Gene fusions and gene variants associated with cancer

By providing specific probes and kits, identifying and detecting gene fusions and variants, the inaccuracy of cancer diagnosis in the prior art is solved, and the accurate identification of cancer samples and personalized treatment is achieved.

CN118910253APending Publication Date: 2024-11-08LIFE TECHNOLOGIES CORP

Patent Information

Application Number
CN202410707725.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2014-02-14
Filing Date
2014-03-14
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The prior art is difficult to effectively identify and diagnose gene fusions and gene variants in specific cancers, which affects the accuracy and effectiveness of cancer treatment.

Method used

Specific probes and kits are provided for identification and detection of gene fusions listed in Tables 1-3, 19 and 22 and gene variants listed in Tables 7 and 11, combining polymerase, reverse transcriptase and Uracil DNA deglycosylase can accurately identify genetic abnormalities in cancer samples through PCR analysis and other nucleic acid detection technologies.

Benefits of technology

It improves the diagnostic accuracy and targeted treatment of specific cancers, and can identify specific gene fusions and variants in the sample, indicate the existence of cancer, and provide a basis for personalized treatment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118910253A_ABST
    Figure CN118910253A_ABST
Patent Text Reader

Abstract

The present invention provides gene fusions, gene variants, and novel correlations to disease conditions, as well as kits, probes, and methods of using the same.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This application is a divisional application of CN201480034585.7. Technical Field

[0002] The present invention generally relates to gene fusions and gene variants associated with cancer. Background Art

[0003] Abnormalities such as chromosomal translocations and gene variants are often found in human cancer cells. Chromosomal translocations may produce chimeric genes that express fusion transcripts, which are then translated into fusion proteins that affect normal regulatory pathways and stimulate cancer cell growth. Gene variants can also produce abnormal proteins that affect normal regulatory pathways.

[0004] The identification of novel fusion genes, novel variants of known fusion genes, and gene variants or alleles provides opportunities for additional diagnostic and cancer therapeutic targets. Summary of the invention

[0005] The present invention provides novel gene fusion variants and gene fusion-disease condition correlations. The gene fusions provided herein are associated with specific cancers. The present invention further provides probes (such as amplification primer sets and detection probes) and methods and systems for detection, diagnosis and treatment, and kits including or detecting the gene fusions disclosed herein.

[0006] In one embodiment, the present invention provides a reaction mixture comprising a probe or a probe set that specifically recognizes a gene fusion selected from Tables 1-3, 19, and 22. The probe set can be, for example, an amplification primer set. In another embodiment, the present invention provides a reaction mixture comprising a primer set flanking a gene fusion selected from Tables 1-3, 19, and 22 in a target nucleic acid. For example, the primer set can each bind to a target sequence in the human genome within 1000, 750, 500, 250, 100, 90, 80, 75, 70, 65, 50, or 25 nucleotides on opposite sides of one of the fusion breakpoints identified in Tables 4-6, 20, and 23. The reaction mixture of this embodiment can further include a detector probe that binds to either side of a breakpoint in a gene fusion selected from Tables 1-3, 19, and 22, or to a binding region spanning a breakpoint in a gene fusion selected from Tables 1-3, 19, and 22, including specific embodiments in which the breakpoints are identified in Tables 4-6, 20, and 23. In exemplary embodiments, the detector probe binds to a target sequence in the human genome within 1000, 750, 500, 250, 100, 90, 80, 75, 70, 60, 50, or 25 nucleotides of one of the fusion breakpoints identified in Tables 4-6, 20, and 23. The reaction mixture including or not including the detector probe can further include a polymerase, a reverse transcriptase, dNTPs, and / or uracil DNA deglycosylase (UDG). The polymerase, reverse transcriptase, and UDG are typically not from human sources. The polymerase in the illustrative embodiments is a thermostable polymerase, such as Taq polymerase. In certain embodiments, the dNTP in the reaction mixture includes dUTP, and in certain instances, the reaction mixture may not have dTTP. In addition, the reaction mixture may include an amplicon, such as a DNA amplicon, which includes one or more deoxyuridine ("dU") residues. In certain embodiments, the reaction mixture includes a DNA amplicon, which includes one or more dU residues for each deoxythymidine residue in the corresponding human genome sequence. In certain embodiments, the amplicon includes a section that the corresponding sequence is not found in the human genome, such as a DNA barcode sequence. The length of the non-human section may be, for example, 5-10,000, 5-5000, 5-1000, 5-500, 5-100, 5-50, 5-25, 5-10, 10-10,000, 10-5000, 10-1000, 10-500, 10-100, 10-50 or 10-25 nucleotides. In certain embodiments, the amplicon includes a segment corresponding to a region spanning an intron in the human genome, but the amplicon does not include a segment corresponding to an intron.The reaction mixture can further include a target nucleic acid, such as a human target nucleic acid.The human target nucleic acid can be isolated, for example, from a biological sample of an individual suspected of having a cancer selected from the group consisting of BLCA = bladder cancer, BRCA = breast cancer, CESC = cervical cell carcinoma, COAD = colon adenocarcinoma, GBM = glioblastoma multiforme, HNSC = head and neck squamous cell carcinoma, KIRK = clear cell renal cell carcinoma, KIRP = kidney renal papillary cell carcinoma, LAML = acute myeloid leukemia, LGG = brain low-grade glioma, LIHC = liver hepatocellular carcinoma, LUAD = lung adenocarcinoma, LUSC = squamous cell lung cancer, OV = ovarian serous adenocarcinoma, PRAD = prostate adenocarcinoma, READ = rectal adenocarcinoma, SKCM = skin melanoma, STAD = gastric adenocarcinoma, THCA = thyroid carcinoma, and UCEC = uterine corpus endometrioid carcinoma. In certain embodiments, the target nucleic acid is from a tumor, such as a tumor of one of the cancer types listed in the previous sentence.

[0007] In another embodiment, a probe set is provided that specifically recognizes a nucleic acid comprising at least one of SEQ ID NOs: 1-257 (gene fusion). In another embodiment, a primer set is provided herein that specifically amplifies a target nucleic acid comprising at least 25, 30, 40, 50, 75, 100, 125, 150, 200, or all of SEQ ID NOs: 1-257, or amplifies up to 25, 30, 40, 50, 75, 100, 125, 150, 200, or all of SEQ ID NOs: 1-257. In another embodiment, a qPCR assay is provided herein, such as TaqMan TM Analysis or MolecularBeacons TM An assay that specifically amplifies and detects target nucleic acids comprising at least 25, 30, 40, 50, 75, 100, 125, 150, 200 or all of SEQ ID NOs: 1-257.

[0008] The present invention also provides an isolated nucleic acid comprising at least one sequence selected from a segment comprising at least 25, 30, 40, 50, 75, 100, 125, 150, 200 or all of SEQ ID NOs: 1-257 or comprising up to 25, 30, 40, 50, 75, 100, 125, 150, 200 or all of SEQ ID NOs: 1-257. The isolated nucleic acid may include a first primer at the 5' end. In addition, the nucleic acid may be single-stranded or double-stranded. In certain embodiments, the isolated nucleic acid includes a segment whose corresponding sequence is not found in the human genome, such as a DNA barcode sequence. A segment can be, for example, 5-10,000, 5-5000, 5-1000, 5-500, 5-100, 5-50, 5-25, 5-10, 10-10,000, 10-5000, 10-1000, 10-500, 10-100, 10-50, or 10-25 nucleotides in length.

[0009] In other embodiments, the present invention provides a kit comprising a detector probe and / or a probe set, such as an amplification primer set, which specifically recognizes a nucleic acid comprising a breakpoint of a gene fusion selected from Tables 1-3, Table 19, and Table 22. For example, in certain embodiments, the detector probe or amplification primer set is designed to amplify and / or detect nucleic acids up to 25, 30, 40, 50, 75, 100, 125, 150, 200, or all of at least one of SEQ ID NOs: 1-257. The kit may further include at least one component from an amplification reaction mixture in one or more separate containers or in the same container, such as a polymerase, dNTP, reverse transcriptase, and / or UDG, typically the reverse transcriptase, polymerase, and UDG are not from human sources. In certain embodiments, the dNTPs include dUTP, and in illustrative examples, there is no dTTP. The polymerase in the illustrative embodiments is a thermostable polymerase, such as Taq polymerase. In addition, the kit may include a control nucleic acid. For example, the control nucleic acid can include a sequence comprising a breakpoint in a gene fusion selected from Tables 1-3, Table 19, and Table 22, such as a nucleic acid comprising at least 25, 30, 40, 50, 75, 100, 125, 150, 200 or all of SEQ ID NOs: 1-257 or a nucleic acid comprising up to 25, 30, 40, 50, 75, 100, 125, 150, 200 or all of SEQ ID NOs: 1-257.

[0010] A method for detecting cancer is provided, comprising amplifying a nucleic acid spanning a breakpoint in a gene fusion selected from Tables 1-3, Tables 19 and 22, for example, the nucleic acid may include a sequence selected from SEQ ID NO: 1-257; and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that the cancer is present in the sample. In another method, a method for detecting a cancer selected from bladder cancer, colon cancer, breast cancer, endometrial cancer, melanoma, ovarian cancer, glioblastoma, glioma, leukemia, renal cell carcinoma, thyroid cancer and prostate adenocarcinoma is provided herein, comprising generating an amplicon comprising a sequence selected from SEQ ID NO: 1-257, and detecting the presence of the amplicon, wherein the presence of the amplicon indicates that bladder cancer, colon cancer, melanoma, ovarian cancer, glioblastoma, lung cancer, glioma, leukemia, renal cell carcinoma, thyroid cancer, endometrial endometrioid adenocarcinoma, breast cancer and prostate adenocarcinoma are present in the sample. The amplicon typically includes a primer that is extended to form the amplicon. The cancer is selected from bladder urothelial carcinoma, breast cancer, endometrial endometrioid adenocarcinoma, colon adenocarcinoma, multiforme glioblastoma, clear cell renal cell carcinoma, papillary renal cell carcinoma, acute myeloid leukemia, brain low-grade glioma, lung adenocarcinoma, ovarian serous cystadenocarcinoma, prostate adenocarcinoma, rectal skin melanoma and thyroid gland cancer. In a specific illustrative embodiment, the amplicon produced is a DNA amplicon, which includes dU residues, and in some instances does not include dT residues. In the method provided in this paragraph, the amplicon can be produced using the reaction mixture provided herein. In certain embodiments, the method includes detecting the expression of a nucleic acid spanning a breakpoint in a gene fusion selected from Table 1-Table 3, Table 19 and Table 22. The method for detecting expression generally includes a step of separating RNA from a sample (such as a tumor sample, which can be a sample fixed with formalin in an illustrative embodiment).

[0011] In one embodiment, the reaction mixture includes a dye selected from SYBR Green, SBYR Greener, Fluorescein, Oregon Green, FAM, TET, JOE, VIC, Yakima Yellow, HEX, Cy3, Bodipy TMR, NED, TAMRA, Cy3.5, ROX, Texas Red, LightCycler Red, Bodipy 630 / 650, Alexa Fluor 647, Cy5, Alexa Fluor 660, or Cy 5.5. In certain embodiments, the dye is connected to a detectably labeled probe in the reaction mixture. In other embodiments, the dye is directly or via a detectably labeled probe bound to an amplicon.

[0012] A kit is provided, comprising a probe or a probe set, such as a detectable probe or an amplification primer set, that specifically recognizes a nucleic acid comprising a breakpoint from Tables 4-6, 20, and 23. The kit may further include components from an amplification reaction mixture, such as a polymerase (usually not from human sources), dNTPs, and / or UDG, in the same container or in certain preferred embodiments in another container. In addition, the kit may include a control nucleic acid. For example, a control nucleic acid may include a sequence comprising a breakpoint selected from Tables 4-6, 20, and 23.

[0013] In another embodiment, provided herein is a gene fusion comprising a gene fusion identified in Tables 1-3, 19, and 22. In illustrative embodiments, the gene fusion comprises one of the breakpoints identified in Tables 4-6, 20, and 23. Thus, provided herein is an isolated gene fusion nucleic acid that is between 100 and 10,000 nucleotides in length and comprises at least 25 nucleotides on either side of one of the breakpoints in Tables 4-6, 20, and 23.

[0014] In a related embodiment, provided herein is an isolated gene fusion nucleic acid comprising at least one of the breakpoints in Tables 4-6, 20, and 23. In certain embodiments, the isolated gene fusion nucleic acid comprises at least 25, 30, 40, 50, 75, 100, 125, 150, 200, or all of SEQ ID NOs: 1-257 or includes up to 25, 30, 40, 50, 75, 100, 125, 150, 200, or all of SEQ ID NOs: 1-257. The length of isolated gene fusion nucleic acid can be for example between 50 and 100,000 nucleotides, between 100 and 50,000 nucleotides, between 100 and 25,000 nucleotides, between 100 and 10,000 nucleotides, between 100 and 5,000 nucleotides, between 100 and 2500 nucleotides, between 100 and 1,000 nucleotides, between 100 and 500 nucleotides, between 100 and 250 nucleotides, between 100 and 200 nucleotides, between 250 and 10,000 nucleotides, between 250 and 5,000 nucleotides, between 250 and 1,000 nucleotides or between 250 and 500 nucleotides. In some aspects, isolated gene fusion nucleic acid is DNA. In some illustrative embodiments, isolated nucleic acid gene fusion does not have intron sequences but spans the region in the genome comprising one or more introns. In certain embodiments, the isolated gene fusion nucleic acid is cDNA.

[0015] In another embodiment, an isolated gene fusion nucleic acid is provided that comprises at least one of the breakpoints in Tables 4-6, 20, and 23.

[0016] In another embodiment is a method of detecting a cancer selected from bladder urothelial carcinoma, breast cancer, endometrial endometrioid adenocarcinoma, colon adenocarcinoma, glioblastoma multiforme, clear cell renal cell carcinoma, papillary renal cell carcinoma, acute myeloid leukemia, brain low-grade glioma, lung adenocarcinoma, ovarian serous cystadenocarcinoma, prostate adenocarcinoma, rectal skin melanoma, and thyroid gland cancer in a sample by detecting the presence of a gene fusion selected from Tables 1-3, Table 19, and Table 22.

[0017] The present invention provides novel gene variants and gene variant-disease condition correlations. Gene variants may have one or more mutations that produce variant proteins. Gene variants provided herein are associated with certain cancers. Gene variants produce protein variants. The present invention further provides probes (such as amplification primer sets and detection probes) and methods of detection, diagnosis and treatment and kits comprising or detecting the gene variants disclosed herein.

[0018] In one embodiment, the present invention provides a composition and a kit comprising a probe set that specifically recognizes a nucleotide sequence encoding a gene variant selected from Table 7 and / or Table 11. The probe set may, for example, be an amplification primer set. In another embodiment, a composition is provided herein, comprising a primer set flanking a gene variant encoding one or more variants in Table 7 and / or Table 11. The reaction mixture of this embodiment may further include a detector probe that is bound to a nucleotide sequence comprising a gene variant selected from Table 7 and / or Table 11. The reaction mixture including or not including the detector probe may further include a polymerase, dNTP and / or uracil DNA deglycosylase (UDG). Polymerase and UDG are generally not derived from human sources. The reaction mixture may further include a target nucleic acid, such as a human target nucleic acid. The human target nucleic acid may, for example, be isolated from a biological sample from an individual suspected of having cancer. The cancer may be selected from the group consisting of: BLCA = bladder cancer, BRCA = breast cancer, CESC = cervical cell carcinoma, COAD = colon adenocarcinoma, GBM = glioblastoma multiforme, HNSC = head and neck squamous cell carcinoma, KIRK = clear cell renal cell carcinoma, KIRP = kidney renal papillary cell carcinoma, LAML = acute myeloid leukemia, LGG = brain low-grade glioma, LIHC = liver hepatocellular carcinoma, LUAD = lung adenocarcinoma, LUSC = squamous cell lung cancer, OV = ovarian serous adenocarcinoma, PRAD = prostate adenocarcinoma, READ = rectal adenocarcinoma, SKCM = cutaneous melanoma, STAD = gastric adenocarcinoma, THCA = thyroid carcinoma, and UCEC = uterine corpus endometrioid carcinoma.

[0019] The nucleotide sequence encoding one or more gene variants in Table 7 and / or Table 11 can be any size that encompasses the variant. For example, the nucleotide sequence can be any size that can be easily copied using primers and / or detected using probes.

[0020] In another embodiment, a probe set is provided that specifically recognizes a nucleic acid encoding a gene variant selected from Table 7 and / or Table 11 (gene variants). In another embodiment, a primer set is provided herein that specifically amplifies a target nucleic acid encoding a gene variant selected from Table 7 and / or Table 11. In another embodiment, a qPCR assay is provided herein, such as (but not limited to) TaqMan TM Analysis, Scorpions Analysis or MolecularBeacons TM An assay that specifically amplifies and detects a target nucleic acid encoding a gene variant selected from Table 7 and / or Table 11.

[0021] The present invention also provides an isolated nucleic acid comprising at least one sequence encoding one or more gene variants selected from Table 7 and / or Table 11. The isolated nucleic acid may include a first primer at the 5' end. In addition, the nucleic acid may be single-stranded or double-stranded.

[0022] In other embodiments, the present invention provides a kit comprising a detector probe and / or a probe set, such as an amplification primer set, which specifically recognizes a nucleic acid encoding a gene variant selected from Table 7 and / or Table 11. For example, in certain embodiments, the detector probe or amplification primer set is designed to amplify and / or detect nucleic acids encoding variants in Table 7 and / or Table 11. The kit may further include components from an amplification reaction mixture in another container or in the same container, such as a polymerase (usually not from a human source), dNTPs and / or UDG. In addition, the kit may include a control nucleic acid. For example, the control nucleic acid may include a sequence comprising a gene variant selected from Table 7 and / or Table 11.

[0023] A method for detecting cancer is provided, comprising amplifying a nucleic acid encoding a gene variant selected from Table 7 and / or Table 11, for example, the nucleic acid may include a sequence from one of the accession numbers in Table 7 and / or Table 11, except that the sequence contains a variant encoding a gene variant in Table 7 and / or Table 11; and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that cancer is present in the sample. In another method, a method for detecting cancer is provided herein, comprising generating an amplicon comprising a sequence encoding a variant selected from Table 7 and / or Table 11, and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that cancer or cancer cells are present in the sample. The amplicon typically includes a primer that is extended to form the amplicon. The cancer is selected from bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, clear cell renal cell carcinoma, kidney renal papillary cell carcinoma, acute myeloid leukemia, brain low-grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, squamous cell lung cancer, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer and uterine corpus endometrioid carcinoma.

[0024] A kit is provided, comprising a probe set (e.g., an amplification primer set) that specifically recognizes a nucleic acid comprising a gene variant from Table 7 and / or Table 11. The kit may further include components from an amplification reaction mixture in another container or in the same container, such as a polymerase (usually not from a human source), dNTPs, and / or UDG. In addition, the kit may include a control nucleic acid. For example, a control nucleic acid may include a sequence comprising a gene variant from Table 7 and / or Table 11.

[0025] In certain embodiments, a probe set is provided that specifically recognizes a nucleic acid comprising a gene variant from Table 7 and / or Table 11.

[0026] In another embodiment, a genetic variant is provided, comprising at least one of the genetic variants in Table 7 and / or Table 11.

[0027] In another embodiment is a method of detecting a cancer selected from bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, clear cell renal cell carcinoma, kidney renal papillary cell carcinoma, acute myeloid leukemia, brain low grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, squamous cell lung carcinoma, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer, and uterine corpus endometrioid carcinoma in a sample by detecting the presence of a genetic variant selected from Table 7 and / or Table 11. The genetic variants may include, but are not limited to, ZNF479 variants R11Q, R295K, R295T, R295I, R345I, R345T, K438T, and T466K.

[0028] In another embodiment, a method for delivering a drug to a subject is provided, wherein the method comprises a gene event identified in detection table 15, and treats the subject with a drug, wherein the drug is considered to actively affect the clinical outcome of the patient with the gene event. In an illustrative embodiment, the gene event is related to the gene found in table 8, and the drug is listed as the partner of the gene in table 8. In another embodiment, a method for determining whether a subject accepts a drug is provided herein, wherein the method comprises a gene event identified in detection table 15, and then if the detected gene event is listed as being related to a poor prognosis in table 15, then the drug is delivered to the subject, wherein the drug is considered to actively affect the clinical outcome of the patient with the gene event. In an illustrative embodiment, the gene event is related to the gene found in table 8, and the drug is listed as the partner of the gene in table 8.

[0029] In one embodiment, a kit is provided, wherein the kit comprises a probe set, wherein each probe specifically hybridizes to a nucleic acid comprising a breakpoint from Tables 4-6, 20, and 23.

[0030] In one embodiment, a method is provided, comprising: amplifying a nucleic acid from a sample comprising at least one gene fusion from Tables 1-3, 19, and 22; and detecting the presence of the at least one gene fusion by at least one of the following: contacting the composition with at least one probe, wherein each probe specifically hybridizes to the nucleic acid, or observing the presence of non-natural or non-native chemical structures in the nucleic acid; wherein detecting the presence of the at least one gene fusion indicates that at least one cancer from Tables 1-3, 19, and 22 is present in the sample.

[0031] In one embodiment, a system is provided, comprising a nucleic acid amplifier configured to amplify nucleic acid from a sample comprising at least one gene fusion from Tables 1-3, 19 and 22 to obtain an amplified nucleic acid; a detector configured to detect the presence of the at least one gene fusion in the amplified nucleic acid by at least one of the following: (i) contacting a composition with at least one probe, wherein each probe specifically hybridizes to the nucleic acid, or (ii) observing the presence of non-natural or non-native chemical structures in the nucleic acid, and further configured to send a detection indication; and a computer system configured to receive the detection indication and determine, based on the detection indication, that at least one cancer from Tables 1-3, 19 and 22 is present in the sample.

[0032] In one embodiment, a non-transitory computer-readable program storage unit is provided, the non-transitory computer-readable program storage unit being encoded with instructions, which, when executed by a computer, perform a method, the method comprising: receiving an input comprising at least one cancer type and an event type, wherein the cancer type is selected from Table 15 and the event type is selected from Table 15; querying a database for at least one entry comprising a plurality of fields, wherein the plurality of fields comprises at least one of the cancer type and the event type; and sending an output comprising at least one field of the plurality of fields from the at least one entry, wherein the at least one field comprises at least one gene, at least one druggable gene, at least one drug targeting the at least one druggable gene, or a prognosis.

