-MEM vO.7.15- rl 14040 with options “-K 100000000 -p -v 3 -t 16 -Y -T 0”, where we used the “-Y” option to retain clipped read sequences for supplementary alignments that are clipped (i.e., “soft-clipping”), since clipped read information is the key evidence that Total ReCall leverages (see below). Duplicate reads were marked with GATK v4.1.8.1 tool MarkDuplicates using optionsOPTICAL DUPLICATE PIXEL DISTANCE 2500 -VALID ATION STRINGENCY SILENT”.Quality control data filtering

[0122] Variability of sequencing library quality across TCGA WGS data (which was generated over the span of years) can obscure the differentiable signal of biologicalrelationships. With that in mind, the entire WGS dataset (4,314 samples from 2,106 patients) was filtered to only those samples sequenced with paired-end reads, where both mates had been retained in the alignment files stored in NCI Genomic Data Commons (4,142 samples from 2,058 patients). Each of these samples was realigned to the hg38 reference genome as described above. To de-duplicate individual patients with multiple tumor or normal samples, only the primary tumor and a single normal sample were retained, resulting in 1,848 tumornormal pairs from 1,848 patients. Using all 1,848 tumor-normal pairs agnostic to additional data quality parameters, no correlation could be found between LI RNA expression, p53 mutational status, and LI retrotransposition, due to very large confounding differences between the libraries. To account for these differences and better capture biological relationships, the quality of each WGS library was evaluated using a modified Picard41tool to quantify the total number of reads, average base quality, average read length, and fractions of reads with split or discordant alignments (the two sources of information on which Total ReCall relies). All reads that did not have a matched pair, that were marked as duplicates, that had mapping quality equal to zero, or whose base and quality score strings were inconsistent (i.e., of differing lengths) were removed from the analysis. From the remaining reads, for each sample, it was calculated that 1) the fraction of chimeric reads, 2) the fraction of overall clipped bases, 3) the average read length, 4) the average base quality, and 5) the total coverage. Note that a chimeric read is defined per Picard / GATK v4.1.8.1 as a read pair aligned in an unexpected orientation or significantly further apart than expected (with maximum insert size set to 100,000). Because Total ReCall will depend on signal from split-read alignments (i.e., clipped reads) to nominate candidate insertion sites, and because bwa mem is intended for use with reads a minimum of 70bp, tumor-normal pairs where either sample was sequenced with reads shorter than 70bp were removed from the analysis dataset. In addition, all paired samples with a chimeric read fraction in the tumor or normal sample greater than 2% were removed, resulting in a total of 765 tumor-normal pairs from unique patients across 22 tumor types in the final analysis dataset.

[0123] Due to their importance, the five metrics for each tumor and paired normal sample calculated here were also used to adjust the retrotransposition count estimates for each tumor sample, as shown in FIG. 8 and described below.Identifi cation of somatic LINE-1 retrotransposition

[0124] The Total ReCall method for identification of somatic LI insertions uses both soft-clipped reads, as well as reads not properly mapped as a pair (“discordant” read pairs) as a signal. The clipped reads were used as the primary signal because of their higher specificity: the last mapping point as well as the clipped sequence are the same (up to sequencing errors) for all the clipped reads supporting the same breakpoint. Interpretation of these types of signals arises from the mechanistic biochemistry of the life cycle of retrotransposons.

[0125] Class I transposable elements replicate via a reverse transcribed RNA intermediate (“copy-and- paste”) in a process known as retrotransposition, whereas class II elements replicate via a DNA intermediate (“cut-and-paste”). Further subdividing based on distinct mechanisms of mobility, there are three major categories of transposable elements: class I LTR retrotransposons (including endogenous retroviruses), class I non-LTR retrotransposons, and class II DNA transposons42. Among the transposable elements, only some non-LTR retrotransposons are known to be able to move within the genome. LI elements are capable of autonomous retrotransposition, and SINE (e.g., Alu) and SVA elements are “parasites” that rely on the LI machinery for mobility.

[0126] The difference between the mechanisms utilized by LTR and non-LTR retrotransposons is that the LTR retrotransposons first synthesize the double stranded DNA from RNA transcript in a complex process involving the long terminal repeats (LTRs) and then integrate that sequence into the host genome. On the other hand, enzymes of non-LTR retrotransposons (LINE elements) reverse transcribe their RNA directly into the genome using the so-called target-primed reverse transcription (TPRT) process43.

[0127] Unlike the LTR retrotransposons, which always integrate the complete provirus with the LTRs into the genome (sometimes, the internal provirus part may be removed by homologous recombination leaving a solo LTR), reverse transcription of a non-LTR retrotransposon may be terminated prematurely resulting in a partial integration. Apart from such a 5’ truncation, another common structural variant of LI is inversion resulting from double priming and simultaneous reverse transcription of the same LI transcript into bothstrands of the genome (FIG. ID)44,45. In addition to that, it is not uncommon for there to be 3’ transductions, where during LI RNA expression the polymerase overruns the poly (A) signal at the 3’ end of the L I element resulting in transcription and retrotransposition of some genomic sequence downstream of the 3’ end of the source LI element in addition to the LI element itself23,46. In the Total ReCall method developed here, the focus became detecting these three most common features of the structural variations that result from LI retrotranspositions: 5’ truncation, LI inversion, and 3’ transduction. It is the belief that the analysis of even rarer and more complex structural variations arising from LI retrotransposons (e.g., instances with more than one template switching during the reverse transcription process, which would result in a series of inversions) is better left for long-read data.

[0128] The Total ReCall algorithm (FIG. 1) was then designed to detect both “canonical” retrotransposition events and “inversion-containing” retrotranspositions. In particular, “single-stranded” retrotransposition (FIG. 1C) leads to a canonical retrotransposition with a single segment of LI inserted into the genome (FIG. IE), with the two ends of the insertion detectable in the clipped short-reads (FIG. 1G). On the other hand, a “double-stranded” retrotransposition event (FIG. ID), where each strand of the target site genomic DNA incorporates reverse transcribed sequence of a single LI mRNA, leads to an inversion-containing retrotransposition with a DNA insertion that contains part of a LI sequence and another part of it that is inverted, i.e., contains reverse complementary LI sequence, in an adjacent position in the genome (FIG. IF), which will be reflected in the shortreads data (FIG. 1G).