[0033] In one embodiment, a method is provided, wherein the method comprises administering to a patient having at least one gene fusion selected from the gene fusions listed in Tables 1-3, 19, and 22 at least one drug selected from the drugs listed in Tables 8, 16-17, 21, and 24.

[0034] In one embodiment, a method is provided, wherein the method comprises contacting a nucleic acid sample from a patient with a reaction mixture comprising a first primer complementary to a first gene and a second primer complementary to a second gene, wherein a fusion of the first gene and the second gene is detectable by the presence of an amplicon generated by the first primer and the second primer, wherein the fusion comprises a breakpoint selected from the breakpoints listed in Tables 4-6, 20, and 23.

[0035] In one embodiment, a non-transitory computer-readable program storage unit is provided, the non-transitory computer-readable program storage unit being encoded with instructions, which when executed by a computer perform a method, the method comprising: receiving RNA sequence data from at least one cancer cell line; running at least one gene fusion caller on the sequence data to identify possible breakpoints between fusion genes in the processed data; filtering the possible breakpoints to retain candidate breakpoints, wherein each candidate breakpoint is in a 5' untranslated region (UTR) or a coding DNA sequence (CDS) of a functional gene region, and each candidate breakpoint does not occur in an intron; annotating the candidate breakpoints with at least one annotation suitable for determining the relevance of the gene fusion for at least one of cancer diagnosis, cancer prognosis, or cancer treatment, wherein the gene fusion comprises the candidate breakpoint.

[0036] In one embodiment, a non-transitory computer-readable program storage unit is provided, the non-transitory computer-readable program storage unit being encoded with instructions, the instructions when executed by a computer performing a method, the method comprising: receiving mutation data from at least one cancer cell line; annotating the mutation data with at least one of variant classification, variant position, or variant change to obtain annotated mutation data; filtering the annotated mutation data to obtain gene region mutation data; classifying the gene region mutation data as hotspot, deleterious, or other; and naming a gene comprising the gene region mutation as gain of function, loss of function, or recurrent other gene based on the relative frequency of mutations in the gene and the classification of all gene region mutations in the gene.

[0037] In one embodiment, a method is provided, the method comprising detecting one or more gene fusions in a sample from a subject to obtain gene fusion detection data, wherein at least one of the gene fusions is selected from the gene fusions listed in Tables 1-3, 19 and 22, receiving the gene fusion detection data through a computer system, and confirming at least one treatment option recommended to the subject based on the gene fusion detection data through the computer system.

[0038] In one embodiment, a system is provided, comprising a detector configured to (i) detect one or more gene fusions in a sample from a subject to obtain gene fusion detection data, wherein at least one of the gene fusions is selected from the gene fusions listed in Tables 1-3, 19 and 22 and (ii) send the gene fusion detection data; and a computer system configured to receive the gene fusion detection data and confirm at least one treatment option recommended to the subject based on the gene fusion detection data.

[0039] In another embodiment, a novel TP53 WT gene signature and methods for detecting the expression level of one or more of the TP53 WT gene signature genes in Table 40 are provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 Provides a workflow for processing gene fusion RNASeq data.

[0041] Figure 2 A pipeline showing the classification of genetic variants for gain-of-function and loss-of-function genes.

[0042] Figure 3 An overview of the data flows that integrate various data types into the Genetic Event Database (GEDB).

[0043] Figure 4 This is a roll-up flow chart showing genetic events.

[0044] Figure 5 is a graph demonstrating that the TP53 WT expression signature is significantly elevated in TP53 WT breast cancers compared to breast cancer samples containing TP53 point mutations.

[0045] Figure 6 is a graph showing that the TP53 WT expression signature is significantly elevated in TP53 WT lung cancers compared to lung cancer samples containing TP53 mutations.

[0046] Figure 7 is a graph demonstrating that the TP53 WT expression signature is significantly elevated in HP53 WT ovarian cancer compared to ovarian cancer samples containing TP53 mutations.

[0047] Fig. 8A -D is a graph depicting crude RPKM expression values ​​(AB) versus z-score normalized values ​​(CD) for PLXNB21 and COL7A1 in serous ovarian cancer patients. Extensive population drops in PLXNB1 expression at exons 12, 17, and 23 are smoothed in the normalized data. Samples predicted to have fusions between these genes, with red diamonds indicating breakpoint exons predicted by the caller.

[0048] Figure 9 is a table of common TP53 mutations by amino acid position. Mutations that occur at an overall frequency of >0.25% in patients in the pan-cancer analysis are shown. Recurrent splice site mutations were identified at intron-exon junctions affecting T-125

[0049] Figure 10 is a table of Tp53 in-frame insertion and deletion mutations. The largest detected in-frame insertion-deletion identified was 21 bp. More than 99% of non-transposon insertions and deletions on the genome were <100 bp.

[0050] Fig.11 is an overview of methods to identify the most recurring peak regions and genes across multiple cancer types. DETAILED DESCRIPTION

[0051] The present invention provides novel gene fusions and variants, as well as novel associations of gene fusions and / or gene variants with certain cancer types. Further provided are probes, reaction mixtures, assays and kits involving the gene fusions and / or variants disclosed herein.

[0052] definition

[0053] The term "marker" or "biomarker" refers to a molecule (usually a protein, nucleic acid, carbohydrate or lipid) expressed in a cell, expressed on the surface of a cancer cell or secreted by a cancer cell compared to a non-cancerous cell, and is suitable for diagnosing cancer, for providing a prognosis, and for the preferred targeting of drugs to cancer cells. Often, the marker is a molecule that is overexpressed in a cancer cell compared to a non-cancerous cell, for example, 1-fold overexpression, 2-fold overexpression, 3-fold overexpression or more compared to a normal cell. In addition, the marker may be a molecule that is inappropriately synthesized in a cancer cell, for example, a molecule that contains a deletion, addition or mutation compared to a molecule expressed on a normal cell. Alternatively, such biomarkers are molecules that are underexpressed in a cancer cell compared to a non-cancerous cell, for example, 1-fold underexpression, 2-fold underexpression, 3-fold underexpression or more. In addition, the marker may be a molecule that is inappropriately synthesized in a cancer cell, for example, a molecule that contains a deletion, addition or mutation compared to a molecule expressed on a normal cell.

[0054] The skilled artisan will appreciate that the markers can be used in combination with other markers or tests for any of the uses disclosed herein (eg, prediction, diagnosis, or prognosis of cancer).

[0055] "Biological samples" include tissue sections, such as biopsy and autopsy samples, and frozen sections obtained for histological purposes. For example, biological samples can include fresh frozen paraffin-embedded (FFPE) samples. Alternatively, biological samples can include blood and blood fractions or products (e.g., serum, plasma, platelets, red blood cells, etc.), sputum, bronchoalveolar lavage fluid, cultured cells (e.g., primary cultures, explants, and transformed cells), feces, urine, etc. Biological samples are usually obtained from eukaryotic organisms, most preferably mammals, such as primates, such as chimpanzees or humans; cattle; dogs; cats; rodents, such as guinea pigs, rats, mice; rabbits; or birds; reptiles; or fish.

[0056] "Biopsy" refers to the process of removing a tissue sample for diagnosis or prognostic assessment, as well as the tissue sample itself. Any biopsy technique known in the art can be applied to the diagnostic and prognostic methods of the present invention. The biopsy technique used will depend on the type of tissue to be evaluated (e.g., lung, etc.), the size and type of the tumor, and other factors. Representative biopsy techniques include, but are not limited to, excisional biopsy, incisional biopsy, needle biopsy, surgical biopsy, and bone marrow biopsy. "Excisional biopsy" refers to removing the entire tumor mass with a small margin of normal tissue surrounding the tumor mass. "Incisional biopsy" refers to removing a wedge of tissue from the tumor. Diagnosis or prognosis guided by endoscopy or radiography may require a "core needle biopsy" or "fine needle aspiration biopsy," which typically obtains a cell suspension from the target tissue. Biopsy techniques are discussed, for example, in Harrison's Principles of Internal Medicine, edited by Kasper et al., 16th edition, 2005, Chapter 70, and throughout Part V.

[0057] The term "overexpression" or "overexpressed" refers interchangeably to proteins or nucleic acids (RNA) that are usually translated or transcribed at a detectably greater level in cancer cells compared to normal cells. The term includes overexpression due to transcription, post-transcriptional processing, translation, post-translational processing, cellular localization (e.g., organelles, cytoplasm, nucleus, cell surface) and RNA and protein stability compared to normal cells. Overexpression can be detected using conventional techniques for detecting mRNA (i.e., RT-PCR, PCR hybridization) or protein (i.e., ELISA, immunohistochemical techniques). Overexpression can be 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or more compared to normal cells. In some cases, overexpression is 1 times, 2 times, 3 times, 4 times or higher levels of transcription or translation compared to normal cells.

[0058] The term "underexpress / underexpression" or "underexpressed" or "downregulation" refers interchangeably to a protein or nucleic acid that is translated or transcribed at a detectably lower level in a cancer cell compared to a normal cell. The term includes underexpression due to transcription, post-transcriptional processing, translation, post-translational processing, cellular localization (e.g., organelles, cytoplasm, nucleus, cell surface), and RNA and protein stability compared to a control. Underexpression can be detected using conventional techniques for detecting mRNA (i.e., RT-PCR, PCR hybridization) or protein (i.e., ELISA, immunohistochemical techniques). Underexpression can be 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or less compared to a control. In some cases, underexpression is 1-fold, 2-fold, 3-fold, 4-fold or lower levels of transcription or translation compared to a control.

[0059] The term "differentially expressed" or "differentially regulated" generally refers to a protein or nucleic acid that is overexpressed (upregulated) or underexpressed (downregulated) in one sample (in the context of the present invention, typically in a cancer patient compared to a sample of non-cancerous tissue) compared to at least one other sample.

[0060] The term "system" describing a set of components (real or abstract) comprises a whole in which each component interacts with or is related to at least one other component within the whole.

[0061] The terms "polypeptide," "peptide," and "protein" are used interchangeably herein to refer to polymers of amino acid residues. These terms apply to amino acid polymers in which one or more amino acid residues is an artificial chemical mimetic of a corresponding naturally occurring amino acid, as well as to naturally occurring amino acid polymers and non-naturally occurring amino acid polymers.

[0062] The term "amino acid" refers to naturally occurring and synthetic amino acids, as well as amino acid analogs and amino acid mimetics that function in a manner similar to naturally occurring amino acids. Naturally occurring amino acids are amino acids encoded by the genetic code, as well as those amino acids that are later modified, such as hydroxyproline, γ-carboxyglutamate, and O-phosphoserine. Amino acid analogs refer to compounds that have the same basic chemical structure (i.e., an alpha carbon bound to a hydrogen, carboxyl, amino, and R groups) as naturally occurring amino acids, such as homoserine, norleucine, methionine sulfoxide, methionine methylsulfonium. Such analogs have modified R groups (e.g., norleucine) or modified peptide backbones, but retain the same basic chemical structure as naturally occurring amino acids. Amino acid mimetics refer to compounds that have a structure that is different from the general chemical structure of an amino acid but function in a manner similar to naturally occurring amino acids.

[0063] Amino acids may be referred to herein by either their commonly known three letter symbols or by the one-letter symbols recommended by the IUPAC-IUB Biochemical Nomenclature Commission. Likewise, nucleotides may be referred to by their commonly accepted single-letter codes.

[0064] With respect to amino acid sequences, the skilled artisan will recognize that individual substitutions, deletions or additions to a nucleic acid, peptide, polypeptide or protein sequence that alters, adds or deletes a single amino acid or a small percentage of amino acids in the encoding sequence are "conservatively modified variants," wherein the changes result in substitution of an amino acid with a chemically similar amino acid. Conservative substitution tables providing functionally similar amino acids are well known in the art. Such conservatively modified variants are in addition to and do not exclude polymorphic variants, interspecies homologs and alleles of the invention.

[0065] The following eight groups each contain amino acids that are conservative substitutions for each other: 1) alanine (A), glycine (G); 2) aspartic acid (D), glutamic acid (E); 3) asparagine (N), glutamine (Q); 4) arginine (R), lysine (K); 5) isoleucine (I), leucine (L), methionine (M), valine (V); 6) phenylalanine (F), tyrosine (Y), tryptophan (W); 7) serine (S), threonine (T); and 8) cysteine ​​(C), methionine (M). See, e.g., Creighton, Proteins (1984).

[0066] When referring to proteins, nucleic acids, antibodies or small molecules, the phrase "specifically (or selectively) binds" refers to a binding reaction that determines the presence of proteins or nucleic acids (such as differentially expressed genes of the present invention) in a heterogeneous population of proteins or nucleic acids and other biological agents. In the case of antibodies, under specified immunoassay conditions, the specified antibody can bind to a specific protein at least twice the background and more usually more than 10 to 100 times the background. Under such conditions, specific binding to antibodies requires antibodies selected for their specificity to specific proteins. For example, polyclonal antibodies can be selected to obtain only those polyclonal antibodies that specifically immunoreact with selected antigens but not with other proteins. This selection can be achieved by removing antibodies that cross-react with other molecules. A variety of immunoassay formats can be used to select antibodies that specifically immunoreact with specific proteins. For example, solid phase ELISA immunoassays are routinely used to select antibodies that specifically immunoreact with proteins (see, for example, Harlow and Lane, "Antibodies, A Laboratory Manual" (1988) for descriptions of immunoassay formats and conditions that can be used to measure specific immunoreactions).

[0067] In the context of an analysis of a test compound that modulates a marker protein, the phrase "functional effect" includes determining a parameter, such as a chemical or phenotype, that is indirectly or directly affected by the biomarker of the invention. Thus, functional effects include, among other things, ligand binding activity, transcriptional activation or inhibition, the ability of cells to proliferate, the ability to migrate. "Functional effect" includes in vitro, in vivo, and ex vivo activity.

[0068] "Determining a functional effect" means an analysis of a compound directed to increasing or decreasing a parameter that is indirectly or directly affected by a biomarker of the invention, such as measuring physical and chemical or phenotypic effects. Such functional effects can be measured by any means known to those skilled in the art, such as changes in spectral characteristics (e.g., fluorescence, absorbance, refractive index); hydrodynamics (e.g., shape), chromatography; or solubility properties of a protein; ligand binding analysis, such as binding to an antibody; measuring an inducible marker or transcriptional activation of a marker; measuring changes in enzyme activity; the ability to increase or decrease cell proliferation, apoptosis, cell cycle arrest, measuring changes in cell surface markers. Functional effects can be assessed by many means known to those skilled in the art, such as microscopy for quantitative or qualitative measurement of changes in morphological characteristics, measuring changes in RNA or protein levels of other genes expressed in placental tissue, measuring RNA stability, identifying downstream or reporter gene expression (CAT, luciferase, β-gal, GFP, etc.), such as via chemiluminescence, fluorescence, colorimetric reactions, antibody binding, inducible markers, etc.

[0069] "Inhibitors," "activators," and "modulators" of markers are used to refer to activating, inhibitory, or modulatory molecules identified using in vitro and in vivo analysis of cancer biomarkers. Inhibitors are compounds that, for example, bind to partially or completely block activity, reduce, prevent, delay activation, inactivate, passivate, or downregulate the activity or expression of a cancer biomarker. "Activators" are compounds that increase, turn on, initiate, promote, increase activation, sensitize, excite, or upregulate the activity of a cancer biomarker, such as agonists. Inhibitors, activators, or modulators also include genetically modified forms of cancer biomarkers (e.g., forms with altered activity) as well as naturally occurring and synthetic ligands, antagonists, agonists, antibodies, peptides, cyclic peptides, nucleic acids, antisense molecules, ribozymes, RNAi and siRNA molecules, small organic molecules, and the like. Such analysis for inhibitors and activators includes, for example, expressing cancer biomarkers in vitro, in cells or cell extracts, applying putative modulator compounds, and then determining the functional effects on activity, as described above.

[0070] "Probe" refers to a polynucleotide that is at least eight (8) nucleotides in length and that forms a hybrid structure with a target sequence due to complementarity of at least one sequence in the probe with a sequence in a target region. The polynucleotide may be composed of DNA and / or RNA. In certain embodiments, the probe is detectably labeled, as discussed in more detail herein. The size of the probe may vary significantly. In general, the length of the probe is, for example, at least 8 to 15 nucleotides. Other probes are, for example, at least 20, 30, or 40 nucleotides long. Other probes are slightly longer, being at least, for example, 50, 60, 70, 80, 90 nucleotides long. Other probes are still longer, and are at least, for example, 100, 150, 200, or more nucleotides long. The probe may also be any specific length that falls within the aforementioned range. Preferably, the probe does not contain a sequence that is complementary to the sequence of a primer used for a target sequence during a polymerase chain reaction.

[0071] The terms "complementary" or "complementarity" are used in reference to polynucleotides (i.e., sequences of nucleotides) that are related by the base pairing rules. For example, the sequence "AGT" is complementary to the sequence "TCA". Complementarity can be "partial", in which only some of the bases of the nucleic acids match according to the base pairing rules. Alternatively, there may be "complete" or "complete" complementarity between nucleic acids. The degree of complementarity between nucleic acid strands has a significant effect on the efficiency and strength of hybridization between nucleic acid strands.

[0072] The term "oligonucleotide" or "polynucleotide" refers to a polymeric form of nucleotides of any length, either deoxyribonucleotides or ribonucleotides. These terms include, but are not limited to, single-stranded, double-stranded, or triple-stranded DNA, genomic DNA, cDNA, RNA, DNA-RNA hybrids, or polymers containing purine and pyrimidine bases or other natural chemically, biochemically modified non-natural or derived nucleotide bases.

[0073] An "amplification detection assay" refers to a primer pair and a matched probe that define an amplicon, wherein the primer pair flanks a target nucleic acid region (usually a target gene), and wherein the probe binds to the amplicon.

[0074] The terms "gene variant" and "nucleotide variant" are used interchangeably herein to refer to changes or alterations at a specific locus relative to a reference human gene or cDNA sequence, including, but not limited to, nucleotide base deletions, insertions, inversions, and substitutions in coding and non-coding regions. The deletion may be a single nucleotide base, a portion or region of the nucleotide sequence of a gene, or a deletion of the entire gene sequence. The insertion may be an insertion of one or more nucleotide bases. "Gene variants" or "nucleotide variants" may occur in transcriptional regulatory regions, untranslated regions of mRNA, exons, introns, or exon / intron junctions. "Gene variants" or "nucleotide variants" may or may not produce stop codons, frame shifts, amino acid deletions, altered gene transcript splicing forms, or altered amino acid sequences.

[0075] The term "gene" refers to a polynucleotide (e.g., a DNA segment) that encodes a polypeptide and includes regions before and after the coding region and intervening sequences (introns) between individual coding segments (exons). Parent gene or protein sequences are presented in Entrez Gene ID or accession number format. For example, the ZNF479 Entrez Gene ID is 90827. If any changes have been made to the sequence in the Gene ID in Entrez, then a decimal point and the number of changes are indicated after the Gene ID (e.g., 90827.1). In addition, for example, TPM1 has accession number NM_004304.

[0076] The term "allele" or "gene allele" is used herein to generally refer to a naturally occurring gene having a reference sequence or a gene containing specific nucleotide variants.

[0077] As used herein, "haplotype" is a combination of gene (nucleotide) variants in a region of mRNA or genomic DNA on a chromosome found in an individual. Therefore, a haplotype includes polymorphic variants of many genes that are usually inherited together as a unit.

[0078] As used herein, the term "amino acid variant" is used to refer to amino acid changes relative to a reference human protein sequence produced by a "gene variant" or "nucleotide variant" of a reference human gene encoding a reference protein. The term "amino acid variant" is intended to cover not only single amino acid substitutions, but also amino acid deletions, insertions, and other significant changes in the amino acid sequence in a reference protein. The variants of the present invention are described by the following nomenclature: [original amino acid residue / position / substituted amino acid residue]. For example, a substitution of leucine for arginine at position 76 is represented as R76L.

[0079] As used herein, the term "genotype" means the nucleotide characters at a specific nucleotide variant marker (or locus) in any allele or two alleles of a gene (or a specific chromosome region). For the specific nucleotide position of a related gene, the nucleotides at the locus or its equivalent in one or two alleles form the genotype of the gene at the locus. Genotype can be homozygous or heterozygous. Therefore, "genotyping" means measuring genotype, that is, nucleotides at a specific locus. Genotyping can also be carried out by measuring the amino acid variants at the specific positions of the protein, which can be used to derive the corresponding nucleotide variants.

[0080] A probe set generally refers to a primer set, typically a primer pair, and / or a detectably labeled probe for detecting a target genetic variation. The primer pair is used in an amplification reaction to define an amplicon spanning a target genetic variation region of each of the above genes. The amplicon set is detected by a set of matched probes. In an exemplary embodiment, the present invention is a set of TaqMan TM (Roche Molecular Systems, Pleasanton, CA) assays for detecting a panel of target genetic variations for use in the methods of the invention.

[0081] In one embodiment, the probe set is a set of primers for generating amplicons, which are detected by a nucleic acid sequencing reaction (such as a next generation sequencing reaction). For example, in these embodiments, AmpliSEQ TM (Life Technologies / Ion Torrent, Carlsbad, CA) or TruSEQ TM (Illumina, San Diego, CA) technology. In other embodiments, two or more probes are primer pairs.

[0082] A modified ribonucleotide or deoxyribonucleotide refers to a molecule that can be used in place of a naturally occurring base in a nucleic acid, and includes, but is not limited to, modified purines and pyrimidines; rare bases; convertible nucleosides; structural analogs of purines and pyrimidines; labeled, derivatized, and modified nucleosides and nucleotides; conjugated nucleosides and nucleotides; sequence modifiers; terminal modifiers; spacer modifiers; and nucleotides with backbone modifications, including, but not limited to, ribose-modified nucleotides, phosphoramidates, phosphorothioates, phosphonamidates, methylphosphonates, methylphosphoramidites, methylphosphonamidates, 5'-β-cyanoethylphosphoramidite, methylenephosphonates, phosphorodithioates, peptide nucleic acids, achiral, and neutral internucleotide bonds.

[0083] "Hybridization" refers to the combination between nucleic acids. The conditions of hybridization can be changed according to the sequence homology of the nucleic acids to be combined. Therefore, if the sequence homology between the test nucleic acids is high, stringent conditions are used. If the sequence homology is low, mild conditions are used. When the hybridization conditions are strict, hybridization specificity increases, and this increase in hybridization specificity causes a reduction in the productivity of non-specific hybridization products. However, under mild hybridization conditions, hybridization specificity decreases, and this reduction in hybridization specificity causes an increase in the productivity of non-specific hybridization products.

[0084] "Stringent conditions" refers to conditions under which a probe will hybridize to its target subsequence (usually in a complex mixture of nucleic acids) but not to other sequences. Stringent conditions are sequence dependent and will vary from situation to situation. Longer sequences hybridize specifically at higher temperatures. A detailed guide to nucleic acid hybridization can be found in Tijssen, Techniques in Biochemistry and Molecular Biology--Hybridization with Nucleic Probes, "Overview of principles of hybridization and the strategy of nucleic acid assays" (1993). Typically, stringent conditions are selected to be higher than the thermal melting point (T) of a specific sequence at a defined ionic strength pH. m ) is about 5℃-10℃ lower. T m 50% of the probes complementary to the target are in equilibrium with the target sequence (when the target sequence is present in excess, at T mStringent conditions may also be achieved with the addition of destabilizing agents such as formamide. For selective or specific hybridization, a positive signal is at least two times the background, preferably 10 times the background hybridization. Exemplary stringent hybridization conditions may be as follows: 50% formamide, 5×SSC and 1% SDS, incubated at 42°C, or 5×SSC, 1% SDS, incubated at 65°C, washed in 0.2×SSC and 0.1% SDS at 65°C.

[0085] If the polypeptides encoded by the nucleic acids are substantially identical, then the nucleic acids that do not hybridize to each other under stringent conditions are still substantially identical. This occurs, for example, when a nucleic acid copy is formed using the maximum code degeneracy permitted by the genetic code. In such cases, nucleic acids are typically hybridized under moderately stringent hybridization conditions. Exemplary "moderately stringent hybridization conditions" include hybridization at 37°C in a buffer of 40% formamide, 1M NaCl, 1% SDS, and washing at 45°C in 1×SSC. Positive hybridization is at least twice the background. Those skilled in the art will readily recognize that alternative hybridization and washing conditions can be used to provide conditions of similar stringency. Other guidelines for determining hybridization parameters are provided in many references such as and in the Current Protocols in Molecular Biology edition.

[0086] Hybridization between nucleic acids can occur between DNA molecules and DNA molecules, between DNA molecules and RNA molecules, and between RNA molecules and RNA molecules.

[0087] "Mutant protein" or "variant" refers to a polynucleotide or polypeptide that differs from the wild type or most prevalent form in a population of individuals by the exchange, deletion or insertion of one or more nucleotides or amino acids, respectively. The number of nucleotides or amino acids exchanged, deleted or inserted can be 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more, such as 25, 30, 35, 40, 45 or 50. The term mutant protein can also encompass translocations, such as a fusion of polypeptides encoded by the ALK and TPM1 genes (TPM1 / ALK).

[0088] "Gene fusion" refers to a chimeric genomic DNA produced by the fusion of at least a portion of a first gene with a portion of a second gene. The transition point between the sequence from the first gene in the fusion and the sequence from the second gene in the fusion is called the "breakpoint" or "fusion point".

[0089] Transcription of the gene fusion produces a chimeric mRNA.

[0090] "Single nucleotide polymorphism" or "SNP" refers to a DNA sequence variation that occurs when a single nucleotide (A, T, G or C) in the genome differs between paired chromosomes of members of a biological species or humans.

[0091] A "mutation" is defined herein as a specific change at a genomic location, ie, chromosome, start, stop, reference base, alternative base, variant type (SNP, INS, DEL), etc.

[0092] An "annotation" is defined herein as a set of transcript-specific properties that describe the effect of a mutation, i.e.: gene, transcript, variant classification, variant change, variant codon position, etc.

[0093] "Primer" or "primer sequence" refers to an oligonucleotide that hybridizes with a target nucleic acid sequence (e.g., a DNA template to be amplified) to initiate a nucleic acid synthesis reaction. A primer can be a DNA oligonucleotide, an RNA oligonucleotide, or a chimeric sequence. A primer can contain natural, synthetic, or modified nucleotides. The upper and lower limits of the primer length are determined empirically. The lower limit of the primer length is the minimum length required to form a stable double helix with the target nucleic acid under nucleic acid amplification reaction conditions during hybridization. Very short primers (usually less than 3-4 nucleotides long) do not form thermodynamically stable double helices with the target nucleic acid under such hybridization conditions. The upper limit is usually determined by the possibility of forming a double helix in the target nucleic acid in addition to the predetermined nucleic acid sequence. In general, suitable primer lengths are in the range of about 10 to about 40 nucleotides in length. In certain embodiments, for example, a primer can be 10-40, 15-30, or 10-20 nucleotides long. When placed under appropriate conditions, a primer can serve as a starting point for synthesis on a polynucleotide sequence.

[0094] The primer will be completely or substantially complementary to the region of the target polynucleotide sequence to be copied. Thus, under conditions conducive to hybridization, the primer will adhere to the complementary region of the target sequence. After adding suitable reactants (including but not limited to polymerase, nucleotide triphosphates, etc.), the primer is extended by a polymerization agent to form a copy of the target sequence. The primer can be single-stranded or alternatively can be partially double-stranded.

[0095] "Detection", "detectable" and their grammatical equivalents refer to a manner of determining the presence and / or amount and / or identity of a target nucleic acid sequence. In some embodiments, detection is performed to amplify the target nucleic acid sequence. In other embodiments, sequencing of the target nucleic acid can be characterized as "detecting" the target nucleic acid. The label attached to the probe can include any of a variety of different labels known in the art that can be detected by, for example, chemical or physical means. Labels that can be attached to the probe can include, for example, fluorescent and luminescent materials.

[0096] "Amplifying" and grammatical equivalents thereof refer to any method by which at least a portion of a target nucleic acid sequence is replicated in a template-dependent manner, including, but not limited to, a wide range of techniques for linear or exponential amplification of nucleic acid sequences. Exemplary ways of performing the amplification step include ligase chain reaction (LCR), ligase detection reaction (LDR), ligation followed by Q-replicase amplification, PCR, primer extension, strand displacement amplification (SDA), hyperbranched strand displacement amplification, multiple displacement amplification (MDA), nucleic acid strand-based amplification (NASBA), two-step multiplex amplification, rolling circle amplification (RCA), recombinase-polymerase amplification (RPA) (TwistDx, Cambridge, UK), and self-sustained sequence replication (3SR), including multiplex formats or combinations thereof, such as, but not limited to, OLA / PCR, PCR / OLA, LDR / PCR, PCR / PCR / LDR, PCR / LDR, LCR / PCR, PCR / LCR (also known as combined chain reaction-CCR), etc. Descriptions of such techniques can be found in Sambrook et al., Molecular Cloning, 3rd ed.; Ausbel et al.; PCR Primer: A Laboratory Manual, Diffenbach, ed., Cold Spring Harbor Press (1995); The Electronic Protocol Book, Chang Bioscience (2002), Msuih et al., J. Clin. Micro. 34:501-07 (1996); The Nucleic Acid Protocols Handbook, R. Rapley, ed., Humana Press, Totowa, NJ (2002), and elsewhere.

[0097] Analysis of nucleic acid markers can be performed using techniques known in the art, including, but not limited to, sequence analysis and electrophoretic analysis. Non-limiting examples of sequence analysis include Maxam-Gilbert sequencing, Sanger sequencing, capillary array DNA sequencing, thermal cycle sequencing (Sears et al., Biotechniques, 13:626-633 (1992)), solid phase sequencing (Zimmerman et al., Methods Mol Cell Biol, 3:39-42 (1992)), sequencing coupled with mass spectrometry such as matrix-assisted laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF / MS; Fu et al., Nat. Biotechnol, 16:381-384 (1998)), and hybridization sequencing. Chee et al., Science, 274:610-614 (1996); Drmanac et al., Science, 260:1649-1652 (1993); Drmanac et al., Nature Biotechnology, 16:54-58 (1998). Non-limiting examples of electrophoretic analysis include slab gel electrophoresis (such as agarose or polyacrylamide gel electrophoresis), capillary electrophoresis, and denaturing gradient gel electrophoresis. In addition, next generation sequencing methods can be performed using commercially available kits and instruments (from companies such as Life Technologies / Ion Torrent PGM or Proton, Illumina HiSEQ or MiSEQ, and Roche / 454 Next Generation Sequencing System).

[0098] In some embodiments, the amount of the probe that gives the fluorescent signal in response to the excitation light is generally related to the amount of the nucleic acid produced in the amplification reaction. Therefore, in some embodiments, the amount of the fluorescent signal is related to the amount of the product formed in the amplification reaction. In the described embodiments, the amount of the amplified product can therefore be measured by measuring the intensity of the fluorescent signal from the fluorescent indicator.

[0099] A "detectably labeled probe" or "detector probe" refers to a molecule used in an amplification reaction, typically for quantitative or real-time PCR analysis as well as end-point analysis. The detector probe can be used to monitor the amplification of a target nucleic acid sequence. In some embodiments, the detector probe present in the amplification reaction is suitable for monitoring the amount of amplicon produced over time. Such detector probes include, but are not limited to, 5'-exonuclease assays (described herein). probes (see also U.S. Pat. No. 5,538,848), various stem-loop molecular beacons (see, e.g., U.S. Pat. Nos. 6,103,476 and 5,925,517 and Tyagi and Kramer, 1996, Nature Biotechnology 14:303-308), stemless or linear beacons (see, e.g., WO 99 / 21881), PNA Molecular Beacons TM (see, e.g., U.S. Pat. Nos. 6,355,421 and 6,593,091), linear PNA beacons (see, e.g., Kubista et al., 2001, SPIE 4264:53-58), non-FRET probes (see, e.g., U.S. Pat. No. 6,150,097), / Amplifluor TM probes (U.S. Pat. No. 6,548,250), stem-loop and double-helix scorpion probes (Solinas et al., 2001, Nucleic Acids Research 29:E96 and U.S. Pat. No. 6,589,743), bulge-loop probes (U.S. Pat. No. 6,590,091), pseudoknot probes (U.S. Pat. No. 6,589,250), cyclicons (U.S. Pat. No. 6,383,752), MGB Eclipse TMprobes (Epoch Biosciences), hairpin probes (U.S. Pat. No. 6,596,490), peptide nucleic acid (PNA) optical probes, self-assembled nanoparticle probes, and ferrocene-modified probes, such as those described in U.S. Pat. No. 6,485,901; Mhlanga et al., 2001, Methods 25:463-471; Whitcombe et al., 1999, Nature Biotechnology 17:804-807; Isacsson et al., 2000, Molecular Cell Probes 14:321-328; Svanvik et al., 2000, Analytical Biochemistry 14:331-338; Biochem. 281:26-35; Wolffs et al., 2001, Biotechniques 766:769-771; Tsourkas et al., 2002, Nucleic Acids Res. Research).30:4208-4215; Riccelli et al., 2002, "Nucleic Acids Research" 30:4088-4093; Zhang et al., 2002 Shanghai.34:329-332; Maxwell et al., 2002, "Journal of the American Chemical Society" (J.Am.Chem.Soc.) 124:9606-9612; Broude et al., 2002, "Trends Biotechnol." (Trends Biotechnol.) 20:249-56; Huang et al., 2002, "Chem. Res. Toxicol." (Chem. Res. Toxicol.) 15:118-126; and Yu et al., 2001, "Journal of the American Chemical Society" 14:11155-11161.

[0100] The detector probe may also include a quencher, including but not limited to Black Hole Quencher (Biosearch), Iowa Black (IDT), QSY Quencher (Molecular Probes), and Dabsyl and Dabcel sulfonate / formate quenchers (Epoch).

[0101] The detector probe may also include two probes, where, for example, the fluorescer is on one probe and the quencher is on the other probe, where hybridization of the two probes together on the target quenches the signal, or where hybridization on the target changes the signal characteristics by changing the fluorescence. The detector probe may also include sulfonate derivatives of fluorescein dyes having SO3 instead of carboxylate groups, phosphoramidite forms of fluorescein, phosphoramidite forms of CY5 (commercially available, for example, from Amersham). In some embodiments, intercalating chelator labels are used, such as ethidium bromide, Green I (Molecular Probes) and (Molecular Probes), thereby allowing real-time or endpoint observation of amplification products in the absence of a detector probe. In some embodiments, real-time observation can include insertion of a detector probe and can employ a sequence-based detector probe. In some embodiments, the detector probe is at least partially quenched when it is not hybridized to a complementary sequence in an amplification reaction, and is at least partially unquenched when it is hybridized to a complementary sequence in an amplification reaction. In some embodiments, the Tm of the detector probe taught by the present invention is 63-69°C, but it should be understood that routine experiments can produce detector probes with other Tms guided by the teachings of the present invention. In some embodiments, the probe may further include various modifications, such as a minor groove binder (see, e.g., U.S. Pat. No. 6,486,308), to further provide desired thermodynamic characteristics.

[0102] In some embodiments, detection can be based on the difference in migration rate between different analytes via any of a variety of mobility-related analytical techniques. Exemplary mobility-related analytical techniques include electrophoresis, chromatography, mass spectrometry, sedimentation (e.g., gradient centrifugation), field flow classification, multi-stage extraction techniques, etc. In some embodiments, the motility probe can be hybridized with the amplification product, and the identity of the target nucleic acid sequence is determined via the mobility-dependent analytical technique of eluting the motility probe, as described in, for example, disclosed PCT applications WO04 / 46344 (Rosenblum et al.) and WO01 / 92579 (Wenz et al.). In some embodiments, detection can be achieved by various microarrays and associated software, such as, inter alia, the Applied Biosystems array system and the Applied Biosystems 1700 chemiluminescent microarray analyzer and other commercially available array systems available from Affymetrix, Agilent, Illumina, and Amersham Biosciences (see also Gerry et al., J. Mol. Biol. 292:251-62, 1999; DeBellis et al., Minerva Biotec 14:247-52, 2002; and Stears et al., Nat. Med. 9:14045, including supplements, 2003). It is also understood that detection can include a reporter gene that is incorporated into the reaction product as part of a labeled primer or due to the incorporation of a labeled dNTP during amplification, or attached to the reaction product, for example, but not limited to, via a hybridization tag complementary sequence comprising the reporter gene or via a linker arm integral to or attached to the reaction product. Detection of unlabeled reaction products, such as using mass spectrometry, is also within the scope of the current teachings.

[0103] "Aberration" refers to a structural variation or change in the genome of DNA. Examples include: overexpression / underexpression; copy number amplification / deletion; mutation; gene fusion; etc.

[0104] By "driver event" is meant a genomic aberration representing a gain-of-function (GoF) mutation, fusion, or copy number spike.

[0105] "Recurrent" means that the event occurred in 3 or more tumor samples.

[0106] "Mitelman" refers to a manually curated database of chromosomal aberrations and gene fusions in cancer from the literature. http: / / goo.gl / PnXMT

[0107] Gene fusion

[0108] Table 1: Gene fusions

[0109]

[0110]

[0111] Table 2: Gene fusions

[0112]

[0113]

[0114] Table 3: Gene fusions

[0115]

[0116]

[0117]

[0118]

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130]

[0131] The present invention provides novel gene fusions and gene fusion variants (i.e., different breakpoint locations on one or both of the partner genes) selected from those shown in Tables 1-3, 19, and 22 of gene fusions, such as TPM1 / ALK, PRKAR1A / ALK, NCOA1 / ALK, LPP / CASR, MDM2 / EGFR, FGFR3 / ELAVL3, B2M / GNAS, DOCK8 / JAK2, HNF1B / NOTCH1, NFASC / NTRK1, SSBP2 / NTRK1, SQSTM1 / NTRK1, TBL1XR1 / PIK3CA, AKAP13 / RET, FKBP15 / RET, TBL1XR1 / RET, CEP85L / ROS1, CLCN6 / RAF 1. TRAK1 / RAF1, PRKACA / AKT1, PRKACA / AKT2, MLL / FYN, ECHD1 / FYN, TTC13 / JAK2, SEC16A / NOTCH1, ERC1 / RET, GTF2IRD1 / ALK, HTATSF1 / BRS3, CDH1 / CCDC132, CCDC132 / CDH1 , ERBB2 / SLC29A3, MET / TFG; TFG / MET, NOTCH2 / MNDA, IRF2BP2 / NTRK1, EIF2C2 / PTK2, RARA / HOXB3, STAT3 / ETV4, and GFAP / VIM; VIM / GFAP, TOP1 / C17orf64, and TP53 / KIAA0753. Due to these discoveries, the present invention provides isolated gene fusion nucleic acids and sequences complementary thereto, amplicons, transcripts, reaction mixtures, and probes that specifically recognize nucleic acid sequences of gene fusions, sequences complementary thereto, amplicons, and transcripts. The present invention further encompasses antisense nucleotides for use in treating related diseases.

[0132] Table 1-Table 3, Table 19 and Table 22 provide a list of gene fusions (gene A / gene B), indicating the genes involved (gene A and gene B), chromosome position, breakpoint position, fusion type and distance. Gene fusions are displayed together with related TCGA diseases (Cancer Genome Atlas (The Cancer Genome Atlas)). Cancer is displayed with 3-4 letter abbreviations, which are described in more detail in the diagnostics section.

[0133] In general, Tables 1-3, 19, and 22 provide one or more novel gene fusions and / or correlations of gene fusions with TCGA diseases. For example, Table 19 presents novel gene fusions, and Table 22 presents novel correlations of gene fusions with TCGA diseases.

[0134] Tables 4-6, 20 and 23 provide the breakpoint sequences for the gene fusions in Tables 1-3, 19 and 22. The breakpoint sequences are identified as SEQ ID NOs: 1-257.

[0135] Assays and Kits

[0136] In certain embodiments, analysis and methods of detection are provided. The methods for detecting gene fusions provided herein are known in the art. As non-limiting examples, the analysis may include 5' nuclease PCR analysis (Applied Biosystems, Foster City, CA), next generation sequencing analysis (Ion Torrent, Carlsbad, CA; Illumina, San Diego, CA) or microarray analysis (Skotheim et al., Molecular Cancer 2009, 8:5). In at least one embodiment, the analysis or method includes at least one primer or probe that is complementary to or encodes the gene fusions and / or breakpoints in Tables 1-6.

[0137] In at least one embodiment, an assay and method for quantifying the expression of a gene fusion is provided. The method may involve quantifying the expression of one or more exons. For example, TaqMan TM Gene expression analysis can be designed for a set of known fusion transcripts for quantitative analysis. The analysis can be designed so that primers and probes span the breakpoint region, although in certain illustrative embodiments, primers and probes are not placed directly on the breakpoint.

[0138] In certain embodiments, the invention provides a primer, a probe, or a probe or primer set that specifically recognizes one or more of the gene fusions and / or breakpoints disclosed herein.

[0139] In one embodiment, the present invention provides a composition and a kit comprising a probe set that specifically recognizes a breakpoint in a gene fusion selected from Tables 1-3, 19 and 22 and / or Tables 4-6, 20 and 23. The probe set may, for example, be an amplification primer set. In another embodiment, a composition is provided herein that includes a primer set flanking a gene fusion selected from Tables 1-3, 19 and 22 in a target nucleic acid. The reaction mixture of this embodiment may further include a detector probe that binds to either side of a breakpoint in a gene fusion selected from Tables 1-3, 19 and 22, or binds to a binding region spanning a breakpoint in a gene fusion selected from Tables 1-3, 19 and 22. The reaction mixture including or not including the detector probe may further include a polymerase, dNTPs and / or uracil DNA deglycosylase (UDG). Polymerases and UDG are generally not from human sources. The reaction mixture may further include a target nucleic acid, such as a human target nucleic acid. The human target nucleic acid may, for example, be isolated from a biological sample from an individual suspected of having cancer.

[0140] In another embodiment, qPCR analysis is provided herein, such as TaqMan TM Analysis or MolecularBeacons TM An assay that specifically amplifies and detects target nucleic acids comprising SEQ ID NOs: 1-257.

[0141] The present invention also provides an isolated nucleic acid comprising at least one sequence selected from SEQ ID NO: 1-257. The isolated nucleic acid may include a first primer at the 5' end. In addition, the nucleic acid may be single-stranded or double-stranded.

[0142] In other embodiments, the present invention provides a kit comprising a detector probe and / or a probe set, such as an amplification primer set, that specifically recognizes a nucleic acid comprising a breakpoint of a gene fusion selected from Tables 1-3, 19, and 22. For example, in certain embodiments, the detector probe or amplification primer set is designed to amplify and / or detect a nucleic acid comprising at least one of SEQ ID NOs: 1-257. The kit may further include components from an amplification reaction mixture in another container or in the same container, such as a polymerase (usually not from human sources), dNTPs, and / or UDG. In addition, the kit may include a control nucleic acid. For example, the control nucleic acid may include a sequence comprising a breakpoint in a gene fusion selected from Tables 1-3, 19, and 22.

[0143] In some embodiments, a kit is provided, which encompasses at least 2 primer pairs and 2 detectably labeled probes. In these non-limiting embodiments, the 2 primer pairs and / or 2 detectably labeled probes form 2 amplification detection assays.

[0144] The kits of the present invention may also include instructions for performing one or more of the methods described herein and / or descriptions of one or more compositions or reagents described herein. The instructions and / or descriptions may be in printed form and may be included in a kit insert. The kit may also include a written description of an Internet location that provides such instructions or descriptions.

[0145] In some embodiments, the kits and assays comprise one or more probes that specifically recognize a target (eg, a gene fusion nucleic acid sequence). In at least one embodiment, the kits and assays are diagnostic kits and assays.

[0146] A kit is provided, comprising a probe set (e.g., an amplification primer set) that specifically recognizes a nucleic acid comprising a breakpoint from Tables 4-6, 20, and 23. The kit may further include components from an amplification reaction mixture in another container or in the same container, such as a polymerase (typically not from human origin), dNTPs, and / or UDG. In addition, the kit may include a control nucleic acid. For example, a control nucleic acid may include a sequence comprising a breakpoint selected from Tables 4-6, 20, and 23.

[0147] In another embodiment, a gene fusion is provided that comprises at least one of the breakpoints in Tables 4-6, 20, and 23.

[0148] In some embodiments, a reaction mixture and a kit are provided. In some embodiments, the kit encompasses a detectable probe that selectively binds to a gene fusion. In some embodiments, the gene fusion is any one of the gene fusions in Table 4, Table 5, Table 6, Table 20, or Table 23.