[0129] To identify the breakpoints of the LI insertion event, Total ReCall collects the clipped sequences of the reads in the vicinity, sort them by length in decreasing order and cluster in a cd-hit-like47’48way: take each new sequence and align it against the existing representatives; if a match is found, assign the sequence to the corresponding cluster, otherwise designate it as a new representative. The reads representing the “left” (clipped on the 3’ end w.r.t. the reference genome) and the “right” (clipped on the 5’ end w.r.t. the reference genome) breakpoints were grouped, respectively. The longest clipped sequence in the cluster serves as the representative of the breakpoint. If the clipped sequence was oriented so that the clippingpoint is at its 5’ end, a canonical LI retrotransposition is characterized by one breakpoint with the poly(T) sequence (regardless of whether a 3’ transduction is present) and the other breakpoint with a sequence mapping to the positive strand of the LI sequence. The coordinate of the alignment of the clipped sequence at the breakpoint to the LI sequence can be used to infer the length of the transposon (without the 3’ transduction, if one exists). An inversioncontaining LI retrotransposition is characterized by one breakpoint bearing the poly(T) sequence (regardless of whether a 3’ transduction is present) and the other breakpoint with a sequence mapping to the negative strand of the LI sequence.

[0130] As noted above, the sensitivity was increased by using the “soft clipping” instead of the “hard clipping” for supplementary alignments (bwa option- Y) relevant when the longer part of the read spanning the breakpoint contains the LI sequence (and whose primary mapping is thus to LI genomic sequence elsewhere in the genome). The use of such supplementary alignments is highly desirable since they provide a longer clipped sequence, which can in turn be aligned to the LI sequence more specifically.

[0131] Total ReCall also identifies reads that belong to discordant read pairs and map near the breakpoints and use their mates as a secondary signal to determine the confidence of the call. If such a read is also clipped at the breakpoint location, its mate is given a higher preference over the mates of the other reads.

[0132] Clipped sequences, as well as sequences of the reads belonging to discordant read pairs, were aligned to the LI consensus sequence using LAST aligner49. The consensus sequence of the LI transposon was constructed by merging the three DFAM50 sequences for its 3’, 5’ ends and ORF2 subdomain (DF0000225, DF0000226, DF0000316). The ambiguous characters in the DFAM consensus sequences were resolved using the alignment of the 146 intact LI sequences from LI base24.

[0133] To identify somatic LI retrotranspositions (i.e., in the typical use case of tumorspecific insertions), Total ReCall checked each call in the case sample for the presence in its control sample of the signal of a clipped read with a matching sequence or a breakpoint, retaining as “somatic” (i.e., case-specific) only calls without such signal in the control. Inaddition, filters were implemented for low complexity and regions with large numbers of alignment artifacts (“high entropy regions”), e.g., regions having too many clipped reads with distinct sequence and low complexity regions where genomic sequence matches clipped sequence. Finally, Total ReCall requires each of the 2 breakpoints of a called insertion rely on at least 3 clipped reads.

[0134] It is worth mentioning that the short-read data possess inherent limitations, for example, difficulty identifying 3’ transductions. Long such transductions may be identified via discordant read pairs whose mates map downstream of source LINE-1 element. At the same time, this method will fail to identify short 3’ transductions. Some other shortcomings are the ambiguity of pairing between the left / right breakpoints if multiple ones are present in the same region as well as low mappability of some genomic regions (in particular, peri centromeric regions).RNA sequencingReprocessing of alignment files

[0135] Public RNA-seq data from TCGA was reverted to unaligned FASTQ format using GATK v4.1.8.1 tools RevertSam with options SORT ORDER “queryname” — VALIDATION STRINGENCY SILENT” and SamToFastq with options “INTERLEAVE=false INCLUDE_NON_PF_READS=true.” Any reads that were not paired were dropped. The paired-end FASTQ files were then aligned to the hg38 human reference genome using STAR v2.7.9a.Quality control data filtering

[0136] Alignment files for a total of 10,904 RNA-seq samples from 10,089 individuals (10,174 tumor samples and 730 normal samples) were downloaded from GDC as described above. 28 of these (all tumor samples) could not be reverted to fastq for realignment due to the presence of unpaired reads or otherwise corrupted downloaded alignment files. The remaining 10,876 samples were realigned as described above. 634 of these (623 tumor samples and 11 normal samples) were sequenced with single-end reads, and therefore removed from thedataset. 59 additional tumor samples were removed from the dataset for containing a strandspecific sequencing library. Finally, any metastatic, recurrent, or new primary tumor samples were filtered out and deduplicated the patient samples in the dataset to include no more than one primary tumor and one normal sample per patient, resulting in a dataset of 9,011 tumor samples and 719 normal samples from 9,081 individuals across 32 tumor types.LINE-1 RNA quantification

[0137] LI RNA expression was quantified in the 9,011 tumor and 719 normal RNA-seq samples using L1EM24, which formalizes a framework for quantification of expression that is based on the mechanisms of active transcription of LI elements. Read counts for proper expression at loci corresponding to the LIBase v2 database24of “FLI-Lls” (Lis that are full length and intact in both ORFs) were converted to transcripts per million (TPM) of active, fully -intact LI expression.Counting L1HS alignments

[0138] To compare performance of LI RNA quantification by L1EM, LI expression was also estimated using samtools view vl .3.1 with options “-F 1540” and “-L” to pass coordinates in bed format of L1HS annotations from RepeatMasker open-4.0.5 Repeat Library 201401313.