[0149] Thus, in some embodiments, a kit is provided that encompasses a reaction mixture and a detectable probe that selectively binds to a gene fusion, wherein the gene fusion is any one of the gene fusions in Table 4, Table 5, Table 6, Table 20, or Table 23.

[0150] Diagnostics

[0151] Methods of diagnosing, treating, and detecting gene fusions and related diseases are contemplated herein. The methods may include detecting a gene fusion in a sample from a subject.

[0152] The subject sample can be any body tissue or body fluid comprising nucleic acid from the subject. In certain embodiments, the sample will be a blood sample containing circulating tumor cells or cell-free DNA. In other embodiments, the sample can be a tissue, such as cancerous tissue. The cancerous tissue can be from tumor tissue and can be fresh frozen or formalin fixed, paraffin embedded (FFPE).

[0153] The disease may be cancer or a tumor. Cancer may include, but is not limited to, melanoma, cervical cancer, pancreatic cancer, head and neck squamous cell carcinoma, lung adenocarcinoma, colon adenocarcinoma, uterine cancer, ovarian cancer, glioblastoma, low-grade glioma, lung adenocarcinoma, thyroid cancer, and gastric cancer.

[0154] Cancers may include, but are not limited to, bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, squamous cell carcinoma of the head and neck, clear cell renal cell carcinoma, renal papillary cell carcinoma of the kidney, acute myeloid leukemia, low-grade glioma of the brain, hepatocellular carcinoma of the liver, lung adenocarcinoma, squamous cell lung cancer, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer, and uterine corpus endometrioid carcinoma. As used herein, BLCA = bladder cancer, BRCA = breast cancer, CESC = cervical cell carcinoma, COAD = colon adenocarcinoma, GBM = glioblastoma multiforme, HNSC = head and neck squamous cell carcinoma, KIRK = clear cell renal cell carcinoma, KIRP = kidney renal papillary cell carcinoma, LAML = acute myeloid leukemia, LGG = brain low-grade glioma, LIHC = liver hepatocellular carcinoma, LUAD = lung adenocarcinoma, LUSC = squamous cell lung cancer, OV = ovarian serous adenocarcinoma, PRAD = prostate adenocarcinoma, READ = rectal adenocarcinoma, SKCM = cutaneous melanoma, STAD = gastric adenocarcinoma, THCA = thyroid carcinoma, and UCEC = uterine corpus endometrioid carcinoma.

[0155] In some embodiments, a method of detecting a novel gene variant or gene fusion is provided, the method comprising a reaction mixture, wherein the novel gene variant or gene fusion is detected by generating an extension product.

[0156] In another embodiment, the present invention provides diagnostic and therapeutic targets using disclosed gene fusions and gene variants.Gene fusions, gene variants and related disease conditions provide targets for both diagnosis and treatment.For example, the presence, absence, or expression increase or decrease of a gene fusion target or gene variant can be used to diagnose a disease condition or can be used to predict or detect a disease condition.In at least one embodiment, a gene fusion or gene variant may have a high incidence (frequency), a medium incidence, or a low incidence in a specific cancer.In at least one embodiment, a gene fusion or gene variant may have a high frequency in a cancer or tumor, and a low or medium incidence in another cancer or tumor.In at least one embodiment, a gene fusion or gene variant may have a medium or low frequency associated with a cancer or tumor.In at least one embodiment, a low or medium frequency gene fusion or gene variant may be used in combination with one or more different high frequency biomarkers of cancer to help diagnose, predict or identify the tendency of a disease.The method can be used to screen a patient's cancer or predict the relative outlook of a specific outcome of a cancer. For example, the presence of BRCA1 or BRCA2 mutations can be analyzed in combination with the gene fusion JAK2 / TTC13 in breast cancer.

[0157] A method for detecting cancer is provided, comprising amplifying a nucleic acid spanning a breakpoint in a gene fusion selected from Tables 1-3, 19 and 22, for example, the nucleic acid may include a sequence selected from SEQ ID NO: 1-257; and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that cancer is present in a sample. In another method, a method for detecting cancer is provided herein, comprising generating an amplicon comprising a sequence selected from SEQ ID NO: 1-257, and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that cancer or cancer cells are present in a sample. The amplicon typically includes a primer that is extended to form the amplicon. The cancer is selected from bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, clear cell renal cell carcinoma, kidney renal papillary cell carcinoma, acute myeloid leukemia, brain low-grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, squamous cell lung cancer, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer and uterine body endometrioid carcinoma.

[0158] In another embodiment is a method for detecting in a sample a cancer selected from bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, clear cell renal cell carcinoma, kidney renal papillary cell carcinoma, acute myeloid leukemia, brain low-grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, squamous cell lung cancer, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer, and uterine corpus endometrioid carcinoma by detecting the presence of a gene fusion selected from Tables 1-3, 19 and 22.

[0159] New gene fusion

[0160] Although some of the gene fusions have been previously reported, provided herein are a number of variations of gene fusions in which the breakpoints are different and / or not previously known. Non-limiting examples of gene fusions in which the breakpoints are different and / or not previously known include: TPM1 / ALK, PRKAR1A / ALK, NCOA1 / ALK, LPP / CASR, MDM2 / EGFR, FGFR3 / ELAVL3, B2M / GNAS, DOCK8 / JAK2, HNF1B / NOTCH1, NFASC / NTRK1, SSBP2 / NTRK1, SQSTM1 / NTRK1, TBL1XR1 / PIK3CA, AKAP13 / RET, FKBP15 / RET, TBL1XR1 / RET, CEP85L / ROS1, CLCN6 / RAF1, TRAK1 / RAF1, PRKACA / AKT1, PRKACA / AKT2, MLL / FYN, ECHD1 / FYN, and TTC13 / JAK2 are the fusions having the genes indicated in Tables 4 and 5 by SEQ ID Novel variants of the breakpoints provided in the form of NO:1-257.

[0161] Also provided herein are a number of gene fusion variants associated with one or more cancers.

[0162] Cancer Relevance

[0163] New gene fusions associated with cancer are presented herein. Some of the gene fusions may have been associated with previous specific cancers or disease conditions. The methods herein have identified new correlations that can be used to help diagnose and / or treat specific cancers. The gene fusions shown in Tables 1-3, 19, and 22 provide genes associated with the fusions and the correlation of the gene fusions with one or more specific cancers. For example, the fusion PRKACA / AKT1 is shown to be associated with colon adenocarcinoma and endometrial endometrioid adenocarcinoma.

[0164] The gene fusions shown in Table 3 are previously known gene fusions that have been shown to be associated with new cancers. For example, SEC16A / NOTCH1 has been previously identified as being associated with breast cancer. The current method identifies the association of the gene fusion SEC16A / NOTCH1 with thyroid gland cancer. In addition, ERC1 / RET has been previously identified as being associated with thyroid cancer. The current method identifies the association of the gene fusion ERC1 / RET with invasive breast cancer (see Tables 3 and 6).

[0165] Reaction mixture and amplicons

[0166] In another embodiment, the present invention provides a reaction mixture comprising a probe or a probe set that specifically recognizes a gene fusion selected from Tables 1-3, 19, and 22. The probe set can be, for example, an amplification primer set or a labeled probe. In another embodiment, the present invention provides a reaction mixture comprising a primer set flanking a gene fusion selected from Tables 1-3, 19, and 22 in a target nucleic acid. For example, the primer set can each bind to a target sequence in the human genome within 1000, 750, 500, 250, 100, 90, 80, 75, 70, 65, 50, or 25 nucleotides on opposite sides of one of the fusion breakpoints identified in Tables 4-6, 20, and 23. The reaction mixture of this embodiment can further include a detector probe that binds to either side of a breakpoint in a gene fusion selected from Tables 1-3, 19, and 22, or to a binding region spanning a breakpoint in a gene fusion selected from Tables 1-3, 19, and 22, including specific embodiments in which the breakpoints are identified in Tables 4-6, 20, and 23. In exemplary embodiments, the detector probe binds to a target sequence in the human genome within 1000, 750, 500, 250, 100, 90, 80, 75, 70, 60, 50, or 25 nucleotides of one of the fusion breakpoints identified in Tables 4-6, 20, and 23. The reaction mixture including or not including the detector probe can further include a polymerase, a reverse transcriptase, dNTPs, and / or uracil DNA deglycosylase (UDG). The polymerase, reverse transcriptase, and UDG are typically not from human sources. The polymerase in the illustrative embodiments is a thermostable polymerase, such as Taq polymerase. In certain embodiments, the dNTPs in the reaction mixture include dUTP, and in certain instances, the reaction mixture may be free of dTTP.

[0167] The reaction mixture may further include a target nucleic acid, such as a human target nucleic acid. The human target nucleic acid may be isolated, for example, from a biological sample (such as a tumor sample) of an individual suspected of having a cancer selected from the following: BLCA = bladder cancer, BRCA = breast cancer, CESC = cervical cell carcinoma, COAD = colon adenocarcinoma, GBM = glioblastoma multiforme, HNSC = head and neck squamous cell carcinoma, KIRK = clear cell renal cell carcinoma, KIRP = kidney renal papillary cell carcinoma, LAML = acute myeloid leukemia, LGG = brain low-grade glioma, LIHC = liver hepatocellular carcinoma, LUAD = lung adenocarcinoma, LUSC = squamous cell lung cancer, OV = ovarian serous adenocarcinoma, PRAD = prostate adenocarcinoma, READ = rectal adenocarcinoma, SKCM = skin melanoma, STAD = gastric adenocarcinoma, THCA = thyroid carcinoma, and UCEC = uterine body endometrioid carcinoma. In certain embodiments, the target nucleic acid is from a tumor, such as a tumor of one of the cancer types listed in the previous sentence. Additionally, target nucleic acids can be extracted from biological samples from tumors (eg, FFPE samples).

[0168] Reaction mixture of the present invention can include amplicon. Amplicon can be for example through separation nucleic acid. The length of amplicon can be for example between 25 and 2500 nucleotides, between 25 and 2000 nucleotides, between 25 and 1000 nucleotides, between 50 and 1000 nucleotides, between 50 and 500 nucleotides, between 50 and 250 nucleotides, between 50 and 200 nucleotides, between 50 and 150 nucleotides, between 50 and 100 nucleotides or between 50 and 75 nucleotides.

[0169] The amplicon may have a nucleotide sequence identical or complementary to a target sequence in the human genome within 1000, 750, 500, 250, 100, 90, 80, 75, 70, 65, 50, or 25 nucleotides on the opposite side of one of the fusion breakpoints identified in Tables 4-6, 20, and 23. In certain embodiments, the amplicon comprises Figure 4-6 In some embodiments, the amplicon comprises 25 to 250, 25 to 100, 25 to 75, 50 to 250, 50 to 200, 50 to 150, 50 to 100 or 50 to 75 of the nucleotide sequence provided in or its complementary sequence. In certain embodiments, the amplicon comprises a sequence variant occurring in nature. For example, the amplicon can include a variable nucleotide sequence corresponding to a single nucleotide variant or a naturally occurring allele.

[0170] In certain illustrative embodiments, the amplicons of the invention have chemical structures not found in nature and / or not found in mammals such as humans. For example, certain illustrative amplicons include bases of a type not found in nature or not found in mammals or that are found to be likely not to be bound to the sugar-phosphate backbone of the amplicon. For example, the amplicon may be a DNA amplicon that includes a uracil base bound to a sugar-phosphate backbone, thereby containing at least one position in the template with a thymidine residue and having a uridine residue at all positions in the illustrative examples.

[0171] Thus, in illustrative embodiments, the amplicon is a DNA amplicon that includes one or more deoxyuridine ("dU") residues. The dU residues can be added by including the residues in the primers used to generate the amplicon. In certain embodiments, the reaction mixture includes a DNA amplicon that includes one or more dU residues for each deoxythymidine residue in the corresponding human genome sequence. When an amplicon is generated using an amplification reaction (such as PCR), these amplicons can be generated, for example, by using a dNTP mixture that includes dUTP instead of dTTP.

[0172] In certain embodiments, the amplicon includes a segment whose corresponding sequence is not found in the human genome, such as an oligonucleotide sequence, such as a DNA barcode sequence. The length of the non-human segment can be, for example, 5-10,000, 5-5000, 5-1000, 5-500, 5-100, 5-50, 5-25, 5-10, 10-10,000, 10-5000, 10-1000, 10-50, 10-100, 10-50, or 10-25 nucleotides.

[0173] In certain embodiments, the amplicon includes a segment corresponding to a region of the human genome that spans an intron, but the amplicon does not include a segment corresponding to an intron.

[0174]

[0175]

[0176]

[0177]

[0178]

[0179]

[0180]

[0181]

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193]

[0194]

[0195]

[0196]

[0197]

[0198]

[0199]

[0200]

[0201]

[0202] The present invention provides novel gene variants and gene variant-disease condition correlations. Gene variants may have one or more mutations that produce variant proteins. Gene variants provided herein are associated with certain cancers. Gene variants produce protein variants. The present invention further provides probes (such as amplification primer sets and detection probes) and methods of detection, diagnosis and treatment and kits comprising or detecting the gene variants disclosed herein.

[0203] Variants are shown as amino acid variants in Tables 7 and 11 with the accession number or Entrez nucleotide and / or protein sequence of the parent or wild-type gene provided. The correlation with various cancers is shown in Tables 7 and 11. Tables 7 and 11 provide a list of more than 99 genes identified using the method outlined in Example 2. The variation or mutation is not found in the corresponding normal tissue. This is important because in a typical patient, a tumor sample can have 10's-100's of tumor-specific variation. However, the variation that occurs in the same position (and not in normal tissue) in multiple patients is more significant. Analyze 4445 samples (from 4445 patients) and prepare a hotspot list. Many recurring mutations are found in the same position in 15-20 different cancer types.

[0204] Diagnostics and kits

[0205] Methods for diagnosing, treating and detecting gene variants and related diseases are contemplated herein. The methods may include detecting gene fusions and / or gene variants in a sample from a subject. Any number and combination of gene fusions and / or gene variants may be detected in any of the reaction mixtures, compositions and kits disclosed herein.

[0206] In one embodiment, the present invention provides a composition and a kit comprising a probe set that specifically recognizes a nucleotide sequence encoding a gene variant selected from Table 7 and / or Table 11. The probe set may, for example, be an amplification primer set. In another embodiment, a composition is provided herein, comprising a primer set flanking a gene variant encoding one or more variants in Table 7 and / or Table 11. The reaction mixture of this embodiment may further include a detector probe that is bound to a nucleotide sequence comprising a gene variant selected from Table 7 and / or Table 11. The reaction mixture including or not including the detector probe may further include a polymerase, dNTP and / or uracil DNA deglycosylase (UDG). Polymerase and UDG are generally not derived from human sources. The reaction mixture may further include a target nucleic acid, such as a human target nucleic acid. The human target nucleic acid may, for example, be isolated from a biological sample from an individual suspected of having cancer. The cancer may be selected from the group consisting of: BLCA = bladder cancer, BRCA = breast cancer, CESC = cervical cell carcinoma, COAD = colon adenocarcinoma, GBM = glioblastoma multiforme, HNSC = head and neck squamous cell carcinoma, KIRK = clear cell renal cell carcinoma, KIRP = kidney renal papillary cell carcinoma, LAML = acute myeloid leukemia, LGG = brain low-grade glioma, LIHC = liver hepatocellular carcinoma, LUAD = lung adenocarcinoma, LUSC = squamous cell lung cancer, OV = ovarian serous adenocarcinoma, PRAD = prostate adenocarcinoma, READ = rectal adenocarcinoma, SKCM = cutaneous melanoma, STAD = gastric adenocarcinoma, THCA = thyroid carcinoma, and UCEC = uterine corpus endometrioid carcinoma.

[0207] In some embodiments, a kit is provided, wherein the kit encompasses one or more probes. In some embodiments, the kit encompasses probes for 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 50, 75, 100, 150, 200, 250, 500 or more fusion genes. In some embodiments, the probes are detectably labeled. In some embodiments, the probes hybridize to breakpoints present in the gene fusion.

[0208] In some embodiments, the detection of any one of the gene variants disclosed in Tables 7 and 11 can be combined with the detection of another one of the gene variants disclosed in those tables or any one of the gene fusions disclosed herein. That is, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 50, 75, 100, 150, 200, 250, 500 or more gene variants can be detected in the same reaction. In some embodiments, the gene variants detected are those disclosed in Tables 4-6, 7 and 11, 20 and 23, and can be combined with the detection of another one of the gene fusions disclosed in those tables. That is, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 50, 75, 100, 150, 200, 250, 500 or more gene fusions can be detected in the same reaction.

[0209] The nucleotide sequence encoding one or more gene variants in Table 7 and / or Table 11 can be any size that encompasses the variant. For example, the nucleotide sequence can be any size that can be easily copied using primers and / or detected using probes.

[0210] In another embodiment, a probe set is provided that specifically recognizes a nucleic acid encoding a gene variant selected from Table 7 and / or Table 11 (gene variants). In another embodiment, a primer set is provided herein that specifically amplifies a target nucleic acid encoding a gene variant selected from Table 7 and / or Table 11. In another embodiment, a qPCR assay is provided herein, such as TaqMan TM Analysis or MolecularBeacons TM An assay that specifically amplifies and detects a target nucleic acid encoding a gene variant selected from Table 7 and / or Table 11.

[0211] The present invention also provides an isolated nucleic acid comprising at least one sequence comprising a variation found in one or more gene variants selected from Table 7 and / or Table 11. The isolated nucleic acid may include a first primer at the 5' end. In addition, the nucleic acid may be single-stranded or double-stranded.

[0212] In other embodiments, the present invention provides a kit comprising a detector probe and / or a probe set, such as an amplification primer set, which specifically recognizes a nucleic acid encoding a gene variant selected from Table 7 and / or Table 11. For example, in certain embodiments, the detector probe or amplification primer set is designed to amplify and / or detect a nucleic acid comprising at least one of the nucleic acids encoding the gene variants in Table 7 and / or Table 11. The kit may further include components from an amplification reaction mixture in another container or in the same container, such as a polymerase (usually not from a human source), dNTPs and / or UDG. In addition, the kit may include a control nucleic acid. For example, the control nucleic acid may include a sequence comprising a gene variant selected from Table 7 and / or Table 11.

[0213] A method for detecting cancer is provided, comprising amplifying a nucleic acid encoding a gene variant selected from Table 7 and / or Table 11, for example, the nucleic acid may include a sequence from one of the accession numbers in Table 7 and / or Table 11, except that the sequence contains a variant encoding a gene variant in Table 7 and / or Table 11; and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that cancer is present in the sample. In another method, a method for detecting cancer is provided herein, comprising generating an amplicon comprising a sequence selected from a sequence encoding a gene variant in Table 7 and / or Table 11, and detecting the presence of the nucleic acid, wherein the presence of the nucleic acid indicates that bladder cancer, head and neck cancer, or lung squamous cell carcinoma is present in the sample. The amplicon typically includes a primer that is extended to form the amplicon. The cancer is selected from bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, clear cell renal cell carcinoma, kidney renal papillary cell carcinoma, acute myeloid leukemia, brain low-grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, squamous cell lung cancer, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer and uterine corpus endometrioid carcinoma.

[0214] A kit is provided, comprising a probe set (e.g., an amplification primer set) that specifically recognizes a nucleic acid comprising a gene variant from Table 7 and / or Table 11. The kit may further include components from an amplification reaction mixture in another container or in the same container, such as a polymerase (usually not from a human source), dNTPs, and / or UDG. In addition, the kit may include a control nucleic acid. For example, a control nucleic acid may include a sequence comprising a gene variant from Table 7 and / or Table 11. In certain embodiments, a probe set is provided that specifically recognizes a nucleic acid comprising a gene variant from Table 7 and / or Table 11.

[0215] In another embodiment, a genetic variant is provided, comprising at least one of the genetic variants in Table 7 and / or Table 11.

[0216] In another embodiment is a method of detecting a cancer selected from bladder cancer, breast cancer, cervical cell carcinoma, colon adenocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, clear cell renal cell carcinoma, kidney renal papillary cell carcinoma, acute myeloid leukemia, brain low-grade glioma, liver hepatocellular carcinoma, lung adenocarcinoma, squamous cell lung cancer, ovarian serous adenocarcinoma, prostate adenocarcinoma, rectal adenocarcinoma, skin melanoma, gastric adenocarcinoma, thyroid cancer, and uterine corpus endometrioid carcinoma in a sample by detecting the presence of a genetic variant selected from Table 7 and / or Table 11. Genetic variants may include, but are not limited to, ZNF479 variants R11Q, R295K, R295T, R295I, R345I, R345T, K438T, and T466K (see Table 8).

[0217] Table 18

[0218] Cancer Type Gene symbol Druggability KM Evidence Astrocytoma CXCR2 Y Endometrial endometrioid adenocarcinoma CXCR2 Y Squamous cell lung cancer CXCR2 Y Skin melanoma CXCR2 Y Skin melanoma CXCR2 Y Colorectal adenocarcinoma IL3 Y Gastric adenocarcinoma IL3 Y Skin melanoma KCNK9 Y Favorable results Endometrial endometrioid adenocarcinoma KCNK9 Y Lung adenocarcinoma KCNK9 Y Squamous cell lung cancer KCNK9 Y Poor results Non-small cell lung cancer, NOS S100A8 Y Glioblastoma S100A8 Y Head and Neck Squamous Cell Carcinoma S100A8 Y Papillary thyroid carcinoma S100A8 Y Skin melanoma S H Y Lung adenocarcinoma S H Y Skin melanoma CCDC61 Poor results Skin melanoma CCDC61 Poor results Skin melanoma CNTN5 Poor results Skin melanoma CNTN5 Poor results Skin melanoma CNTN5 Poor results Skin melanoma CNTN5 Poor results Colorectal adenocarcinoma CNTN5 Poor results Colorectal adenocarcinoma CNTN5 Poor results Skin melanoma CNTN5 Poor results Skin melanoma CNTN5 Poor results Skin melanoma CNTN5 Poor results Skin melanoma EDDM3A Poor results Skin melanoma FABP1 Poor results Lung adenocarcinoma OR2L13 Poor results Skin melanoma OR4E2 Poor results Skin melanoma OR4E2 Poor results Skin melanoma OR4E2 Poor results Skin melanoma PRSS37 Poor results Skin melanoma PRSS37 Poor results Skin melanoma SPINK13 Poor results Endometrial endometrioid adenocarcinoma ST6GAL2 Poor results

[0219] Table 18 provides druggability or prognostic correlations filtered from Table 11. Table 18 provides cancer type, gene symbol of the gene identified as druggable in Table 11, druggability (Y=yes), and KM evidence. The KM evidence column provides Kaplan-Meier evidence. KM evidence indicates whether the event type supports a good or poor prognosis in a particular cancer type.

[0220] Targeted therapy

[0221] In at least one embodiment, gene fusions and / or gene variants can be used to identify targeted therapies. Targeted therapies can include identifying agents that specifically interact with gene fusions and / or gene variants. Targeted therapies can include, but are not limited to, antibody therapies, antisense therapies, and small molecule therapies. Antisense therapies are discussed in more detail under the heading "Antisense".

[0222] Compositions and methods for inactivating nucleic acid molecules involve, in part, the use of molecules having a nucleic acid region having a sequence complementary to a nucleic acid molecule that is the subject of desired inactivation (i.e., a target nucleic acid molecule). The methods of the present invention can be used to inactivate gene fusions and / or gene variants associated with specific cancers. Thus, antisense molecules complementary to any of the gene fusions or gene variants identified herein can be identified.

[0223] Small molecules are low molecular weight (<800 Daltons) organic compounds that can serve as enzyme substrates or regulators of biological processes and are approximately 10-9 m. In pharmacology, the term is often used for molecules that bind to proteins or nucleic acids and act as effectors (alter the activity or function of proteins or nucleic acids). The effector function of small molecules can be tested by expressing the gene fusion or variant in a cell assay and identifying small molecules that inhibit the expression or activity of the gene fusion or variant.

[0224] Druggability is a term used in drug discovery to describe biological targets, such as proteins, that are known to bind or predicted to bind to drugs with high affinity. In addition, drug binding to a druggable target will change the function of the target with a therapeutic benefit to the patient. The term "drug" herein includes small molecules (low molecular weight organic substances) and has been expanded to include biomedical products, such as therapeutic monoclonal antibodies. In at least one embodiment, a gene fusion or gene variant can be used to identify a druggable target. Table 8 provides a list of druggable targets that have been identified from Tables 1-3 and 7. For example, the TPM1 / ALK gene fusion is a druggable target because, as shown in Table 8, diseases involving ALK can be treated with crizotinib. Therefore, if the gene fusion includes ALK, then cancer can be treated with crizotinib. In addition, if the gene variant includes a mutation in ALK, then cancer can be treated with crizotinib.

[0225] Similarly, Table 21 provides a list of druggable targets that have been identified from Table 19, and Table 24 provides a list of druggable targets that have been identified from Table 22.

[0226] Table 8: Druggable genes from Table 1

[0227]

[0228]

[0229]

[0230]

[0231]

[0232] Table 8 provides a list of 11 druggable targets identified in the gene fusions in Tables 1-3 or the gene variants in Tables 7 and 11. Tables 16 and 17 provide an analysis of other druggable targets within the gene variants in Tables 1-3 or Tables 7 and 11. Tables 8, 16 and 17 provide information about druggable targets, including the gene name, whether the drug has been approved by the US Food and Drug Administration (FDA) (N = No), and if the drug has not been approved, then which stage of clinical trials is in (pre-registration, Phase III, Phase II, Phase I, and preclinical). For example, a drug related to the NOTCH1 gene has not been approved, but is in Phase 1 clinical trials as of this writing (see OMP-52M51).

[0233] Approved drugs include, but are not limited to, crizotinib for diseases with ALK gene fusions and cinacalcet hydrochloride for diseases with CASR gene fusions. Many approved drugs have been identified for gene fusions with EGFR, including, but not limited to, erlotinib; panitumumab; cetuximab; nepidamine; gefitinib; nimotuzumab; vandetanib; lapatinib ditosylate; and icotinib hydrochloride. Approved drug ponatinib has been identified for diseases with FGFR3, and ruxolitinib has been identified for diseases with JAK2 gene fusions. Many approved drugs have been identified for gene fusions with RET, including, but not limited to, sorafenib; vandetanib; sunitinib malate; cabozantinib; and regorafenib. Approved drug crizotinib has been identified for diseases with ROS1. Other drugs that may prove applicable include (but are not limited to) crizotinib, afatinib, masitinib, zarutumumab, neratinib, dovitinib lactate, XL647, lindopem, lesitumumab, dacomitinib, SAR-302503, paritinib, perifosine, buparcoxib, motesanib diphosphate, and apatinib.

[0234] Method provided herein can include delivering medicine to experimenter or patient.Described medicine can be the medicine that has been approved according to government drug regulatory department (such as FDA), or described medicine can be in any one stage before approval stage.In illustrative aspect, described medicine is the medicine approved by FDA.In other aspects, described medicine can be in preclinical, I phase, II phase, III phase or pre-approval stage.In some aspects, method provided herein includes delivering one or more medicines listed in table 8,16 and 17 to experimenter.When identifying the gene event related to one or more genes listed in table 8,16 and 17 in experimenter, method provided herein can include delivering one or more medicines, especially delivering the medicine related to the different genes affected by identified gene event.

[0235] Antonym

[0236] Antisense technology has been applied to inhibit the expression of various oncogenes. For example, an antisense-oriented Craf-1 cDNA fragment introduced into human squamous carcinoma under the control of the adenovirus 2 late promoter greatly reduced tumorigenic potential relative to cells transfected with a control sense transfectant. Similarly, Cmyc antisense constructs accelerate differentiation and inhibit G1 progression in Friend Murine Erythroleukemia cells. Antisense methods take advantage of the tendency of nucleic acids to pair with "complementary" sequences. Complementary sequences are those polynucleotides that can base pair according to the standard Watson-Crick complementarity rules. Targeting double-stranded (ds) DNA with polynucleotides causes triple helix formation; targeting RNA will cause double helix formation. Antisense polynucleotides specifically bind to their target polynucleotides when introduced into target cells and interfere with transcription, RNA processing, transport, translation and / or stability. Antisense RNA constructs or DNA encoding the antisense RNA can be used to inhibit gene transcription or translation, or both, in host cells in vitro or in vivo (eg, in host animals, including human subjects).

[0237] Antisense can be under the transcriptional control of a promoter. "Promoter" refers to a DNA sequence that is recognized by the synthetic machinery of the cell or introduced synthetic machinery and is required for initiating specific transcription of a gene. The phrase "under transcriptional control" means that the promoter is in the correct position and orientation relative to the nucleic acid to control RNA polymerase initiation and expression of the gene.

[0238] In some cases, the antisense expression construct will comprise a virus or an engineered construct derived from a viral genome. When cDNA is inserted, a polyadenylation signal that realizes the appropriate polyadenylation of the gene transcript may be included. The nature of the polyadenylation signal is not considered to be critical and any of the described sequences may be used. Terminators may be used to enhance messenger levels and minimize synchronous reading from the box to other sequences.

[0239] Antisense constructs can be designed to bind to promoters and other control regions, exons, introns or even exon-intron boundaries of gene fusions or gene variants disclosed herein. The most effective antisense constructs include regions complementary to intron / exon splice junctions. One embodiment includes antisense constructs complementary to regions within 50-200 bases of intron-exon splice junctions. It has been observed that some exon sequences can be included in constructs without seriously affecting their target selectivity. The amount of exon material included will vary, depending on the specific exon and intron sequences used. It can be easily tested by in vitro testing constructs whether there is too much exon DNA simply included to determine whether normal cell function is affected or whether the expression of related genes with complementary sequences is affected.

[0240] The term "complementary" means, in terms of antisense, a polynucleotide sequence that is substantially complementary over its entire length and has very few base mismatches. For example, a sequence of fifteen bases in length may be referred to as complementary when it has complementary nucleotides at thirteen or fourteen positions. Naturally, a fully complementary sequence will be a sequence that is fully complementary over its entire length and has no base mismatches. Other sequences with a lower degree of homology are also contemplated. For example, antisense constructs with limited high homology regions and containing non-homologous regions (e.g., ribonucleases) may be designed. These molecules, despite having less than 50% homology, will still bind to the target sequence under appropriate conditions.

[0241] Antisense in vivo, in vitro or in vitro delivery may involve the use of vectors. An effective vector for antisense delivery is an adenovirus expression vector. "Adenovirus expression vector" means including those constructs containing adenovirus sequences sufficient to support construct packaging and expression of antisense polynucleotides cloned therein. The expression vector may include a genetically engineered form of adenovirus. Adenovirus is particularly suitable as a gene transfer vector because of its medium-sized genome, ease of manipulation, high titer, wide target cell range and high infectivity. The two ends of the viral genome contain 100-200 base pairs of inverted repeats (ITRs), which are cis elements necessary for viral DNA replication and packaging.

[0242] Adenoviruses are easy to grow and manipulate and exhibit a broad host range in vitro and in vivo. This group of viruses can be expressed at high titers (e.g., 10 9 -10 11 The adenovirus vector is a vector that is designed to deliver adenoviruses to the host cell. The adenovirus vector is designed to deliver adenoviruses to the host cell in a manner that allows the delivery of ...

[0243] Retrovirus is a group of single-stranded RNA viruses, which have the characteristics of being able to convert its RNA into double-stranded DNA in infected cells by reverse transcription methods. The resulting DNA is then stably integrated into the cell chromosome in the form of a provirus and guides the synthesis of viral proteins. The integration causes the viral gene sequence to remain in the recipient cell and its offspring. The retrovirus genome contains three genes gag, pol and env, which encode capsid protein, polymerase and envelope components respectively. The sequence found from the upstream of the gag gene contains a signal for packaging the genome into virions. Two long terminal repeat (LTR) sequences are present at the 5' and 3' ends of the viral genome. These contain strong promoter and enhancer sequences, and they are also required for integration in the host cell genome.

[0244] In order to construct a retroviral vector, the nucleic acid encoding the relevant gene is inserted into the position of some viral sequences in the viral genome to produce a virus with replication defects. In order to produce virions, a packaging cell line containing gag, pol and env genes but without LTR and packaging components is constructed. When the recombinant plasmid containing human cDNA and retroviral LTR and packaging sequence is introduced into the cell line (for example, by calcium phosphate precipitation), the packaging sequence allows the RNA transcript of the recombinant plasmid to be packaged into the virion, which is then secreted into the culture medium. The recombinant retrovirus is then collected, optionally concentrated, and used for gene transfer. Retroviral vectors can infect a wide range of various cell types. However, integration and stable expression require the division of host cells.

[0245] A novel approach designed to allow specific targeting of retroviral vectors has recently been developed based on chemical modification of retroviruses by chemical addition of lactose residues to the viral envelope. This modification may allow specific infection of hepatocytes via the sialoglycoprotein receptor.

[0246] Other viral vectors can be used as expression vectors. Vectors derived from viruses such as vaccinia virus, adeno-associated virus (AAV), and herpes virus can be used.

[0247] In order to achieve expression of sense or antisense gene constructs, expression vectors can be delivered to cells. This delivery can be achieved in vitro (such as in laboratory procedures for transforming cell lines) or in vivo or ex vivo (such as in the treatment of certain disease conditions). As described above, one delivery mechanism is via viral infection, in which the expression construct is encapsidated in infectious viral particles.

[0248] Several non-viral methods for transferring expression constructs into cultured mammalian cells are also contemplated. These methods include calcium phosphate precipitation of DEAE-dextran, electroporation, direct microinjection, DNA-loaded liposomes and lipofectamine-DNA complexes, cell sonication, gene bombardment using high-speed microparticles, and receptor-mediated transfection. Some of these techniques can be successfully adapted for in vivo or ex vivo use.

[0249] Pharmaceutical compositions - When clinical applications are contemplated, pharmaceutical compositions, gene delivery vectors or engineered cells can be manufactured in a form suitable for the intended application. Generally, this requires preparing compositions that are substantially free of pyrogens and other impurities that may be harmful to humans or animals.

[0250] Appropriate salts and buffers are used to stabilize the delivery vector and allow it to be taken up by the target cells. When the recombinant cells are introduced into the patient, a buffer will also be used. The aqueous composition of the present invention comprises an effective amount of a carrier to the cell, which is dissolved or dispersed in a pharmaceutically acceptable carrier or aqueous medium. The phrase "pharmaceutically or pharmacologically acceptable" refers to molecular entities and compositions that do not produce adverse, allergic or other adverse reactions when administered to animals or humans. As used herein, "pharmaceutically acceptable carriers" include any and all solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic agents, and absorption delay agents, etc. The use of such media and medicaments for pharmaceutically active substances is well known in the art. Unless any conventional media or reagents are incompatible with the carrier or cells of the present invention, it is expected that they are used in therapeutic compositions.

[0251] The expression vector and delivery vehicle can be administered via any common route, as long as the target tissue is available via the route. This includes oral, nasal, buccal, rectal, vaginal or topical. Alternatively, administration can be by in situ, intradermal, subcutaneous, intramuscular, intraperitoneal or intravenous injection. The composition will generally be administered in the form of a pharmaceutically acceptable composition.

[0252] Based on the established purpose, the effective amount of the therapeutic agent is determined. The term "unit dose" refers to physically dispersed units suitable for use in subjects, each unit containing a predetermined amount of a therapeutic composition calculated to produce the desired response associated with its administration, i.e., the appropriate route and treatment regimen. The amount to be administered (both according to the number of treatments and the unit dose) depends on the subject to be treated, the subject's state, and the protection required. The exact amount of the therapeutic composition also depends on the physician's judgment and is unique to each individual.

[0253] Therapeutic kits - All necessary materials and reagents required to inhibit tumor cell proliferation can be assembled together in a kit. This will generally include the selected expression vector, virus or cell. Also included may be various culture media for replication of the expression vector and host cells for said replication. The kit will contain different containers for each individual reagent. The kit may also include an instruction sheet defining (i) administration of an antisense expression vector construct; (ii) an antisense expression virus; and (iii) an antisense expression cell.

[0254] In some embodiments, interference (iRNA or siRNA) is provided. In some embodiments, the iRNA is complementary to the breakpoint of the fusion gene.

[0255] Methods related to clinical outcome findings

[0256] Tables 15 and 39 provided herein contain more than 100 gene events of various cancer types, including mutations that gain function, mutations that lose function, gene amplification / deletion in peaks and fusion events, which are related to clinical outcomes with high statistical significance (q<0.1). Therefore, provided herein are methods for delivering treatment to a subject, methods for determining whether a subject is receiving treatment, methods for determining whether a treatment is delivered, and methods for delivering reports. In certain illustrative embodiments, the treatment is a drug. As a non-limiting example, the drug may be a drug listed in Tables 8, 16, and 17, particularly when the method involves a gene event that affects the gene listed for the drug in Tables 8, 16, and 17. In other examples, as discussed herein, the drug may be any drug approved by a regulatory agency, or any drug in a development stage before approval.

[0257] Therefore, in another embodiment, a method for delivering treatment to a subject is provided, wherein the method includes the gene event identified in the detection table 15, and treating the subject, wherein the treatment is considered to actively affect the clinical outcome of the cancer patient with the gene event and / or is considered to affect the biological path related to the gene event. This embodiment can be regarded as a method for determining whether a subject is receiving treatment or a method for determining whether to deliver treatment to a subject or perform treatment. Therefore, a method for determining whether a subject is receiving medicine is provided herein, wherein the method includes the gene event identified in the detection table 15 and / or 39, and then if the detected gene event is listed in the table 15 and / or 39, then the subject is delivered with medicine, wherein the medicine is considered to actively affect the clinical outcome of the patient with the gene event. In the illustrative aspects of these embodiments, the gene event is related to the gene found in the table 8, 16 and 17, and the medicine is listed in the table 8, 16 and 17, as the partner of the gene. The subject is typically a subject suffering from the cancer of the type listed in the table 15 and / or 39. In an illustrative aspect of this embodiment, the genetic event is associated with a poorer prognosis in a subject suffering from a cancer, typically a cancer listed in Tables 15 and / or 39, the subject's poorer prognosis being associated with the genetic event.

[0258] In another embodiment, a method for delivering a report is provided herein, wherein the method includes detecting a gene event identified in Table 15 and / or 39, and delivering a report that provides a predicted clinical outcome associated with the gene event of the cancer of a subject to a medical professional. As a non-limiting example, a medical professional may be a physician, a genetic counselor, or other medical professional. Typically, a physician, a genetic counselor, or other medical professional has a professional relationship with the subject, such as a patient / doctor relationship. The report may be a paper report or may be an electronic report delivered to a medical professional via a computer network. The method and report may include one or more of the gene events identified in Table 15 and / or 39 and the associated prognosis.

[0259] In another embodiment, a method for determining which treatment to administer to a subject is provided herein, the method comprising detecting the gene events listed in Table 15, and administering treatment depending on the gene events detected. In an illustrative embodiment, if the detected gene event is associated with a poor prognosis, then the treatment is an aggressive treatment, such as a treatment in which the patient will fall into more pain and suffering due to the treatment. In a related embodiment, if the detected gene event is associated with a poor prognosis, then the treatment is a more aggressive treatment, and if the detected gene event is another gene event, particularly if the detected gene event is identified as indicating a good prognosis in Tables 15 and / or 39, then the treatment is a less aggressive treatment. For example, if AADAC gene deletion, CHD1L gene, FMO5 gene, or PRKAB2 gene amplification or a combination thereof is detected in a lung cancer adenocarcinoma patient, then the patient can be treated with an aggressive chemotherapy drug regimen. If these gene events are not detected in the patient, then the patient can be monitored, but chemotherapy drugs may not be administered.

[0260] In another embodiment, a method for determining whether to treat a cancer patient is provided herein, the method comprising detecting a gene event listed in Table 15 and / or 39, and if a gene event associated with a poor prognosis in Table 15 is detected, the subject is treated. In another embodiment, a method for determining whether to treat a cancer patient is provided herein, the method comprising detecting a gene event listed in Table 15 and / or 39, and if a gene event associated with a good prognosis in Table 15 and / or 23 is detected, the subject is not treated. In another embodiment, a method for determining whether to treat or monitor a cancer patient is provided herein, the method comprising detecting a gene event listed in Table 15 and / or 39, and if a gene event associated with a good prognosis in Table 15 and / or 39 is detected, the subject is monitored but not treated. If monitoring detects recurrence or progression of cancer, treatment may be administered later.

[0261] In certain aspects of these embodiments of the invention involving methods provided herein (e.g., methods of delivering a treatment to a subject, or methods of determining whether to deliver a treatment to a subject, or methods of determining which treatment to administer or deliver, or methods for delivering a report) based on clinical outcomes associated with the genetic events in Tables 15 and / or 39, a subject can be identified as having any type of genetic event and any specific genetic event listed in Tables 15 and / or 39. For example, a genetic event can be a gain-of-function mutation, a loss-of-function mutation, a gene amplification or deletion (typically an intra-peak gene amplification / deletion), or a fusion event. In certain illustrative embodiments, a q value of 1×10 is identified in Tables 15 and / or 39. -3 or smaller, 1×10-4 or smaller or 1×10 -5 Or smaller gene events. In certain aspects, the gene events are listed in Tables 15 and / or 39 as involving druggable genes. For example, the gene events may be gene events listed in Tables 15 and / or 39 that are associated with genes that are preclinical drug targets. As a non-limiting example, a method for determining which treatment or treatment course to administer to a patient with ovarian cancer (e.g., ovarian serous cystadenocarcinoma) is provided herein, wherein the method includes detecting or otherwise determining the amplification of ID1 or BCL2L1 genes and administering the treatment. In illustrative embodiments, the treatment is an approved treatment for BCL2L1, such as a currently FDA-approved BCL2L1 treatment, in which BCL2L1 amplification is detected.

[0262] Skilled professionals know methods for detecting the types of gene events listed in Tables 15 and / or 39. Those methods may include nucleic acid sequencing methods or amplification methods, such as PCR or isothermal amplification methods or combinations thereof. Those methods may include providing primers designed to bind to genes identified in Tables 15 and / or 39 or to the upstream of genes identified in Tables 15 and / or 39. Therefore, reaction mixtures and kits are provided herein, which include nucleic acid samples of subjects and one or more primers that bind to genes identified in Tables 15 and / or 39 or to their upstream. Typically, the genes are associated with gene events in Tables 15 and / or 39, and the subjects suffer from cancers identified in Tables 15 and / or 39 as having a prognosis associated with the gene events. The kits may also include control nucleic acids bound by primers as disclosed herein for various embodiments of the present invention. The reaction mixtures may also include polymerases as disclosed herein for various embodiments of the present invention.

[0263] In certain aspects of these embodiments of the invention involving methods provided herein (e.g., methods of delivering treatment to a subject, or methods of determining whether to deliver treatment to a subject, methods of determining which treatment to deliver, or methods for delivering reports to medical professionals) based on clinical outcomes associated with the genetic events in Tables 15 and / or 39, the genetic events may include more than one genetic event identified in Tables 15 and / or 39. In certain aspects, the method according to this embodiment detects 2, 3, 4, 5, 6, 7, 8, 9, 10 or more genetic events identified in Table 15, particularly those genetic events identified as having the same prognosis for a given cancer type. For example, the method may include detecting genetic events in a breast cancer patient, and administering treatment to the patient when the detected genetic events include gene amplification of two or more of the BRF2, ERLIN2, GPR124, PROSC, and TAB11FI genes. In another example, the method includes detecting two or more genetic events in a subject suffering from low-grade glioma, and administering treatment to the subject, wherein the genetic event is at least two of the following: amplification of EGFR or SEC61G gene; amplification of CDK4, CYP27B1, MARCH9, TSPAN31 or AGAP2 gene; gain-of-function mutation in EGFR gene; or deletion of CDKN2A, CDKN2B or MTAP gene. In another aspect, the method includes detecting a genetic event associated with a poor prognosis, and the genetic event is identified in Table 8, 16, 17, Table 15 and / or 39 as a target of a drug in current preclinical trials or an approved drug (such as an FDA-approved drug).

[0264] In certain aspects of these embodiments of the invention that relate to the methods provided herein (e.g., methods of delivering treatment to a subject, or methods of determining whether to deliver treatment to a subject, or methods of determining which treatment to administer or deliver, or methods for delivering reports) based on clinical outcomes associated with the genetic events in Tables 15 and / or 39, the genetic event may be a specific genetic event identified in one of the other tables herein. A skilled artisan can identify which general type of genetic event in Tables 15 and / or 39 will fall into a specific genetic event in one of the other tables.