[0139] From the output of alignments which overlapped the input L1HS coordinates, unique qnames were counted, and normalized against the total unique qnames in the alignment file to calculate an approximate expression level in reads per million (RPM) for every sample (FIG. 12). The samples are noticeably more homogeneous than when using L1EM (FIG. 3A), and the largest increase in median tumor estimate over median normal estimate occurs in stomach adenocarcinoma, where the tumor median is only 13% higher than the normal median.Categorizing tumors as p53 wild type or mutant

[0140] Categorical designations of p53 alteration were obtained through the public repository cBioPortal (www.cbioportal.org, accessed 23 February 2023) by querying “TCGA PanCancer Atlas Studies” (which includes 10,967 samples from 10,953 patients in 32 studies)for all alterations in “TP53”. A sample will be marked as “altered” in TP53 if any non- synonymous mutations, amplification, deep deletion, or structural variants have been identified in that sample. Shallow deletions or low- level copy number gains of the gene do not contribute to classification of a sample as altered. The altered annotations were used to divide tumor samples into p53 mutant and wild type categories to test significance with Mann- Whitney U tests as shown in FIG. 4, and to demonstrate robustness of the mediation model between p53, LI RNA, and LI RT burden as shown in FIG. 11.Adjusting RNA and RT estimates based on sequencing metrics10141] Intronic rate of RNA-seq, which has previously been shown to be a key confound for LI expression31, was calculated using RNA-SeQC v252. A linear regression between the LI RNA estimates and the intronic rate was then performed in Python v3.7.12 using statsmodels v0.13.0 OLS to perform an ordinary least squares regression. The residuals from this model were then used as the adjusted LI RNA estimates (FIG. 8B).

[0142] Similarly, quality metrics for the WGS samples were calculated as described above. An ordinary least squares model was fitted to the tumor-specific retrotransposition counts as a function of the average read length, depth of coverage, average base quality, chimeric read fraction, and clipped base fraction in both the tumor sample and the paired normal sample. The residuals from this model were then used as the adjusted LI RT estimates (FIG. 8C).Quantifying the functional fitness of p53|0143] Somatic TP53 mutations, copy number variation data, and the variant allele fractions for TP53 mutations in TCGA samples were downloaded from the National Cancer Institute’s Genomic Data Commons repository53. The variant allele fractions were averaged across mutation callers. Functional fitness of missense mutations was scored as defined in previous work21. Reverse- phase protein assay data and copy number alteration data were utilized to infer mutant- and tissue-specific p53 protein concentrations as previously described21.

[0144] The contribution of p53 to tumor fitness was determined as follows. Wild-type p53 acts as a transcription factor and regulates DNA damage by transactivating pro- senescence / apoptotic genes. Mutations in p53 often result in loss of DNA-binding ability to varying degrees. The oncogenic advantage of having a p53 mutation was estimated by the probability of mutant p53 not binding eight principal transcriptional targets (WAFJ, MDM2, BAX, h!433s, AIP1, GADD45, NOXA, and P53R2). The probability of not binding target DNA depends on the concentration of the protein and the affinity of the mutation to the target site. Importantly, the concentration of p53 depends on the tissue. In this work, the method described in Hoyos et al. 2022 for missense mutations21was utilized. For wild-type p53, the fitness in an analogous fashion to the missense mutations was computed, except that wild-type p53 concentrations and wild-type p53 affinities to DNA promoter sites were used. This method extends to other mutations as follows. For silent / non-coding mutations, the fitness was considered to be wild type. For nonsense and frameshift mutations, the lack of binding was considered to be dependent on the position of the first altered residue along the protein.Specifically, the fitness for these mutations was computed as follows:where FMTis the fitness of the mutation, FWTis the fitness of wild-type p53, Fmaxis the maximum fitness (which corresponds to a complete deletion of the p53 gene), and p is the position of the mutation (which may take an integer value between 1 and 393, inclusive). In this way, a position of “1” corresponds to maximum fitness, and a position of “394” (which would be after the protein) corresponds to wild type fitness. Tumors with wild-type p53 are assigned the wild-type p53 fitness that corresponds to the tissue in question.Mediation model

[0145] Statistical analysis was performed in Python v3.7.12 using statsmodels v0.13.0 OLS to evaluate the mediation model. For the 620 samples for which WGS, RNA-seq, and p53 mutation data was available, five linear regressions were fitted (see FIG. 5B):

[0146] (1) adjusted log2(RT) = T X p53 mutation score + ci

[0147] (2) adjusted log2(RT) = 0 x adjusted log2(RNA) + C2

[0148] (3) adjusted log2(RNA) = a x p53 mutation score + C3

[0149] (4) adjusted log2(RT) = ?’ 0’ adjusted log2(RNA) + T’ X p53 mutation score + C4

[0150] (5) adjusted log2(RNA) = 0* x adjusted log2(RT) + a* x p53 mutation score +C5

[0151] Note that only equations (3) and (4) are necessary for quantifying the magnitude and significance of the mediation. Equation (1) is used to normalize the mediated and unmediated effects to the total effect of p53 on LI RT burden. Equation (2) is included here to evaluate the correlation between LI RNA and RT burden alone, without considering p53. Equation (5) (results not shown in FIG. 5 for simplicity) was evaluated to confirm that p53 has a significant effect on LI RNA even when controlling for LI RT burden. In each model, Ci incorporates both the intercept and the error terms. The fitted coefficient values were then standardized based on the estimated standard deviations for each variable34. The ratio of T’ to T gives the estimate for the percentage of the total impact of p53 on LI RT burden that is not mediated by LI RNA. The product of coefficients a and 0’ (which is equivalent to T - T’) gives a coefficient for the effect of p53 on LI RT burden via LI RNA, and similarly the ratio of oc0’ to T gives the estimate for the percentage of the total impact of p53 on LI RT burden that is mediated by LI RNA. 95% confidence intervals for all estimates were calculated using bootstrap resampling of the samples, N=100,000.Example 2: Total ReCall uses clipped sequences to accurately identify somatic LI insertions