[0265] Computer-implemented systems

[0266] Computer system can be used in certain embodiments of the present invention.In various embodiments, computer system can include bus or other communication mechanism for conveying information, and processor for processing information coupled with bus.In various embodiments, computer system 100 can also include memory (which can be random access memory (RAM) or other dynamic storage device) coupled to bus to measure base call and instruction to be executed by processor.Memory can also be used for storing temporary variables or other intermediate information during execution of instruction intended to be executed by processor.In various embodiments, computer system can further include read-only memory (ROM) or other static storage device coupled to bus to store static information and instruction for processor.Storage device (such as disk or CD) can be provided, and it is coupled to bus to store information and instruction.

[0267] In various embodiments, the computer system may be coupled to a display, such as a cathode ray tube (CRT) or a liquid crystal display (LCD), via a bus to display information to a computer user. Input devices, including alphanumeric and other keys, may be coupled to the bus to communicate information and command selections to the processor. Another type of user input device is a cursor control, such as a mouse, trackball, or cursor direction keys, for communicating directional information and command selections to the processor and for controlling cursor movement on the display. This input device typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), which allows the device to specify a position in a plane.

[0268] The computer system can perform the teachings of the present invention. Consistent with certain embodiments of the teachings of the present invention, results can be provided by the computer system 100 in response to a processor executing one or more sequences of one or more instructions contained in a memory. Such instructions can be read into the memory from another computer-readable medium (such as a storage device). The execution of the sequence of instructions contained in the memory can cause the processor to perform the methods described herein. Alternatively, hard-wired circuits can be used instead of or in combination with software instructions to implement the teachings of the present invention. Therefore, the embodiments of the teachings of the present invention are not limited to any specific combination of hardware circuits and software.

[0269] In various embodiments, the term "computer-readable medium" as used herein refers to any medium that participates in providing instructions to a processor for execution. Such media can be in many forms, including (but not limited to) non-volatile media, volatile media, and transmission media. Examples of non-volatile media can include (but are not limited to) optical or magnetic disks, such as storage devices. Examples of volatile media can include (but are not limited to) dynamic memory, such as memory. Examples of transmission media can include (but are not limited to) coaxial cables, copper wire, and optical fiber, including the wires that comprise a bus.

[0270] Common forms of non-transitory computer-readable media include, for example, floppy disks, floppy disks, hard disks, magnetic tape, or any other magnetic media, CD-ROMs, any other optical media, punch cards, paper tape, any other physical media with patterns of holes, RAM, PROM and EPROM, flash EEPROM, any other memory chip or cartridge, or any other tangible media from which a computer can read.

[0271] According to various embodiments, instructions configured to be executed by a processor to perform the method are stored on a computer-readable medium. A computer-readable medium may be a device that stores digital information. For example, a computer-readable medium includes a compact disc-read-only memory (CD-ROM) as known in the art for storing software. The computer-readable medium is accessed by a processor adapted to execute the instructions configured to be executed.

[0272] According to the teachings and principles embodied in the present application, methods, systems and computer-readable media are provided that can effectively collect, analyze, store, transfer, retrieve and / or distribute information (including genomic and / or patient information) among multiple locations and entities.

[0273] In one embodiment, a system for determining whether one or more gene fusions and / or variants are present in a sample is provided. The system can further determine to identify disease conditions (such as cancer) associated with one or more gene fusions and / or gene variants and determine appropriate treatment based on the mutation status. In certain embodiments, the system includes a processor that communicates with a sequencing instrument that receives sequencing data.

[0274] In some embodiments, the processor can perform one or more variant calls.In some embodiments, the processor can provide filtering and / or annotation prediction.

[0275] Examples

[0276] In the following example, methods are used to identify gene fusions and gene variants associated with a panel of 19 cancers in 4,225 cancer patient samples. The gene fusions and gene variants are then used to generate diagnostic methods for confirming cancer predisposition, diagnosing cancer, staging cancer, providing prognosis, and identifying druggable cancers. Methods are provided to provide targeted therapies for cancer based on the identification of gene fusions.

[0277] Example 1: High-throughput systematic analysis of paired-end next-generation sequencing data to characterize the gene fusion landscape in cancer 4,225 cancer patient samples across 19 diseases were processed with deFuse (McPherson et al. "deFuse: an algorithm for gene fusion discovery in tumor RNASeq data" PLoS Comp. Bio. 2011) and TopHat (Kim et al. "TopHat-Fusion: an algorithm for discovery of novel fusion transcripts" Genome Biology 2011) gene fusion calling software using a cloud-based computing infrastructure. Filtering criteria for gene fusion events were identified that enriched for high-confidence chemically validated gene fusion events.

[0278] Gene fusions encode oncogenic driver genes in hematology and solid tumors, and are often associated with significant clinical responses and appropriate targeted agents. Large-scale parallel double-end sequencing can identify structural rearrangements in tumor genomes and transcriptomes. However, the computational methods for identifying gene fusions vary (still evolving) and are essentially trained on cell line data. A systematic approach has been developed to characterize known oncogene fusions and to discover novel gene fusions in cancer. RNASeq data from approximately 3,400 clinical cases of 16 cancer types were obtained from the Cancer Genomics Hub (CGHub) of the Cancer Genome Atlas (TCGA). The performance of several gene fusion callers was investigated, and two (deFuse and TopHat) were selected for further method development with the goal of supporting both single-end and double-end data. An analysis pipeline was developed and executed in parallel on a high-performance computing cluster. Aggregated data were filtered and annotated as a post-processing step to allow exploratory analysis of various filters. To the data set optimization filtering method including known standards (such as TMPRSS2.ERG in prostate adenocarcinoma, PML.RARA in acute myeloid leukemia etc.), thus enrichment has these and other gene fusions of correct 5'-3' orientation, excludes the situation of having fuzzy breakpoint and spanning read segment, comparison error and read-through transcript from adjacent gene simultaneously.Based on the unique gene and unique gene that participate in fusion together with multiple collocations to the occurrence in specific disease separately, summarize the predicted fusion.In the case of predicted fusion and added important verification evidence positive, observe expression raising after the predicted breakpoint of 3' gene.Pan-disease fusion and multi-collocation fusion events widen the clinical population scope of gene fusion events.

[0279] All single-end data were processed using TopHat, and all paired-end data were processed using deFuse. TopHat has been shown to be effective with longer 75bp single-end data. The deFuse algorithm is not compatible with single-end data and has been designed to take advantage of read pairs. Preprocess data and detect fusions: The deFuse TopHat steps for all samples were performed in parallel on a high-performance computing cluster. The aggregated data were filtered and annotated as a post-processing step to enable filtering criteria to minimize false positive fusions. The list of prioritized fusions was validated using RNASeq exon expression data.

[0280] TCGA data source: from The Cancer Genomics Hub (CGHub, a repository of current TCGA genomic data - install Full Hypertext Transfer Protocol: / / cghub.ucsc.edu / ) to obtain all RNASeq data for gene fusion analysis. Table 9 lists the TCGA sample counts downloaded and processed for M2 and M3:

[0281] Table 9: Processed TCGA samples

[0282]

[0283]

[0284] *Single-end TCGA disease

[0285] BAM files were downloaded from CGHub using its Gene Torrent Software

[0286] With the goal of supporting both single-end and paired-end data, 4,374 paired-end samples were processed with deFuse and 584 single-end samples were processed with TopHat.

[0287] Broadly, the analysis pipeline consists of 5 main steps: 1. preprocessing raw data to obtain FASTQ files; 2. running the fusion calling program; 3. filtering breakpoints to relevant gene regions; 4. annotating breakpoints with the Oncomine transcriptome; and 5. outlining and prioritizing novel fusions that may be of interest.

[0288] The input into the fusion caller consists of RNASeq reads in FASTQ format, which requires converting the BAM files provided by TCGA into one or two FASTQ files of (respectively) single-end or paired-end data.

[0289] A custom SamToFastq converter was developed to generate FASTQ files from TCGA BAM files. In addition to allowing systematic conversion of all double-end RNASeq TCGA BAMs, the SamToFASTQ converter has other advantages over other conversion tools. First, it is written in C and compiled to run faster and reduce processing time. Second, it combines several validation steps to ensure proper mate pairing and consistent mate pair ordering in the output FASTQ files, both of which are input requirements for the fusion calling program.

[0290] Three cancer types (COAD, READ, UCEC) were available only as single-end RNASeq data. For single-end BAM file conversion, the program BamTools ( http: / / github.com / pezmaster31 / bamtools ) to generate FASTQ files.

[0291] Integration Figure 1The diagram is the relative level of the result filtering performed by the two calling programs. As part of the analysis, the "Level I" data, the output potential_fusion.txt file from TopHat-Fusion Post and the output results.classify.tsv file from deFuse were integrated. The integration step involves converting the reported breakpoints into breakpoints based on the genomics coordinate system and merging into a common file format.

[0292] Breakpoint filtering - The approximately 5.5 million predictions from the "Level I" output of the caller were filtered, retaining only those calls where each breakpoint was in the 5'UTR or CDS region of a RefSeq transcript (refGene obtained from UCSC on ca. 18 July 2012). This was done to enrich the predicted fusions for those containing functional gene regions. Breakpoints predicted to occur in intronic sequences were also excluded, yielding a set of 423,587 predicted chimeras.

[0293] Breakpoint Annotation - For each pair of breakpoints, only one transcript is retained per Entrez ID. This ensures consistency in annotating breakpoints at the same position. However, predicted breakpoints at different positions of the same gene pair may still produce multiple transcripts representing a pair of genes, which is evidence of the possibility of alternative transcripts.

[0294] The basic annotation from the caller is discarded because it is the default annotation source based on each corresponding caller. However, some output fields of both TopHat and deFuse are retained to help prioritize the predicted fusions. In addition, some annotation features that are not explicitly reported by the caller are inferred from other caller features.

[0295] Inferred properties - support and span read counts are obtained from each caller, and are summarized in Reads Span and Reads Span Support. The latter is the sum of the reads spanning the fusion and the reads supporting the fusion. The breakpoint sequence reported by the caller is trimmed to include 50 bases on each side of the fusion, and merged into the breakpoint sequence. The fusion breakpoint is depicted by "|". Since none of the callers provide a decisive '5'' or '3'' mark, the relative 5'-3' orientation of the fusion partner is inferred by combining the caller parameters with the gene chain annotation. If there are inferred 5' and 3' partners for a given gene fusion call, the effective orientation field is marked as "Y".

[0296] RepeatMasker Annotation - Each predicted breakpoint position was also annotated with a RepeatMasker feature near the breakpoint. This was done to identify breakpoints in highly repetitive genomic regions where alignment errors are likely to affect the prediction of chimeric transcripts. In each fusion prediction, if any of the breakpoints of the flanking sequences overlapped with a RepeatMasker element by 12 or more bases, the RepeatMasker overlap field was set to 1. The frequency of overlapping fusion calls was used for Oncomine prioritization described below, so that gene fusion predictions with lower overlap frequencies were considered higher quality.

[0297] Fusion Exon Expression Imbalance - Reproducible Oncomine Prioritized Fusions are observed using RNASeq exon expression data downloaded using the GDACFirehose tool to provide secondary evidence for true positive fusion events by looking for imbalance in exon expression before and after the breakpoint call. Specifically, if the expression of the 3' partner is affected by the promoter region of the 5' partner, then exon expression should increase after the predicted breakpoint. This effect is particularly visible when looking at fusion vs. non-fusion patient samples.

[0298] The RPKM RNASeq values ​​for each patient are listed as Gene Annotation Format (GAF) features corresponding to a composite of UCSC exons from several different gene definitions (including Refseq). Compendia processed fusion breakpoints were mapped to GAF features. 80.8% of the 396,298 Refseq exons mapped perfectly to the GAF features in the curves shown below. The Refseq exon and GAF ​​feature pairs that produced the greatest overlap were selected and reported.

[0299] A value called rg_pct provides a measure of the quality of mapping of a given Refseq exon to a GAF feature based on the formula: rg_pct = overlap / length refseq *Overlap / Length GAF特征

[0300] Mappings with an rg_pct value of 1 overlap perfectly, while values ​​less than 1 indicate that the refseq exon or GAF feature does not map to the exact same genomic region and the RPKM value may be questionable. Due to the disease coverage shown in the bar graphs, RNASeq V2 data were selected for all diseases except OV, STAD and LAML.

[0301] Fusion exon expression was manually reviewed for expression imbalance for a subset of Oncomine priority fusions that met the following criteria: 1. Recurrent Oncomine priority fusions; 2. Oncomine priority fusions listed in the Mitelman database; 3. One fusion partner was an Oncomine gain-of-function oncogene and involved at least 3 Oncomine priority fusions; and 4. One fusion partner was listed in the Sanger Cancer Gene Census (http: / / www.ncbi.nlm.nih.gov / pubmed / 21696664). : / / www.sanger.ac.uk / genetics / CGP / Census / ) and involving at least 3 Oncomine priority fusions.

[0302] A total of 994 gene fusions met these criteria and were manually reviewed for exon expression imbalance by assigning a rating of "supported," "refuted," "neutral," or "not tested" to the gene fusion calls.

[0303] Experts use the following criteria to assign ratings: Supported: The fused sample has a highly expressed 3' fusion partner after the breakpoint, making the fused sample an outlier for the patient population. Before the breakpoint, expression of the 3' partner should be lower than after the breakpoint. Refutation: Very low average expression of the 5' partner (<5 RPKM) or the average expression of one partner is much lower (about 1 / 10) compared to the other. Neutral: Neither satisfied support Not satisfied refute Guidelines. Assign fusions that have not been manually reviewed Untested Rating.

[0304] Fusion Summary - Summarizes fusions within a disease based on the occurrence of unique gene pairs as well as based on the occurrence of individual genes (possibly with multiple partners).

[0305] Fusion level summary - For unique fusion pairs (unique by Entrez ID pair), the number of samples within a disease that had at least one fusion prediction according to either caller is the sample count of the fusion. Since multiple breakpoints for the same gene pair can be reported in one sample and across multiple samples, the number of unique fusion pairs within each disease represented by 424K+ fusion calls is 49,588. Table 10 shows the properties summarized for a given fusion partner pair across individual predictions:

[0306] Table 10:

[0307] If genes were <1 Mb apart on the genome, a proximity flag was set for the fusion, and a defuse flip flag was set in individual fusion predictions for ≤75% of these fusion partners.

[0308] Mitelman Cross-reference - Individual unique fusions cross-referenced to the Mitelman database of genomic aberrations (HTTP : / / cgap.nci.nih.gov / Chromosomes / Mitelman , downloaded on February 25, 2013). Matching is based on gene name rather than disease type. Therefore, a gene fusion reported in a disease may have appeared in a different disease type in the TCGA data set. The gene fusions summarized at the gene level are cross-referenced to the Mitelman database based on gene name. Therefore, genes reported in Mitelman as having different histology or completely different aberration types (e.g., larger chromosome level deletions rather than fusions) compared to predicted unique fusions have more potential.

[0309] Normal Sample Fusion Blacklist - To reduce the number of false positive fusions, 344 double-ended normal samples across 10 diseases were processed using the same deFuse pipeline described above. A total of 56,579 total fusion calls consisting of 6,024 unique fusions were observed. Of the 49,588 unique gene fusion events, 11,801 of these calls were observed in normal samples. These normal sample fusion calls were used to generate a blacklist, and thereby remove these false positives from the Oncomine priority gene fusions.

[0310] Paralogous fusion partner blacklist - The blacklist of fusions between members of a paralogous gene family was compiled using two strategies: 1) manually detecting high frequency fusion partner gene names and 2) comparing the first 3 characters of all priority fusion partner gene names. In the latter strategy, fusion partners were validated as "paralogous" using HomoloGene, Ensembl and SIMAP and subsequently included in the final blacklist. The blacklist consists of 375 unique paralogous gene fusions and is used to remove false positives from Oncomine priority gene fusions.

[0311] Example 2: NGS mutation method for identifying cancer-associated genetic variants

[0312] The goal of data integration of genetic variants is to form the most complete set of mutation data currently available from TCGA.

[0313] Data Sources - For this version, the following were integrated: TCGA mutation data from the Broad GDAC Mutation_Packager 2013_02_22stddata build, Level 2 (public, experimentally unvalidated) data available from the TCGA DCC as of March 1, 2013, and mutation data generated by Compendia from TCGA raw data for prostate adenocarcinoma.

[0314] CBI mutation calls - There is a problem that the prostate adenocarcinoma mutation calls available from TCGA are of low quality and produce false positive 'gain of function' predictions. Therefore, all calls for this disease are derived from CBI's own mutation calling pipeline, which is closely similar to the method used by the TCGA cancer type working group to generate publicly available mutation calls.

[0315] Table 12: Data source selection

[0316]

[0317] Data cleaning - Perform some simple cleansing operations to remove duplicate mutation records present in the source data. Remove duplicate mutations from various tumor / normal aliquot pairs from the same patient sample. A total of 25 "hypermutation gene" samples (mutation counts > 5,000 per sample) were also excluded from the downstream analysis pipeline. In some diseases (such as uterine body endometrioid carcinoma), several highly mutated samples may dominate the overall mutation counts and dilute the results of the mutation reproducibility analysis necessary for the mutation and gene classification process of the Compendia company.

[0318] Mutation Annotation: A. Compendia Annotation - The Compendia method for defining mutations therefore relies on accurate variant annotation; mutations are re-annotated using a standard annotation pipeline that ensures mutations across disease types are consistently assessed and subject to common interpretation during nomination of potential oncogenes or tumor suppressor genes.

[0319] Mutations obtained from TCGA were processed by KenPendia according to the following general steps: 1. Each mutation was first re-annotated using the KenPendia transcriptome. Successfully annotated mutations received annotations from KenPendia, while the rest retained annotations obtained from TCGA. Annotations included: variant classification, variant position, variant change. 2. Redundant annotations for mutations in multiple transcripts were removed. 3. Mutations located outside the relevant gene region were removed. 4. Mutations without valid gene Entrez IDs were removed.

[0320] A "mutation" is defined herein as a specific change at a genomic location, ie, chromosome, start, stop, reference base, alternative base, variant type (SNP, INS, DEL), etc.

[0321] An "annotation" is defined herein as a set of transcript-specific properties that describe the effect of a mutation, i.e.: gene, transcript, variant classification, variant change, variant codon position, etc.

[0322] In the mutation annotation step, mutations obtained from TCGA were re-annotated against a set of standard transcripts compiled by Compendia, Inc. This transcript set included RefGene transcripts from the hg18 and hg19 genome builds obtained from UCSC.

[0323] Each mutation was individually mapped against contigs in the CBI transcriptome within a defined genome build. SNP mutations were mapped directly to their starting position, while for small insertion (INS) and deletion (DEL) mutations relevant positions were selected for mapping.

[0324] For mutations that were successfully mapped to transcripts, CBI mutation annotations were inferred relative to the transcripts. For mutations that could not be mapped, the more limited TCGA annotations were retained and variant positions for hotspot calculations were constructed based on genomic coordinates.

[0325] Below is a description of the criteria used to annotate mutations mapped to the CBI transcriptome:

[0326] Variant classification: For each mutation successfully mapped to a transcript, the variant classification was inferred using the position and sequence variant type of the mutation. This approach identified the following main mutation variant classifications:

[0327] Table 13: Classification of major mutation variants:

[0328]

[0329]

[0330] Variant position: The variant position of a mutation is the position used to identify genes with hotspot mutations, which are mutations of a certain classification observed at the same position in multiple tumor samples. In order to effectively identify the hotspots of each mutation, a mutation point identifier covering the mutation position, the amino acid or base identity affected, and the variant classification is constructed. Mutations that occur at the same position and are unrelated to the specific base changes they produce are aggregated. Therefore, only reference bases or amino acids are used to define variant positions. This ensures that mutations that affect the same codon or genomic position will be counted for possible hotspots, even if the alternative alleles they produce are different. For example, for a given gene, missense mutations V600E, V600F, and V600G will all have a variant position of V600, and will thus be gathered together when identifying hotspot mutations. When the amino acid level position is not available, the RNA level or genomic level position is used.

[0331] For mutations that do not map to the CBI transcriptome and therefore do not have a transcript-based position, the genomic position (start position) and reference nucleotide (reference allele) are used as variant positions independent of coding region or splice site proximity. The TCGA annotated variant classification is then added as a suffix. The variant changes (see below) for these mutations are undefined.

[0332] Variant changes: Variant changes provide HGVS-like information about the alternative allele changes of mutations (e.g., V600E). For SNP mutations in coding regions, variant changes are full HGVS protein-level sequence variant descriptions, indicating the alternative amino acid. For SNPs outside of coding regions, alternative allele nucleotide bases are provided. For mutations that do not map to the CBI transcriptome, the variant classification from TCGA is retained.

[0333] Transcript filtering: To avoid retrieving multiple transcripts, and therefore avoid retrieving multiple annotations for a single mutation within a gene, only one transcript per mutation per gene (unique Entrez ID) is maintained. If a mutation maps to several transcripts of a gene, only one transcript is selected. However, if a mutation maps to several genes, only one transcript is selected for each gene. Therefore, a mutation can receive two different annotations, but only if it comes from transcripts with different EntrezIDs. In fact, any mutation of the same variant classification at the same genomic position is always assigned to the same transcript, and therefore will be in the same reference frame when calculating the reproduction of hotspot identification.

[0334] Gene region filtering: All mutations were further filtered by variant type and category to avoid including less relevant mutations into gene function analysis. Mutations that did not resolve to gene regions were filtered out, either because they fell significantly outside the transcript or because they were in a position not associated with the RefSeq gene. These mutations were evident by their lack of gene identifiers or membership in the following variant categories: introns, 5' flanking sequences, IGRs, and miRNAs. Mutations with variant types of DNP, TNP, ONP, Complex_substitution, and Indel were also filtered out, as their annotations were not supported by the pipeline.

[0335] Classify mutations as hotspot, deleterious, or other - The next step in the analysis pipeline identifies mutations that recur in multiple samples based on the variant position of the mutation and classifies them into hotspot, deleterious, or other variant categories. For this step and subsequent frequency calculations, mutations for each disease type are processed independently. Only mutations of the same variant classification are calculated together, so for example, missense mutations and silent mutations at the same position are counted separately.

[0336] To identify driver gene events, each mutation given an Entrez Gene Id was classified as "harmful" or "hotspot". If a mutation was observed in the same variant position in 3 or more tumor samples, the mutation was considered to be 'recurring'. If a mutation is recurring and annotated with one of the following variant classifications: in-frame insertion / deletion, no stop, missense, non-coding exon, the mutation belongs to the "hotspot" variant category. If a mutation is annotated with one of the following variant classifications: frameshift insertion / deletion, nonsense, the mutation belongs to the "harmful" category. If a mutation does not meet the above criteria, the mutation is considered to be in the "other" variant category.