[0152] The Total ReCall algorithm (FIG. 1) was designed to analyze patterns in paired- end whole-genome sequencing (WGS) data to detect “canonical” retrotransposition events (where a single stretch of LI starting from the 3’ end is inserted in the genome; FIG. IE), as well as “inversion -containing” retrotranspositions (a genomic insertion that contains part of an LI sequence adjacent to an inverted part of LI; FIG. IF; see Methods). To derive somatic insertions in tumors, the algorithm looks for read patterns (FIG. 1G) that are specific to the tumor sample and not found in normal tissue from the same patient. For validation, The TotalReCall was tested using a child-mother-father trio sequenced with both Illumina short -read and PacBio long-read technologies as part of the NIST-led Genome in a Bottle (GIAB) Consortium22, using the latter as ground truth for retrotransposition events (see Methods). The Total ReCall on the short-read data was found to have had high sensitivity (71%, FIGs. 6A-6B) for detection of long-read-based trio retrotransposition calls, and the inferred lengths of the retrotransposed sequence insertions (estimated for canonical events only) were highly correlated (Pearson correlation R > 0.99, p < IxlO'10, FIG. 6C).

[0153] Applying Total ReCall to TCGA, a total of 4,357 tumor-specific somatic retrotranspositions were found in the quality-controlled (see Methods) dataset of 765 tumors (mean of 5.7 calls per sample; FIG. 2A). As a negative control, insertions found in the normal tissue but not present in the matched tumor were looked for, which yielded only 100 such instances (mean of 0.13 calls per sample), consistent with a low number of false positives. Approximately two thirds (65.5%) of tumor samples were found to have no somatic retrotransposition calls (in the negative control, 91.6% of normal samples had no normalspecific calls). Because many tumor samples were sequenced at higher depths than their paired normal samples, The tumor samples were confirmed to have had a significantly higher rate of somatic LI calls than the normal samples even when adjusting for sequencing depth in various ways (FIG. 7).

[0154] Among the somatic retrotranspositions detected, slightly over half (2,416 of 4,357, 55.5%) had a canonical orientation of inserted LI sequence that is consistent with single-stranded priming (FIG. 2B, FIG. 1C), while the remaining (1,941, 44.5%) showed evidence of inversion-containing integrations that result from double-stranded priming (FIG. ID). This inversion-containing frequency is higher than previous estimates of approximately 19-25% within germline LI polymorphisms23,15, possibly reflecting the higher levels of genomic instability in tumors. As expected, the length of inserted LI sequence (estimated from the 2,416 canonical insertions, FIG. 1G) shows a pattern of 5’ decay, with over 94% containing a 5 ’-truncated LI sequence, and only 6% approaching full length of a reverse- transcribed LI transcript (defined here as having at least 97% of the LIHs consensus [5,833 of 6,032 bases], FIG. 2C). The “mortality rate” for the LI life cycle in cancer (i.e., the fraction ofnew LI integrations that fail to insert the full LI sequence) was estimated as being nearly 19 out of 20.Example 3: Pan-cancer survey reveals highest LI retrotransposition “burden ” in lung squamous cell carcinoma

[0155] In the analysis of 765 tumors across the 22 tissue types available in the quality- controlled TCGA data, A high degree of heterogeneity was found in the number of somatic retrotranspositions per sample, dubbed the “RT burden”. In particular, lung squamous cell carcinoma had the highest median RT burden of 13 (FIG. 2D). The next highest, head and neck squamous cell carcinoma tumors had a median RT burden of 8, and the highest mean burden (32.6 RTs / sample) of any tumor type in the dataset. After these, colon adenocarcinoma samples had a median RT burden of 7, and uterine corpus endometrial carcinoma and rectal adenocarcinoma samples both had a median of 2. Stomach, ovarian, cervical, and bladder tumors all had a median RT burden of 1, while breast cancers had a median of 0, but 44% of samples did have at least one somatic RT. In contrast, for the remaining tumor types there were no detection of any somatic retrotranspositions in 364 of 388 samples.

[0156] Note, however, that without sufficient quality control (QC, see Methods) of the whole-genome sequencing data, the patterns observed were obscured. In particular, the QC- pipeline removed tumor-normal pairs for which sequencing had been performed earlier in the TCGA project, when read lengths were shorter and thus would not allow the detection of clipped reads (the critical component of Total ReCall), giving the appearance of 0 retrotranspositions in these samples (see Methods). Unfortunately, this filtering removed all esophageal carcinoma and uveal melanoma data from the analysis, cases which are likely to have many RT events11,14.Example 4: LI RNA expression also differs by tumor types but is related to RT burden

[0157] Throughout TCGA, high-quality tumor RNA-seq data is available from 9,011 distinct individuals across 32 tumor types, and normal RNA-seq from 719 individuals across 23tumor types (see Methods). Locus-level LI RNA “active” expression (transcription driven by the LI promoter) were quantified from these 9,730 samples using L1EM tool16and aggregated expression levels (transcripts per million, TPM) across fully intact loci (as curated by LIBase24) to calculate the total relative abundance of intact LI RNA present in each sample.

[0158] Lung squamous cell carcinoma was found to have the greatest median gain of LI expression in tumor samples with respect to paired normal (tumor median 33.0 TPM, normal median 2.5 TPM, p < 1x1 O’10one-sided Mann-Whitney U test), while esophageal samples have the highest expression of LI RNA overall among both tumor and normal samples (tumor median 41.9 TPM, normal median 11.1 TPM, p = 5xl0-5one-sided Mann-Whitney U test, FIG. 3A). Prostate tissue had the next highest levels of LI RNA in normal samples after esophagus, and no significant gain of LI RNA in the prostate adenocarcinoma samples (tumor median 10.0 TPM, normal median 9.5 TPM, p = 0.12 one-sided Mann- Whitney U test).

[0159] Next, to relate LI retrotransposition to LI expression, signals within tumor types were started by aggregating and comparing summary statistics . For the 21 tumor types with at least 5 tumor samples in both the WGS and RNA-seq datasets, each tumor type was ranked twice, first based on average LI RNA expression, then by average RT burden. Tumor types with higher retrotransposition ranks tended to also have higher LI RNA rank (Pearson correlation coefficient R = 0.57, FIG. 8A). The 10 tumor types with an average RT burden of at least 1 per sample have higher median LI RNA expression (median of 14.8 TPM), while the 11 tumor types with an average below one retrotransposition per sample have lower LI RNA expression (median of 3.3 TPM, p = 0.022, two-sided Mann-Whitney U test, FIG. 3B). The subset of 620 cancer patients were compared where both retrotransposition and expression could be assessed, i.e., both WGS and RNA-seq had been performed. At the individual sample level, tumors with at least one retrotransposition had significantly higher LI RNA expression (median 18.0 TPM, N=189) than those with none (median 3.5 TPM, N=431; p < IxlO'10, two- sided Mann-Whitney U test, FIG. 3C).Example 5: Pan-cancer statistical model finds that p53 limits retrotransposition by repression of LINE- 1 transcription and regulation of integration

[0160] To incorporate the orthogonal signal of TP53 mutational status into the analyses, the estimates were first adjusted for LI RNA and RT burden per sample based on sequencing quality metrics to minimize technical biases that may make distinguishing biological relationships more difficult (see Methods, FIGs. 8B-8C). Consistent with the relationship between LI RNA and LI RT in FIGs. 3B-3C, the adjusted values for these also significantly correlate with each other for the tumors with both WGS and RNA-seq data (N = 620, Pearson correlation R = 0.55, p < IxlO'10, FIG. 8D). The TP53 alteration status as defined by cBioPortal25,26was then used to group the WGS and RNA-seq datasets into p53 mutant and wild type categories. The p53 mutant groups exhibited significantly higher LI RNA and LI RT burden than the wild type samples (p < 1 xlO'10for both, two-sided Mann-Whitney U test, FIG. 4)

[0161] A “fitness model” was previously created of the selective advantage a TP53 mutation gives to a tumor cell to quantify the pro-tumor advantage of loss of native p53 function conferred by missense mutations in a mathematical model (defined by the ability to bind 8 key transcriptional targets)21. The fitness model has been further extended here to represent the corresponding impact of additional types of alterations other than point mutations to TP53 (see Methods) for the first time, enabling us to compute p53 functional fitness for every tumor sample in the dataset (FIG. 5A). We observe that LI RNA expression and RT burden are each significantly correlated with mutant p53 fitness (R=0.30, p < IxlO'10; R=0.25, p < IxlO'10, respectively, Pearson correlation, FIG. 9). Additionally, the correlation between LI RNA and mutant p53 fitness remains significant when using LI RT burden as a covariate, and the correlation between LI RT burden and fitness similarly remains significant when using LI RNA as a covariate.