[0337] Naming of "Gain of Function" and "Loss of Function" Genes - Individual genes were classified into predicted functional categories, namely "Gain of Function", "Recurrent Other", and "Loss of Function", to reflect their relative enrichment in potentially activating or deleterious mutations.

[0338] Frequency of mutations: The mutation frequency of each gene was calculated relative to a given variant classification and variant class in all samples within a disease type. The overall mutation frequency of a gene within a disease was calculated by combining all mutations.

[0339] Mutation significance: A hotspot p-value is calculated for each gene within a disease by selecting the most reproduced mutation m and using sampling to determine the probability p of observing r or more mutations at that position. More specifically, where cm is the count of replicates with maximum multiplicity m. The P value for a transcript with a maximum multiplicity of one is defined as 1.0. The P value for a transcript with a maximum multiplicity that has never been observed is defined as 1e-5.

[0340] The hotspot Q value within each disease was calculated by counting the number of transcripts (N) that were mutated at least once and calculating the rank for each p-value. The q value for a given p-value is then Q = p*N / rank.

[0341] To assess whether a gene is significantly enriched for deleterious mutations compared to other genes, Fisher's exact test was performed comparing the deleterious mutation frequency of the gene in question with that of other genes given the background mutation rate. Nonsense mutations, frameshift insertions, and frameshift deletions were classified as deleterious mutations, while mutations of any other type (missense, etc., but not intergenic) were counted as others.

[0342] The deleterious Q value within each disease is calculated by counting the number of genes (N) that harbor deleterious mutations and calculating the rank for each. The q value for a given p value is then Q = p*N / rank.

[0343] Gene classification: Once the mutations are classified, individual genes are named into one of three categories: "Gain of Function", "Loss of Function" and "Recurrent Other". The classification is based on a combination of the relative frequency and significance of the mutations observed in the gene. The significance of each gene mutation is assessed by the p-value. Figure 2 The classification process in specifies the criteria used for gain-of-function and loss-of-function genes.

[0344] "Gain of function" genes will have a relatively high frequency of hotspot missense mutations and a low frequency of deleterious mutations, while "loss of function" genes contain a larger fraction of deleterious mutations. "Recurrent Others" tend to contain recurrent indel mutations, some of which (e.g., recurrent frameshift indels of 1 base) show signs of potential false positive calls that may be caused by local alignment errors.

[0345] Pan-cancer analysis - To summarize mutations across diseases, the same calculations were performed for the within-disease analysis, but without stratifying mutation records by disease. For the pan-disease gene classification, genes (unique by EntrezID) across all cancer types were summarized.

[0346] Example 3: Diagnostic Assays for Identifying Gene Fusions and / or Gene Variants in Cancer

[0347] Library preparation

[0348] PCR amplification of genomic DNA targets

[0349] The disclosed variants and fusion polynucleotides can be detected by sequencing of nucleic acids. This can be achieved by next generation sequencing (described below). The nucleic acid source for next generation sequencing can include fresh frozen paraffin embedded (FFPE) samples.

[0350] Multiplex polymerase chain reactions were performed to amplify 384 individual amplicons within one genomic DNA sample. A pool of more than 32,000 primers was developed, covering more than 100 gene variants or fusion polynucleotides. Each primer in the primer pool was designed to contain at least one uridine nucleotide near the end of each primer. Each primer was also designed to selectively hybridize with a specific gene, gene variant, or fusion polypeptide of a nucleic acid sample by forming a primer pair, and promote amplification.

[0351] To a single well of a 96-well PCR plate, add 5 μl of a primer pool (containing 384 primer pairs at a concentration of 15 μM in TE), 10-50 ng of genomic DNA, and 10 μl of an amplification reaction mixture (2× AmpliSeq HiFi Master Mix, which can include glycerol, dNTPs, and Taq High Fidelity DNA Polymerase, Invitrogen, Catalog No. 11304) and DNase / RNase-free water (Life Technologies, CA, Part No. 600004) were brought to a final volume of 20 μl.

[0352] Seal the PCR plate and load it into a thermal cycler ( The PCR system 9700 Dual 96-well Thermal Cycler (Life Technologies, CA, part numbers N8050200 and 4314445) was run using the following temperature profile to generate the pre-amplified amplicon library.

[0353] The initial hold phase was performed at 98° C. for 2 minutes, followed by 16 cycles of denaturation at 98° C. for 15 seconds, and ligation and extension phases for 4 minutes at 60° C. After cycling, the pre-amplified amplicon library was held at 4° C. until proceeding to the purification step outlined below.

[0354] Purification of amplicons from input DNA and primers

[0355] Found two wheels XP reagent (Beckman Coulter, CA) binding, washing, and elution at 0.6× and 1.2× volume ratios removes genomic DNA and unbound or excess primer. The amplification and purification steps outlined herein produce amplicons ranging from about 100 bp to about 600 bp in length.

[0356] In a 1.5 ml LoBind tube (Eppendorf, part number 022431021), the pre-amplified amplicon library (20 μl) was mixed with 12 μl (0.6× volume) of XP reagent (Beckman Coulter, CA). The bead suspension was pipetted up and down to thoroughly mix the bead suspension with the pre-amplified amplicon library. The samples were then pulse spun and incubated at room temperature for 5 minutes.

[0357] Place the tubes containing the samples on a magnetic rack (such as DynaMag TM -2 spin magnet (Life Technologies, California, part number 123-21D) for 2 minutes to capture the beads. Once the solution is clear, the supernatant is transferred to a new tube, to which 24 μl (1.2× volume) of XP beads (Beckman Coulter, California). The mixture was pipetted to ensure that the bead suspension was mixed with the pre-amplified amplicon library. The sample was then pulse-spun and incubated at room temperature for 5 minutes. The test tube containing the sample was placed on a magnetic rack for 2 minutes to capture the beads. Once the solution was clear, the supernatant was carefully discarded without disturbing the bead pellet. The desired pre-amplified amplicon library was then bound to the beads. Without removing the test tube from the magnetic rack, 200 microliters of freshly prepared 70% ethanol were introduced into the sample. The sample was incubated for 30 seconds while the test tube was gently rotated on the magnetic rack. After the solution was clear, the supernatant was discarded without disturbing the pellet. A second ethanol wash was performed and the supernatant was discarded. Any remaining ethanol was removed by pulse-spinning the test tube, and the residual ethanol was carefully removed without disturbing the pellet. The pellet was air-dried at room temperature for about 5 minutes.

[0358] Once the test tube is dry, the test tube is removed from the magnetic rack and 20 microliters of DNase / RNase-free water (Life Technologies, California, part number 600004) are added. The test tube is vortexed and pipetted to ensure that the sample is thoroughly mixed. The sample is pulse-spun and placed on the magnetic rack for two minutes. After the solution is clarified, the supernatant containing the eluted DNA is transferred to a new test tube.

[0359] Phosphorylated Amplicon

[0360] To the eluted DNA (approximately 20 μl), 3 μl DNA ligase buffer (Invitrogen, catalog number 15224041), 2 μl dNTP mix and 2 μl FuP reagent were added. The reaction mixture was mixed thoroughly to ensure homogeneity and incubated at 37°C for 10 minutes.

[0361] Ligating adaptors to amplicons and purifying ligated amplicons

[0362] After incubation, the reaction mixture proceeded directly to the ligation step. Here, the reaction mixture, now containing the phosphorylated amplicon library, was combined with 1 microliter of A / P1 adapter (20 μm each) (sold as a component of the Ion Fragment Library Kit, Life Technologies, part number 4466464) and 1 microliter of DNA ligase (sold as a component of the Ion Fragment Library Kit, Life Technologies, part number 4466464) and incubated at room temperature for 30 minutes.

[0363] After the incubation step, 52 μl (1.8× sample volume) The ATCC® reagent (Beckman Coulter, CA) was added to the ligated DNA. The mixture was pipetted thoroughly to mix the bead suspension with the ligated DNA. The mixture was pulse spun and incubated at room temperature for 5 minutes. The sample was pulse spun again and placed on a magnetic rack (such as DynaMag TM -2 spin magnet (Life Technologies, California, Part No. 123-21D)) for two minutes. After the solution is clear, discard the supernatant. Without removing the test tube from the magnetic rack, introduce 200 microliters of freshly prepared 70% ethanol to the sample. Incubate the sample for 30 seconds while gently spinning the test tube on the magnetic rack. After the solution is clear, discard the supernatant without disturbing the pellet. Perform a second ethanol wash and discard the supernatant. Remove any remaining ethanol by pulse spinning the test tube and carefully remove the residual ethanol without disturbing the pellet. Air dry the pellet at room temperature for about 5 minutes.

[0364] The pellet was resuspended in 20 microliters of DNase / RNase-free water (Life Technologies, California, part number 600004) and vortexed to ensure that the sample was thoroughly mixed. The sample was pulse spun and placed on a magnetic rack for two minutes. After the solution was clarified, the supernatant containing the connected DNA was transferred to a new Lobind tube (Eppendorf, part number 022431021).

[0365] Gap shifting and amplification of amplicon libraries and library purification

[0366] The ligated DNA (about 20 μl) was mixed with 76 μl PCR SuperMix High Fidelity (Life Technologies, California, part number 12532-016, sold as a component of the Ion Fragment Library Kit, Life Technologies, part number 4466464) and 4 microliters of Library Amplification Primer Mix (5 μM each) (Life Technologies, California, part number 602-1068-01, sold as a component of the Ion Fragment Library Kit, Life Technologies, part number 4466464) were combined and the mixture was pipetted thoroughly to ensure a uniform solution. The solution was applied to a single well of a 96-well PCR plate and sealed. The plate was loaded into a thermal cycler ( The PCR system 9700 Dual 96-well Thermal Cycler (Life Technologies, CA, part numbers N8050200 and 4314445) was run on the following temperature profile to generate the final amplicon library.

[0367] Nick translation was performed at 72° C. for 1 minute, followed by an enzyme activation phase at 98° C. for 2 minutes, followed by 5-10 cycles of denaturation at 98° C. for 15 seconds, and ligation and extension phases at 60° C. for 1 minute. After cycling, the final amplicon library was kept at 4° C. until proceeding to the final purification step outlined below.

[0368] In a 1.5 ml LoBind tube (Eppendorf, part number 022431021), the final amplicon library (approximately 100 μl) was mixed with 180 μl (1.8× sample volume) of XP reagent (Beckman Coulter, CA). The bead suspension was pipetted up and down to thoroughly mix the bead suspension with the final amplicon library. The samples were then pulse spun and incubated at room temperature for 5 minutes.

[0369] Place the tube containing the final amplicon library on a magnetic rack (such as DynaMag TM -2 spin magnet (Life Technologies, California, part number 123-21D)) for 2 minutes to capture the beads. Once the solution is clear, the supernatant is carefully discarded without disturbing the bead pellet. Without removing the test tube from the magnetic rack, 400 microliters of freshly prepared 70% ethanol are introduced to the sample. The sample is incubated for 30 seconds while the test tube is gently rotated on the magnetic rack. After the solution is clear, the supernatant is discarded without disturbing the pellet. A second ethanol wash is performed and the supernatant is discarded. Any remaining ethanol is removed by pulse spinning the test tube, and the residual ethanol is carefully removed without disturbing the pellet. The pellet is air-dried at room temperature for about 5 minutes.

[0370] Once the tube is dry, the tube is removed from the magnetic rack and 20 microliters of Low TE (Life Technologies, California, part number 602-1066-01) are added. The tube is aspirated and vortexed to ensure that the sample is thoroughly mixed. The sample is pulse-spun and placed on the magnetic rack for two minutes. After the solution is clarified, the supernatant containing the final amplicon library is transferred to a new Lobind tube (Eppendorf, part number 022431021).

[0371] Assess library size distribution and determine template dilution factor

[0372] The final amplicon library is quantified to determine the library dilution (template dilution factor) that yields a concentration within the optimized target range for template preparation (e.g., PCR-mediated addition of library molecules to the Ion Sphere TM The final amplicon library is typically prepared for downstream template preparation procedures using the Ion Library Quantification Kit (qPCR) (Life Technologies, part number 4468802) and / or the Bioanalyzer TM (Agilent Technologies, Agilent 2100 Bioanalyzer) quantification to determine the molar concentration of the amplicon library, from which the template dilution factor is calculated. For example, instructions for determining the template dilution factor by quantitative real-time PCR (qPCR) can be found in the Ion Library Quantification Kit User Guide (Life Technologies, Part No. 4468986), which is incorporated herein by reference in its entirety.

[0373] In this example, the 2100 Bioanalyzer TM One microliter of the final amplicon library preparation was analyzed using the Agilent High Sensitivity DNA Kit (Agilent Technologies, part number 5067-4626) to generate amplicon libraries with a size range of 135-205 bp and a concentration of approximately 5 × 10 9 The peak value of copies.

[0374] Proceed to template preparation

[0375] An aliquot of the final library was used to prepare DNA templates that were TMThe templates in the examples of the present invention were prepared using the Ion Xpress Template Kit (Life Technologies, Part No. 4466457) according to the manufacturer's instructions, which are incorporated herein by reference in their entirety. Once the template-positive Ion Sphere particles were enriched, aliquots of the Ion Spheres were loaded onto the Ion 314 TM The DNA was loaded onto a PCR chip (Life Technologies, part number 4462923) and subjected to analysis and sequencing as described in the Ion Torrent PGM Sequencer User Guide (Life Technologies, part number 4462917, which is incorporated herein in its entirety).

[0376] Example 4: Oncomine NGS integrated analysis approach to identify genetic events associated with clinical outcomes

[0377] Oncomine NGS integrated analyses are designed to bring together the largest possible set of core NGS data, enabling scientific workflows that interrogate relationships between data types and diseases, outlining the analyses at multiple levels of abstract biology, such as genes and pathways.

[0378] Data source (Oncomine is available from Life Technologies / Compendia Biosciences, Ann Arbor, Michigan and at http: / / www.oncomine.org)

[0379] The data for the integrated analysis were obtained from the following sources:

[0380] Fusion: Oncomine driver gene fusion

[0381] Mutation: Oncomine Pan-Cancer Driver Gene Mutation

[0382] CNV: Peak amplification and deletion data obtained from copy number data processed by Oncomine

[0383] DNA: DNA sequence (DNA-seq) continuous data processed by Oncomine

[0384] RNA: Normalized gene-level RNA sequence (RNAseq) continuous data

[0385] Clinical: Clinical and outcomes metadata managed by Oncomine

[0386] Pathway: Oncomine Pathway Definition

[0387] Fusion and filtering

[0388] Fusion data for integrative analysis were obtained from Oncomine NGS fusion data. Oncomine prioritized fusions are a priority process developed at Compendia that captures the properties of known true positive fusion events and characterizes a subset of observed gene fusions as high confidence priority fusions. Criteria for defining priority fusions include: valid 5' to 3' orientation, non-adjacent fusion partners, uniquely mapped spanning reads, non-parallel homologous fusion partners, not observed in normal tissues, and not overlapping with redundant regions in the genome.

[0389] Fusions were included and considered driver gene fusions if they were called by deFuse or Tophat, had "supporting" or "neutral" evidence for exon expression, and met one of the following 4 criteria:

[0390] Oncomine Prioritized Fusions + Reproducible

[0391] Oncomine prioritized fusions + Mitelman annotated

[0392] Oncomine Prioritized Fusions + One Partner is an Oncomine Gain of Function Gene involved in 3 or more Pan-Disease Prioritized Fusions

[0393] Oncomine Prioritized Fusions + One partner is a Sanger oncogene involved in 3 or more pan-disease priority fusions (http: / / goo.gl / JQBw9)

[0394] Mutation data and filtering

[0395] Mutation data for integrative analysis were obtained from Oncomine NGS mutation data. Individual genes were classified into predicted functional categories, i.e., "gain of function" and "loss of function", to reflect their relative enrichment in potential activating or deleterious mutations. This classification is based on a combination of the relative frequency and significance of mutations observed in genes assessed by p-values. "Gain of function" genes will have a relatively high frequency of hotspot missense mutations and a lower frequency of deleterious mutations, while "loss of function" genes contain a larger portion of deleterious mutations.

[0396] Copy number segmentation and quantification

[0397] DNA copy number data for each TCGA sample was obtained from Oncomine. Measurements of multiple reporters for a single gene were averaged.

[0398] Minimum Common Region (MCR) peak generation

[0399] In genes with reproducible amplification (4 or more copies) or deletion (1 or less copy), peak values ​​were independently identified in 25 cancer types by applying MCR analysis to Oncomine clinical samples. In order to define the peak value, continuous genomic regions (common regions) with significant abnormalities of multiple genes were first identified. In each common region, the peak value was defined as one or more genes where the abnormal sample count met or exceeded the peak threshold. In each cancer, the common region was defined as the region where the abnormal sample count met or exceeded the common region threshold. The baseline average of abnormal samples observed in all genes of each arm of each chromosome in each cancer was calculated.

[0400] mRNA gene expression

[0401] Expression data were obtained from TCGA standard data of Broad GDAC.

[0402] Clinical Data Management

[0403] Patient clinical data were obtained from TCGA and managed by Compendia. The types of data managed included demographic data, major clinical and histological disease subtypes, and clinical outcome data. All characteristics were standardized to be consistent across diseases.

[0404] Construction of clinically relevant subsets

[0405] Curated clinical data obtained from TCGA and Oncomine NGS data were used, and the rules in Table 14 were applied to define clinical subsets:

[0406] Table 14 - Rules for defining clinical subsets

[0407]

[0408] path

[0409] Manually curated Compendia pathway definitions were used to summarize gene-level aberrations in the integrative analysis. The pathways represent clinically relevant pathway modules, and several modules may cover major biological pathways, and a single gene may be present in one or more pathway module definitions.

[0410] Data Integration

[0411] Figure 3 The diagram in summarizes the data flow of integrating various data types into the Genetic Event Database (GEDB). All further analyses were performed using GEDB. The method has 4 main steps.

[0412] Mapping data to internal IA gene and patient dimensions

[0413] Define events and driver events for each data type

[0414] Rolling up individual events to the gene and pathway level

[0415] Merge events into the gene event database.

[0416] Gene and patient dimensions

[0417] Construct a single gene and patient dimension covering all patients and genes measured in all diseases and data types. Give genes and patient internal identifiers, and all data in IA are referenced to these identifiers to achieve gene name and patient barcode consistency. The unique identifier of a gene is the Entrez ID of the gene. The unique identifier of a patient is the TCGA patient barcode (the first 12 digits of the TCGA barcode).

[0418] Driver gene event definition

[0419] Mutations, fusions, and copy number events are defined based on the following criteria for genomic events:

[0420] Fusion: Priority Fusion Reproduced by Oncomine

[0421] Mutations: Oncomine driver gene mutations from pan-cancer driver genes

[0422] CNV: CBI-identified peaks and gene amplification / deletion within peaks

[0423] Genetic event definition and volume

[0424] A genetic event is a genomic aberration that represents an individual mutation, fusion, or copy number event, or a combination of events at the gene or pathway level. Figure 4 Flowchart 'rollup' shown in . When multiple events are combined to construct a rolled-up event, the measured set of patients for the rolled-up event becomes the intersection of patients measured for all 3 data types. Patient positives are only included if they are fully measured.

[0425] analyze

[0426] Once all driver genetic events were constructed, a set of analyses were performed on each genetic event, calculating frequencies, correlations, and relationships within disease (and pan-cancer when appropriate). Below is a brief description of each analysis:

[0427] frequency

[0428] Frequency is the occurrence of a driver event among the patients in whom it was measured. Frequencies were calculated within disease and pan-cancer.

[0429] Clinical relevance analysis

[0430] Each driver gene event is tested for its relevance relative to a set of available clinical subtypes. Each relevance is tested using Fisher's exact test by comparing the incidence of gene events in patients with one clinical subtype versus another clinical subtype. For example, overexpression of loss-of-function mutations can be tested in smokers versus non-smokers or in phase I versus phase II lung cancer. A total of 136 subtype pairs are tested for each event, and the characteristics defining the subtypes (some characteristics may be disease-specific) are listed below. A total of at least 4 patients (and at least 1 patient in each category) are required to perform the test.

[0431] Clinical outcome analysis

[0432] The correlation of each driver gene event with clinical outcome was tested using the log-rank test. Only the patient group with available clinical data was used for calculation, so the number of patients included in the test may be less than the number of patients measured for the driver gene event. At least 4 patients with positive driver gene events are required to perform the test. Survival time is presented in years, and individual survival / death events are clearly marked on the Kaplan-Meier curve. Correct the p-values ​​(q values) of multiple tests. Consider events with q values ​​less than 0.1.

[0433] The results of the analysis are presented in Tables 15 and 39. In Tables 15 and 39, the columns provide the following information:

[0434] The "Subset" column provides clinically relevant cancer types.

[0435] The p-value column is the p-value.

[0436] The q value column is the corrected p value. Events with q < 0.1 are included in the table.

[0437] The Number of Positives column is the number of patients who were positive for an event type.

[0438] The Total Patients column is the total number of patients evaluated.

[0439] The Cytoband column is the chromosomal location of the gene.

[0440] The Gene (Entrez ID) column is a list of genes and their corresponding Entrez IDs.

[0441] The Druggable Gene column indicates whether any gene is a target for a drug that is in active trial, approved, or otherwise marketed.

[0442] The KM Evidence column provides the Kaplan-Meier evidence. The KM evidence indicates whether the event type supports a good or poor prognosis in a particular cancer type.

[0443] Tables 15 and 39 contain more than 100 gain-of-function mutations, loss-of-function mutations, intra-peak gene amplification / deletion, and fusion events for various cancer types with q < 0.1. The genes within each event and cancer type are included along with their chromosomal location, druggability information, and clinical outcome associations, as indicated in the column information above.

[0444] Example 5 - Integrated data analysis

[0445] Oncomine NGS integrated analyses. Oncomine NGS integrated analyses are designed to bring together the largest possible core integrated genomic and phenomic data sets, enabling scientific workflows that interrogate relationships between data types and diseases, outlining the analyses at multiple abstract biological levels such as genes and pathways.

[0446] the term:

[0447] Aberration - a genomic structural variation or change in DNA; examples include: mRNA over- / under-expression, copy number amplification / deletion, mutation, and gene fusion.

[0448] Driver Genes - Aberrations identified as potential cancer driver genes by the Oncomine method described in this document; examples include gain-of-function mutations, gene amplifications in peak amplification regions, or gene fusions.