[0162] It was then tested whether LI RNA expression mediates the relationship between p53 mutational fitness and LI RT burden. Statistical mediation analysis (see Methods) was performed on the subset of tumor samples for which WGS, RNA-seq, and p53 mutation data were all available (N = 620 tumor samples). The results revealed a significant mediated effect of p53 mutational fitness on LI RT burden through LI RNA (a[3’= 0.20, p <1x10-5, FIG. 5B). The unmediated effect of p53 mutational fitness on LI RT burden was also significant (T’ = 0.10, p = 1.2xl0’3). The combined linear regression model of LI RT burden as a function of L I RNA and p53 functional fitness has a Pearson correlation coefficient R = 0.56 (p < IxlO'10, FIG. 10). This shows a complementary partial mediating role of LI RNA expression on the relationship between p53 mutational status and LI RT burden, suggesting that p53 regulates somatic retrotransposition via at least two mechanisms: one that is LI RNA- dependent, and one that is not. This modeling at scale is thus highly consistent with the proposed dual regulatory role of p53 in restraining LI retrotransposition27, with the non-Ll RNA mechanism likely mediated through regulation of genomic instability-associated processes.REFERENCES1. Kazazian, H. H. & Moran, J. V. Mobile DNA in Health and Disease. New Engl J Med3TJ, 361-370, doi:10.1056 / NEJMral510092 (2017).2. Hoyt, S. J. et al. From telomere to telomere: The transcriptional and epigenetic state of human repeat elements. Science 376, 57, doi:ARTN eabk31 1210.1 126 / science.abk31 12 (2022)3. Smit, A. F. A., Hubley, R . & Green, P. RepearMasker 4.0. (2013-2015).4. Leonova, K. I. et al. p53 cooperates with DNA methylation and a suicidal interferon response to maintain epigenetic silencing of repeats and noncoding RNAs. P Natl Acad Set USA 110, E89-E98, doi:10.1073 / pnas, 1216922110 (2013).5. Menendez, L., Benigno, B. B. & McDonald, J. F. LI and HERV-W retrotransposons are hypomethylated in human ovarian carcinomas. Mol Cancer 3, doi 10.1 186 / 1476-4598-3- 12 (2004).6. Burns, K. H. Transposable elements in cancer. Nat Rev Cancer 17, 415-424, doi: 10.1038 / nrc.2017.35 (2017).7. Ting, D. T. et al. Aberrant Overexpression of Satellite Repeats in Pancreatic andOther Epithelial Cancers. Science 331, 593-596, doi: 10.1126 / science.1200801 (2011).8. Chiappinelli, K. B et al. Inhibiting DNA Methylation Causes an Interferon Response in Cancer via dsRNA Including Endogenous Retroviruses. Cell 162, 974-986, doi: 10.1016 / j. cell.2015.07.011 (2015).9. Solovyov, A. et al. Global Cancer Transcriptome Quantifies Repeat Element Polarization between Immunotherapy Responsive and T Cell Suppressive Classes. Cell Rep 23, 512- 521, doi:10.1016 / j. celrep.2018.03.042 (2018).10. Roulois, D. et al. DNA-Demethylating Agents Target Colorectal Cancer Cells by Inducing Viral Mimicry by Endogenous Transcripts. Cell 162, 961-973, doi: 10.1016 / j. cell.2015.07.056 (2015).11. Rodriguez-Martin, B. et al. Pan-cancer analysis of whole genomes identifies driver rearrangements promoted by LINE-1 retrotransposition. Nat Genet 52, 306-319, doi : 10.1038 / s41588-019-0562-0 (2020).12. Tanne, A. et al. Distinguishing the immunostimulatory properties of noncoding RNAs expressed in cancer cells. P Natl Acad Sci USA 112, 15154-15159, doi: 10.1073 / pnas.1517584112 (2015).13. Burns, K. H. & Boeke, J. D. Human Transposon Tectonics. Cell 149, 740-752, doi: 10.1016 / j. cell.2012.04.019 (2012).14. Tubio, J. M. C. et al. Extensive transduction of nonrepetitive DNA mediated by LI retrotransposition in cancer genomes. Science 345, 531-531, doi:ARTN125134310.1126 / science.1251343 (2014).15. Chu, C. et al. Comprehensive identification of transposable element insertions using multiple sequencing technologies. Nat Commun 12, doi:ARTN 3836 10.1038 / s41467- 021- 24041-8 (2021).16. McKerrow, W. & Fenyo, D. L1EM: a tool for accurate locus specific LINE-1RNA quantification. Bioinformatics 36, 1167-1173, doi: 10.1093 / bioinformatics / btz724 (2020).17. Hollstein, M., Sidransky, D., Vogelstein, B. & Harris, C. C. P53 Mutations in Human Cancers. Science 253, 49-53, doi: 10.1126 / science.1905840 (1991).18. Haoudi, A., Semmes, O. J., Mason, J. M. & Cannon, R. E. Retrotranspositi on- competent human LINE-1 induces apoptosis in cancer cells with intact p53. J Biomed Biotechnol, 185-194, doi: 10.1155 / S1110724304403131 (2004).19. Tiwari, B. et al. p53 directly represses human LINE1 transposons. Gene Dev 34, 1439- 1451, doi: 10.1101 / gad.343186.120 (2020).20. Wylie, A. et al. p53 genes function to restrain mobile elements. Gene Dev 30, 64-77, doi:10.1101 / gad.266098.115 (2016).21. Hoyos, D. et al. Fundamental immune-oncogenicity trade-offs define driver mutation fitness. Nature 606, 172-178, doi: 10.1038 / s41586-022-04696-z (2022).22. Zook, J. M. et al. Extensive sequencing of seven human genomes to characterize benchmark reference materials. SciData 3, doi: 10.1038 / sdata.2016.25 (2016).23. Szak, S. T. et al. Molecular archeology of LI insertions in the human genome. Genome Biol 3 doi: 10.1186 / gb-2002-3-10-research0052 (2002).24. Penzkofer, T. et al. LIBase 2: more retrotransposition-active LINE-ls, more mammalian genomes. Nucleic Acids Res 45, D68-D73, doi: 10.1093 / nar / gkw925 (2017).25. Cerami, E. et al. The eBio Cancer Genomics Portal: An Open Platform for Exploring Multidimensional Cancer Genomics Data. Cancer Discov 2, 401-404, doi: 10.1158 / 2159- 8290.Cd-12-0095 (2012).26. Gao, J. I. et al. Integrative Analysis of Complex Cancer Genomics and Clinical Profiles Using the cBioPortal. Sci Signal 6, doi: 10.1126 / sci signal.2004088 (2013).27. Tiwari, B., Jones, A. E. & Abrams, J. M. Transposons, p53 and Genome Security. Trends Genet 34, 846-855, doi: 10.1016 / j .tig.2018.08.003 (2018).28. Ardeljan, D. et al. Cell fitness screens reveal a conflict between LINE-1 retrotransposition and DNA replication. Nat Struct Mol Biol 27, 168-178, doi:10.1038 / s41594- 020-0372-1 (2020).29. Harris, C. R. et al. p53 responsive elements in human retrotransposons. Oncogene 28, 3857-3865, doi: 10.1038 / onc.2009.246 (2009).30. Valente, L. J. et al. p53 Efficiently Suppresses Tumor Development in the Complete Absence of Its Cell-Cycle Inhibitory and Proapoptotic Effectors p21, Puma, and Noxa. CellRep , 1339-1345, doi: 10.1016 / j.celrep.2013.04.012 (2013).31. Li, T. Y. et al. Tumor Suppression in the Absence of p53-Mediated Cell-Cycle Arrest, Apoptosis, and Senescence. Cell 149, 1269-1283, doi:10.1016 / j.cell.2012.04.026 (2012).32. Kaiser, A. M. & Attardi, L. D. Deconstructing networks of p53-mediated tumor suppression in vivo. Cell Death Differ 25, 93-103, doi:10.1038 / cdd.2017.171 (2018).33. Shukla, R. et al. Endogenous Retrotransposition Activates Oncogenic Pathways in Hepatocellular Carcinoma. Cell 153, 101-111, doi: 10.1016 / j.cell.2013.02.032 (2013).34. Faulkner, G. J. et al. The regulated retrotransposon transcriptome of mammalian cells. Nat Genet 41, 563-571, doi:10.1038 / ng.368 (2009).35. Han, J. S., Szak, S. T. & Boeke, J. D. Transcriptional disruption by the LI retrotransposon and implications for mammalian transcriptomes. Nature 429, 268-274, doi: 10.1038 / nature02536 (2004).36. McKerrow, W. et al. LINE-1 expression in cancer correlates with p53 mutation, copy number alteration, and S phase checkpoint. P Natl Acad Sci USA 119, doi:ARTN e2115999119 10.1073 / pnas.2115999119 (2022).37. Della Valle, F. et al. LINE-1 RNA causes heterochromatin erosion and is a target for amelioration of senescent phenotypes in progeroid syndromes. Sci Transl Med 14, doi: 10.1126 / scitranslmed.abl6057 (2022).38. Lindholm, H. T., Chen, R. & De Carvalho, D. D. Endogenous retroelements as alarms for disruptions to cellular homeostasis. Trends Cancer 9, 55-68, doi: 10.1016 / j.trecan.2022.09.001 (2023).39. Zhou, W. et al. Identification and characterization of occult human-specific LINE-1 insertions using long-read sequencing technology. Nucleic Acids Res 48, 1146-1163, doi: 10.1093 / nar / gkzl l73 (2020).40. Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows- Wheeler transform. Bioinformatics 26, 589-595, doi:10.1093 / bioinformatics / btp698 (2010).41. Broad Institute. Picard Tools version 2.27.4. Broad Institute, GitHub Repository (2022).42. Levin, H. L. & Moran, J. V. Dynamic interactions between transposable elements and their hosts. Nat Rev Genet 12, 615-627, doi:10.1038 / nrg3030 (2011).43. Kazazian, H. H., Jr. & Moran, J. V. The impact of LI retrotransposons on the human genome. Nat Genet 19, 19-24, doi: 10.1038 / ng0598-19 (1998).44. Ostertag, E. M. & Kazazian, H. H., Jr. Twin priming: a proposed mechanism for the creation of inversions in LI retrotransposition. Genome Res 11, 2059-2065, doi: 10.1101 / gr.205701 (2001).45. Ostertag, E. M. & Kazazian, H. H., Jr. Biology of mammalian LI retrotransposons. Annu Rev Genet 35, 501-538, doi : 10.1146 / annurev. genet.35.102401.091032 (2001).46. Richardson, S. R. etal. The Influence of LINE-1 and SINE Retrotransposons on Mammalian Genomes. Microbiol Spectr 3, MDNA3-0061-2014, doi : 10.1128 / microbiolspec.MDNA3 -0061-2014 (2015).47. Li, W. & Godzik, A. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22, 1658-1659, doi : 10.1093 / bioinformatics / btl 158 (2006).48. Fu, L., Niu, B., Zhu, Z., Wu, S. & Li, W. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics 28, 3150-3152, doi : 10.1093 / bioinformatics / bts565 (2012).49. Frith, M. C., Wan, R. & Horton, P. Incorporating sequence quality data into alignment improves DNA read mapping. Nucleic Acids Res 38, doi:ARTN elOO 10.1093 / nar / gkq010 (2010).50. Hubley, R. et al. The Dfam database of repetitive DNA families. Nucleic Acids Res 44, D81-D89, doi: 10.1093 / nar / gkv!272 (2016).51. Faulkner, G. J. Elevated LI expression in ataxia telangiectasia likely explained by an RNA- seq batch effect. Neuron 111, 610-611, doi: 10.1016 / j. neuron.2023.02.007 (2023).52. Graubert, A., Aguet, F., Ravi, A., Ardlie, K. G. & Getz, G. RNA-SeQC 2: efficient RNA-seq quality control and quantification for large cohorts. Bioinformatics 37, 3048- 3050, doi:10.1093 / bioinformatics / btabl35 (2021).53. Grossman, R. L. et al. Toward a Shared Vision for Cancer Genomic Data. N EnglJMed lS, 1109-1112, doi: 10.1056 / NEJMp 1607591 (2016).54. Menard, S. Six approaches to calculating standardized logistic regression coefficients. Am Stat 58, 218-223, doi: 10.1198 / 000313004x946 (2004).EQUIVALENTS

[0163] The present technology is not to be limited in terms of the particular embodiments described in this application, which are intended as single illustrations of individual aspects of the present technology. Many modifications and variations of this present technology can be made without departing from its spirit and scope, as will be apparent to those skilled in the art. Functionally equivalent methods and apparatuses within the scope of the present technology, in addition to those enumerated herein, will be apparent to those skilled in the art from the foregoing descriptions. Such modifications and variations are intended to fall within the scope of the present technology. It is to be understood that this present technology is not limited to particularmethods, reagents, compounds compositions or biological systems, which can, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting.

[0164] In addition, where features or aspects of the disclosure are described in terms of Markush groups, those skilled in the art will recognize that the disclosure is also thereby described in terms of any individual member or subgroup of members of the Markush group.

[0165] As will be understood by one skilled in the art, for any and all purposes, particularly in terms of providing a written description, all ranges disclosed herein also encompass any and all possible subranges and combinations of subranges thereof. Any listed range can be easily recognized as sufficiently describing and enabling the same range being broken down into at least equal halves, thirds, quarters, fifths, tenths, etc. As a non-limiting example, each range discussed herein can be readily broken down into a lower third, middle third and upper third, etc. As will also be understood by one skilled in the art all language such as “up to,” “at least,” “greater than,” “less than,” and the like, include the number recited and refer to ranges which can be subsequently broken down into subranges as discussed above. Finally, as will be understood by one skilled in the art, a range includes each individual member. Thus, for example, a group having 1-3 cells refers to groups having 1, 2, or 3 cells. Similarly, a group having 1-5 cells refers to groups having 1, 2, 3, 4, or 5 cells, and so forth.

[0166] All patents, patent applications, provisional applications, and publications referred to or cited herein are incorporated by reference in their entirety, including all figures and tables, to the extent they are not inconsistent with the explicit teachings of this specification.

Claims

WHAT IS CLAIMED IS:

1. A method of selecting cancer therapies based on magnitudes and frequencies of retrotransposon (RT) insertions in subject genomes, comprising: obtaining, by one or more processors, for a subject at risk of or diagnosed with cancer, a first dataset identifying genomic sequences derived from a first biological sample of the subject, wherein the genomic sequences are obtained via next-generation sequencing and comprise a plurality of sequencing reads; identifying, by the one or more processors, using the first dataset, (i) an RT burden metric corresponding to a RT indicating a level of RT activity in the genomic sequences of the subject and (ii) a plurality of p53 sequences in the genomic sequences of the subject, each of the plurality of p53 sequences including a plurality of non-silent p53 mutant sequences and a plurality of wild-type p53 sequences; for each p53 sequence of the plurality of p53 sequences: identifying, by the one or more processors, a mutated position of each p53 mutant sequence of the plurality of non-silent p53 mutant sequences; generating, by the one or more processors, a wild-type p53 fitness metric based on transactivation levels of at least one p53 target gene induced by a wild-type p53 polypeptide encoded by the plurality of wild-type p53 sequences, a binding affinity of the wild-type p53 polypeptide to a promoter region of the at least one p53 target gene, and expression levels of the wild-type p53 polypeptide; generating, by the one or more processors, maximum p53 fitness metric based on transactivation levels of the at least one p53 target gene caused by a complete deletion of p53 gene; determining, by the one or more processors, a mutant p53 fitness metric based on the wild-type p53 fitness metric, the maximum p53 fitness metric, and the mutated position of each p53 mutant sequence of the plurality of non-silent p53 mutant sequences; determining, by the one or more processors, for each p53 mutant sequence in the plurality of p53 sequences, a score indicating a degree of correlation between a RT burden and mutant p53 fitness, as a function of the RT burden metric and the mutant p53 fitness metric;selecting, by the one or more processors, from a plurality of cancer therapies, a cancer therapy for the subject based on the score; and storing, by the one or more processors, using one or more data structures, an association between the subject and the cancer therapy.

2. The method of claim 1, wherein the RT activity comprises canonical insertions, inversion-containing insertions, 5’ truncated insertions, full-length insertions, and 3’ transductions.

3. The method of claim 1 or 2, wherein the at least one p53 target gene is WAF1, MDM2, BAX, hl433s, AIP1, GADD45, NOXA, or P53R2.

4. The method of any one of claims 1-3, wherein the plurality of non-silent p53 mutant sequences comprise missense mutations, nonsense mutations, and frameshift mutations.

5. The method of any one of claims 1-4, further comprising: obtaining, by the one or more processors, a second dataset identifying Ribonucleic acid (RNA) sequences from a second biological sample of the subject; determining, by the one or more processors, using the second dataset, an RT RNA metric indicating expressions levels of the RT in the subject; and wherein determining the score further comprises determining, for each p53 mutant sequence in the plurality of p53 sequences, the score indicating the degree of correlation among at least two of the RT burden, RT RNA expression, and the mutant p53 fitness, as a function of at least two of the RT burden metric, the RT RNA metric, and the mutant p53 fitness metric.

6. The method of claim 5, wherein determining the score further comprises determining the score indicating the degree of correlation between the RT burden and the RT RNA expression, as a function of the RT burden metric and the RT RNA expression metric.

7. The method of claim 5, wherein determining the score further comprises determining the score indicating the degree of correlation between the RT RNA expression and the mutant p53 fitness, as a function of the RT RNA expression metric and the mutant p53 fitness metric.

8. The method of claim 5, wherein determining the score further comprises determining the score indicating the correlation of the RT burden with respect to the RT RNA expression and the mutant p53 fitness, as a function of the RT burden metric, the RT RNA metric, and the mutant p53 fitness metric.

9. The method of claim 5, wherein determining the score further comprises determining the score indicating the correlation of the RT RNA expression with respect to the RT burden and the mutant p53 fitness, as a function of the RT burden metric, the RT RNA metric, and the mutant p53 fitness metric.

10. The method of any one of claims 1-9, wherein the function comprises a regression model of the RT burden metric and the mutant p53 fitness metric.

11. The method of any one of claims 1-10, further comprising detecting, by the one or more processors, using the first dataset, the RT activity corresponding to at least one of a long terminal repeat (LTR) retrotransposon, a non-LTR retrotransposon, or deoxyribonucleic acid (DNA) retrotransposon in the genomic sequences of the subject.

12. The method of any one of claims 1-11, further comprising detecting, by the one or more processors, using the first dataset, the RT activity corresponding to at least one of a long interspersed nuclear element-1 (LINE-1), a short interspersed nuclear element (SINE), or a SINE- Variable number of tandem repeat (VNTR)-Alus (SVA).

13. The method of any one of claims 1-12, further comprising identifying, by the one or more processors, using the first dataset, a plurality of RT insertion sites using soft- clipped reads and / or discordant read pairs in the plurality of sequencing reads as a signal.

14. The method of claim 13, wherein identifying the RT burden metric further comprises generating the RT burden metric based on the plurality of RT insertion sites identified using soft-clipped reads and / or discordant read pairs in the plurality of sequencing reads.

15. The method of any one of claims 13-14, further comprising identifying, by the one or more processors, using the first dataset, at each insertion site of the plurality of RT insertion sites, an orientation of a corresponding RT sequence.

16. The method of claim 15, wherein the orientation comprises a single-stranded priming integration or an inversion-containing integration.

17. The method of claim 15 or 16, wherein identifying the RT burden metric further comprises generating the RT burden metric based on the orientation at each insertion site.

18. The method of any one of claims 1-17, further comprising removing, by the one or more processors, from the first dataset, a subset corresponding to a portion of genomic sequences based on at least one of: (i) lack of matched pair, (ii) duplicity, (iii) a mapping quality score not satisfying a threshold, or (iv) inconsistent base and quality score.

19. The method of any one of claims 1-17, further comprising removing, by the one or more processors, from the first dataset, a subset corresponding to a portion of genomic sequences based on a chimeric read fraction exceeding a threshold.

20. The method of any one of claims 1-19, wherein the first biological sample and / or the second biological sample comprise tissue from lung, head, neck, colon, uterus, stomach, ovary, cervix, bladder, or breast.

21. The method of any one of claims 1-20, wherein the subject is diagnosed with or suffers from a cancer.

22. The method of claim 21, wherein the cancer is selected from among lung cancer, brain cancer, head and neck cancer, colon cancer, rectal cancer, uterine cancer, endometrial cancer, stomach cancer, ovarian cancer, cervical cancer, bladder cancer, and breast cancer.

23. The method of any one of claims 1-22, wherein the cancer therapy comprises one or more of immune checkpoint inhibitor therapy, adoptive cell therapy, or prophylactic cancer vaccine therapy.

24. A system comprising one or more processors coupled with memory configured to perform any one or more of the methods in the preceding claims.

Images