[0449] Volume Up - Summary of all mutations, fusions or copy number aberrations for a gene or pathway; only patients measured for all three aberration types are included in the volume up.

[0450] Hotspot mutations - mutations that are recurrent (n≥3) and classified as in-frame indels, nonstops, or missense.

[0451] Patient null group - group of patients measured for one genetic aberration

[0452] Patient positive group - group of patients with gene aberrations

[0453] Gene null group - the group of genes measured by the experimental platform used to assess genetic aberrations

[0454] Mitelman - a manually curated database of chromosomal aberrations and gene fusions in cancer from the literature (http: / / hypertext transfer protocol : / / goo.gl / PnXMT )

[0455] RPKM - "reads per kilobase per million"; a method for quantifying RNASeq data that normalizes the total read length and number of sequenced reads (Mortazavi et al. 2008)

[0456] RSEM - "RNA-Seq by Expectation Maximization" is a method for RNASeq data quantification used to estimate the best possible distribution of reads between expected transcripts providing relative transcript abundance as a fraction of the total read pool (Li and Dewey 2011)

[0457] Data sources. Efforts are made to collect the largest overlapping set of data available for each sample. The data in this version of the NGS integrated analysis browser are obtained directly or after Oncomine processing and analysis methods from The Cancer Genome Atlas (TCGA), Cancer Cell Line Encyclopedia (CCLE), COSMIC cell line project, and many research publications. Due to the uneven coverage of all data types in the source data set, some cancer types have more patients covered in multiple data types.

[0458] The Oncomine NGS mutation version for integrative analysis contains many manually curated datasets obtained from NGS mutation studies in peer-reviewed publications. For a complete list of publications that contributed mutation data for integrative analysis, see the Oncomine NGS mutation method data. The following datasets contain multi-dimensional NGS data, providing both mutation and copy number data. The copy number data of these datasets are processed in the same way as the copy number data obtained from TCGA.

[0459] Cell line data include mutation, fusion, and copy number datasets. Cell line data are processed in the same way as clinical tumor data, and mutation and fusion cell line data are obtained from Oncomine TM NGS Mutation and Oncomine TM NGS fusion power tools were obtained. Copy number data for cell lines were processed using the standard Oncomine copy number pipeline. Despite using two very different cell line datasets, CCLE and COSMIC, our standardization of cell line disease types and names has enabled us to cross-reference the two datasets and merge CCLE copy number data, COSMIC mutation data, and Oncomine fusion calls (based on CCLE RNASeq data). Therefore, the exomes of many cell lines in this release have been systematically characterized for all three types of aberrations. Cell line data are outlined using Oncomine cancer type definitions that will be directly comparable to tumor data, but the overview is performed separately for tumors and cell lines.

[0460] Phenomic Data

[0461] Clinical patient metadata management. Patient clinical data is obtained from primary sources and managed by Compendia. Data types managed include demographics, major clinical and histological disease subtypes, and clinical outcome data. All cancer type-independent characteristics (such as age or survival) are standardized to achieve consistency within cancer types. Certain disease stages are merged to obtain higher patient counts within one stage. For example, stage Ia and stage Ib can be combined into a modified stage I.

[0462] The following is a list of the most densely populated features and corresponding values ​​captured by the management method. Not all features are available for all patients.

[0463]

[0464]

[0465]

[0466] Features with the prefix "TCGA" are derived and curated from TCGA publications defining molecular subtypes of invasive breast cancer, glioblastoma, squamous cell lung cancer, and colorectal cancer.

[0467] Genomics Event Data: Fusion Data Filtering. Fusion data for the Integrative Analysis Browser was obtained from the Oncomine NGS fusion data release November 2013. Only fusions identified as Oncomine priority fusions were included in the Integrative Analysis Browser.

[0468] Oncomine Prioritized Fusions is a prioritization pipeline developed at Compendia that captures the properties of known true positive fusion events and characterizes a subset of observed gene fusions as high confidence priority fusions. Criteria used to define priority fusions include: valid 5' to 3' orientation, non-adjacent fusion partners, uniquely mapped spanning reads, non-parallel homologous fusion partners, not observed in normal tissues, and not overlapping redundant regions in the genome.

[0469] The patient null set of fusion data is the complete set of patient tumor samples processed in the fusion analysis; data for only one tumor sample (preferably, a primary, non-recurrent tumor) is retained for each patient. The gene null set is the genome in RefGene as of May 2012. Fusions are included in the integrated analysis browser if they are Oncomine priority fusions, have "supportive" or "neutral" exon expression evidence, and meet one of the following criteria:

[0470] Recurrence (occurrence in 2 or more patients)

[0471] Annotation in the Mitelman database of known structural variants

[0472] Contains gene partners that are Oncomine gain-of-function genes involved in 3 or more pan-disease priority fusions

[0473] Sanger oncogenes that contain 3 or more pan-disease priority fusions (HT : / / goo.gl / JQBw9 )'s genetic partners.

[0474] Mutation data filtering. Mutation data for the integrative analysis were obtained from the Oncomine NGS mutation data release in November 2013. Only non-silent mutations in coding gene regions were included in the integrative analysis browser.

[0475] The patient null set is the complete set of patients processed in the mutation analysis; data from only one tumor sample (preferably, a primary, non-recurrent tumor) is retained for each patient. The gene null set is the genome in RefGene as of March 2012.

[0476] Mutations with the following variant classifications are not included in the integrated analysis browser: silent, 5'UTR, 3'UTR, RNA, noncoding exon.

[0477] Calling Amplification / Deletion. DNA copy number data for each sample was obtained from the 2013Q4 Oncomine standard data build, in which all copy number data available from TCGA and manually curated publications up to October 2013 were normalized.

[0478] The patient blank group for this analysis is the patient group measured for copy number data up to October 2013 and the patient group measured in the manually managed publication. Each patient retains only the data of one tumor sample (preferably, primary, non-recurrent tumor). The gene blank group for this data is the Oncomine DNA copy number platform, based on the RefSeq coordinates (hg18) provided by UCSC RefGene in July 2009, and measures 18,796 genes. The measured values ​​of multiple reporter genes of a single gene are averaged.

[0479] The log2 of the estimated copy value was used to make amplification / deletion (amp / del) calls with cutoffs of >1.0 and <-1.0, respectively. No amp / del calls were made with a log2(estimated copies) of ≥-1.0 or ≤+1.0.

[0480] Genomics continuous data: copy number binning and quantification. DNA copy number data for each sample was obtained from the 2013Q4 Oncomine standard data construction, in which all copy number data available from TCGA up to September 2013 and all copy number data from manually curated publications were normalized.

[0481] The patient blank group for this analysis is the TCGA patient group measured for copy number data up to October 2013 and the patient group measured in the manually managed publication. Each patient retains only the data of one tumor sample (preferably, primary, non-recurrent tumor). The gene blank group of this data is the Oncomine DNA copy number platform, based on the RefSeq coordinates (hg18) provided by UCSC RefGene in July 2009, and 18,796 genes are measured. The measured values ​​of multiple reporter genes of a single gene are averaged.

[0482] Copy number data was segmented and quantified using the standard Oncomine processing pipeline. Segmentation is a method used to identify contiguous regions of amplification or deletion. These regions or "segments" can include multiple genes or a single gene. The copy number value for each segment is calculated based on the average of the reporter genes contained in the segment. Genes are mapped to segments and assigned a value. This gene-level data is then reported. For more information, please see the Oncomine DNA Processing Pipeline White Paper.

[0483] mRNA expression data. Expression data were obtained from the TCGA standard data build from Broad GDAC September 2013.

[0484] The patient null set for this data is the set of patients with available RNASeq data in the Broad GDAC 2013_08_09stddata build; data for only one tumor sample (preferably, a primary, non-recurrent tumor) is retained for each patient. The gene null set for this data varies by disease and corresponds to the TCGA gene annotation file (GeneAnnotation File, GAF) used for RNASeq quantification.

[0485] TCGA currently uses two RNASeq quantitative methods, V1 (RPKM) and V2 (RSEM), which are not directly comparable in value. In order to avoid possible inaccurate numerical conversions, we use data from a single quantitative method on a disease-by-disease basis, selecting the format based on maximum coverage. In line with TCGA's efforts to process (and reprocess) all available RNASeq data using RSEM (V2), RSEM (V2) data can be used for most samples. The exception is gastric adenocarcinoma, where RPKM (V1) data is used. Normalized gene-level quantitative values ​​for both RSEM and RPKM are obtained and converted to log2 values ​​(the minimum non-zero RPKM or RSEM value is set at -12). If the log2 value of a gene is >-12, then the gene is considered to be expressed.

[0486] Oncomine driver gene reference data: Minimum common region (MCR) peak generation and gene selection. To identify cancer driver genes that undergo amplification and deletion, a peak clustering method is performed to select genes that are frequently abnormal in multiple cancer types. The first copy number peak in the largest available copy number data set (i.e., data beyond it is included in the integrated analysis) within many cancer types is defined. Next, the gene list defined by the peak is clustered to identify genes that appear in the copy number peak in multiple samples and multiple diseases. Some of the methods are described in more detail below.

[0487] Aberrations can be classified as "driver" aberrations, or considered as potentially interesting aberrations according to one of the Oncomine classification methods specific to the data type. Driver aberrations will be captured as events independent of other aberrations (non-driver aberrations are referred to as "any"). For example, a patient with a "driver" mutation will be positive for two aberrations (a "driver" mutation and an "any" mutation). Each of the measured data types has a set of rules to determine driver events.

[0488] A set of contiguous genomic regions undergoing amplification or deletion were identified using the Oncomine MCR assay by applying it to 10,249 clinical samples from Oncomine grouped into 25 cancers.

[0489] The peak defined patient null set was 10,249 clinical samples from Oncomine (see table below). The gene null set for this data was the Oncomine DNA copy number platform, based on RefSeq coordinates provided by the UCSC refGene build July 2009 (hg18), and measured 18,796 genes.

[0490] The data for the minimum common region (MCR) analysis comes from the Oncomine DNA copy number browser containing >20,000 clinical samples, xenografts and cell lines across different cancer types. MCR analysis identifies regions of recurrent copy number amplification or deletion by analyzing data at three levels: pan-cancer (in all cancer types), common cancer types (in multiple cancer types) and intermediate cancer types or specific cancer subtypes. In brief, the method first calculates the common regions (CRs) defined as continuous genomic regions that are amplified or deleted in 2 or more samples. The minimum thresholds for amplification and deletion are set to log2≥0.9 (3.7 copies or more) and log2≤-0.9 (1 copy or less), respectively. Next, the peak regions within these common regions are defined as (i) one or more genes that are abnormal in the highest number of samples (n) and one or more genes that are abnormal in samples (n-1) that are one less than the highest number and (ii) genes that are abnormal in 90% of the highest abnormal sample counts.

[0491] Cluster analysis of common peak regions was identified. MCR analysis was performed. The peak regions identified by MCR analysis were further filtered in three analysis types (i.e., pan-cancer, general cancer type, and specific cancer type analysis) using the criteria listed in the table below. It should be noted that only a selected number (about 40) of mid-stage or specific cancer types (also listed further below) were included.

[0492] Filtering criteria for identifying highly amplified / deleted regions from MCR analysis:

[0493]

[0494] Selected intermediate or specific cancer types included in the filtering criteria described above:

[0495]

[0496] Next, to identify the most recurring peak regions and genes across multiple cancer types, we used Cytoscape 2.8.3 [Markiel et al. 2003; Smoot et al. 2001] to construct network clusters. Briefly, the analysis compares each gene in a given peak region with genes in other peak regions and clusters peaks with at least one common gene. Next, the most recurring amplified or deleted genes within each cluster were considered potential candidate driver genes. The method is outlined in Fig.11 middle.

[0497] Identification and development of clinically relevant cancer subtypes. To provide patient subsets for more focused analysis, several clinically relevant cancer subtypes were identified and curated using a combination of clinical phenotype and categorical genomic data. Phenotype data were derived from the TCGA Web Portal or supplemental methods from manually curated publications.

[0498] The following rules were applied to define clinical subsets:

[0499]

[0500]

[0501] Pathways. Manually curated Compendia pathway definitions are used to summarize gene-level aberrations in the integrated analysis browser. The pathways represent clinically relevant pathway modules, and several modules can cover major biological pathways. A single gene can be present in one or more pathway definitions, but care should be taken to eliminate substantially redundant pathways where one module is a complete subset of another. There are a total of 67 pathways, ranging in size from 42 genes (e.g., MAPK pathway) to 2 genes (e.g., IGF1 / IGF1R and several other pathways).

[0502] Data Integration. This section outlines the data flow that integrates raw data onto common patient and gene dimensions and constructs a genetic event database (GEDB) containing all aberrations that will be subjected to integrated analysis. The method has 4 main steps: (1) integrating raw data using common gene and patient dimensions; (2) calling aberration events for each data type and defining driver gene aberrations; (3) rolling up individual events to gene and pathway levels and integrating events; and (4) constructing a genetic event database by defining the patient status of each event.

[0503] Constructed and mapped to gene and patient dimensions. The varying data types included in the meta-analysis may have been measured on different experimental platforms and for non-completely overlapping patient groups. Therefore, care was taken to include all patients and genes measured while avoiding duplicate or conflicting entries.

[0504] For each data type, gene and patient "dimensions" are constructed, enumerating the genes and patients measured in the data. The dimensions for each data type may be different, as indicated by the following patient dimension overlay diagram (numbers are for illustration only), in this case, for invasive breast cancer.

[0505] Gene and patient dimensions were collected from each cancer and data type, and a non-redundant superset of all patient and gene identifiers was constructed in the integrated analysis. The bars in the following figures represent data blocks of patient identifiers (if sorted by that identifier) ​​for patients measured for certain aberration types.

[0506] Redundancy for a patient is determined based on a unique patient identifier, currently the first 12 characters of the TCGA tumor sample barcode (eg, TCGA-AB-1234).

[0507] The redundancy of genes is determined based on unique gene identifiers (currently Entrez Gene IDs). Genes are also compared relative to the Oncomine genome, and when gene symbol conflicts are found (one Entrez ID is assigned two or more gene symbols), the gene symbol from Oncomine is used. Several (12) measured genes contain discontinued Entrez IDs and therefore may not represent the most accurate gene model. The gene symbols of these genes are marked with the word "discontinued".

[0508] Once constructed, non-redundant gene and patient dimensions are indexed to provide consistent internal identifiers for each gene and patient in the data set. Then, all unique patients and gene identifiers in the raw data are mapped to the dimensional patient and gene identifiers. Therefore, gene and patient metadata (such as gene symbols and patient clinical data) are always mapped via corresponding dimensions, providing consistency in naming and annotation. The total number of unique genes and patients in the integrated analysis is as follows:

[0509]

[0510] The patient dimension together with the dataset-specific mapping of patients helps to correctly identify completely wild-type patients (patients who have all aberration types measured but do not contain any aberrations).

[0511] A patient may thus measure any number of distortions, and may be abnormal only for those events measured. To determine the type of analysis measuring a patient distortion, the patient will be included in the following:

[0512]

[0513] * "For clinical measures" indicates that clinical metadata is presented for the patient.

[0514] Event model. Each genomic aberration from mutation, fusion, and copy number data sets is identified as an aberration event (a term used to define related events that will undergo various predefined integrative analyses). Each aberration is part of three broad levels of events (data type-specific events, gene-specific but data type-independent events, and pathway-specific but gene- or data type-independent events). The latter two levels are considered "rollup" events.

[0515] The following describes the unique rules for defining aberration events as well as the "layer" and "driver gene" processes.

[0516] Mutation event calling program. Oncomine Pan-Cancer Mutation Classification: A mutation is classified as "hotspot" if it is: recurrent (occurring at the same amino acid position in 3 or more samples) and annotated with one of the following variant classifications: in-frame insertion / deletion, non-stop missense. A mutation is classified as "deleterious" if it is: non-recurrent and annotated with one of the following variant classifications: frameshift insertion / deletion, nonsense. Reproducibility is measured across all cancer types analyzed as part of the Oncomine NGS mutation browser.

[0517] Oncomine Pan-Cancer Gene Classification. As part of the Oncomine NGS Mutation Browser pipeline, individual genes are classified into predicted functional categories, "Gain of Function" and "Loss of Function", to reflect their relative enrichment in potentially activating or deleterious mutations. This classification is based on a combination of the relative frequency and significance of mutations observed in the gene as assessed by p-value. "Gain of Function" genes will have a relatively high frequency of hotspots (reproduced in 3 or more samples) missense mutations and a lower frequency of deleterious mutations, while "Loss of Function" genes contain a larger portion of deleterious mutations. The Pan-Cancer Gene Classification is based on mutations observed in all cancer types.

[0518] Mutation aberration events. For each patient gene mutation (as defined by the mutation data filtering section), one or two mutation events will be formed, depending on whether the mutation is classified as a driver gene aberration. Driver gene mutation aberrations are defined as "hotspot" mutations present in "gain of function" genes, or "hotspots" or "harmful" mutations present in "loss of function" genes. For driver gene mutations, both driver gene events and any events are formed. For non-driver gene mutations, only one any event is formed. Pan-cancer mutations and gene classifications are used for all analysis subsets; therefore, gene classifications may sometimes differ between integrated analysis and Oncomine NGS mutation browsers.

[0519] The following table gives descriptions and examples of mutational events that can be generated for each gene:

[0520]

[0521] Fusion event calling program.Only Oncomine priority fusions are included in the integrated analysis.In priority fusions, driver gene fusions are defined as fusions of known oncogenes marked by Mitelman database or fusions without sufficient exon expression data and reproduced or fusions with exon expression data and exon expression imbalance of two gene collocations (for details of exon imbalance classification, see Oncomine NGS fusion method data).For each gene, an event is formed by combining 5'-3' observed for each unique gene collocation.For example, for PML-RARA balanced translocation, two isoforms are observed, and therefore two fusion events will be called for PML-RARA and RARA-PML respectively.

[0522]

[0523] Copy number event calling procedure. Each amplification / deletion called (see Calling Amplification / Deletion) is defined as any event of an abnormal gene. If the amplification / deletion occurs in a gene that is part of the peak definition (see MCR peak generation) in a certain cancer type, then a driver gene amplification deletion event for the gene is also formed. Therefore, the driver gene definition of the copy number event is specific to the cancer type.

[0524] The following are copy number aberration events that can be called for gene amplification or deletion:

[0525]

[0526] Gene event rollup. Driver genes and any events are "rolled up" to gene-level and pathway-level events to capture data type-independent aberration statistics and correlations. For example, it may be interesting to see the correlation of any aberration in a tumor suppressor gene with clinical outcome, not just the correlation of deleterious mutations.

[0527] Gene level distortion events are formed for each gene with at least one distortion of any data type. Path level distortion events are formed for each path in which at least one of the component genes has distortion of any data type. Driver genes and any distortions are independently rolled up into gene level or path level driver genes or any events. The following diagram shows the hierarchical relationship between various distortion event types.

[0528] Patient Event Status. A patient may measure any number of aberrations, but may be abnormal only for those events that were measured. Thus, the patient status for each event-level aberration was recorded as abnormal, wild-type, or not measured.

[0529] The patient dimension together with the dataset-specific mapping of patients helps to correctly identify completely wild-type patients (patients who have all aberration types measured but do not contain any aberrations).

[0530] When gene-level and path-level events are defined, only patients who have measured all 3 gene data types (mutation, fusion, and copy number) are marked as "abnormal" or "wild type" for the event. This assumption has the effect of reducing the number of patients summarized for gene or path-level aberrations compared to event-level aberrations that are specific to the data type. If at least one of the patient's event-level aberration types (fusion, mutation, amplification, or deletion) is abnormal, the patient is considered to have a gene-level event abnormality. If the patient has an aberration in at least one gene that is part of the path definition, the patient is considered to have a path-level event abnormality. In each case, the patient must have measured all event types.

[0531] In the case of gain of function and fusion, the aberration frequency is about 50%. For driver gene aberration events, the aberration frequency is also about 50%, but only half of the patients are included in the numerator and denominator of the frequency.

[0532] Analysis. Once all driver gene events were constructed, a set of analyses were performed on each gene event to calculate frequencies, correlations, and relationships within cancer types, clinically relevant subtypes, and between cancer types (pan-cancer). Below is a brief description of each analysis, including what data was used and what restrictions (if any) were placed on the reported results: frequency, frequency of expression, co-occurrence and mutual exclusivity, clinical relevance analysis, etc.

[0533] Frequency. Frequency is the rate of occurrence of an aberration among the patients for whom it was measured. Frequencies were calculated within cancer types, clinically relevant cancer subtypes, and pan-cancer. All events with at least one patient with an abnormality were reported.

[0534] Frequency of expression. The frequency of expression is the frequency of expression of a gene among event-positive patients. For each event, the expression level of a gene is used to determine the frequency of expression. Frequencies are calculated within cancer types and clinically relevant cancer subtypes but not pan-cancer.

[0535] Co-occurrence and mutual exclusivity. The co-occurrence and mutual exclusivity of each pair of events were calculated using Fisher's exact test. At least 2 patients positive for each event and 5 patients with all events measured were required for calculation. The co-occurrence or exclusivity of two individual copy number deletion or amplification events was not calculated. In addition, the co-occurrence and mutual exclusivity between pairs of events with "any" driver gene status were not calculated (i.e., comparing only driver gene vs. driver gene and driver gene vs. any). Correlations were calculated within cancer types and clinically relevant cancer subtypes but not pan-cancer.

[0536] Clinical relevance analysis. Test the relevance of each driver gene event relative to a set of available clinical subtypes. By comparing the incidence of gene events in patients with one clinical subtype versus another clinical subtype, each correlation is tested using Fisher's exact test. For example, overexpression of loss-of-function mutations can be tested in smokers versus non-smokers or in phase I versus phase II lung cancer. A total of 136 subtype pairs were tested for each event, and the characteristics defining the subtypes (some characteristics may be disease-specific) are listed below. A total of at least 4 patients (and at least 1 patient in each category) are required to perform the test. Calculate the correlation within cancer types, clinically relevant cancer subtypes, and pan-cancer.

[0537] Clinical subtype characteristic name:

[0538] Race / Ethnicity

[0539] Modified smoking status

[0540] ERBB2 status

[0541] Estrogen receptor status

[0542] Progesterone receptor status

[0543] TCGA PAM50 subtypes

[0544] BRAF mutation status

[0545] Modified T period

[0546] Modified N Period

[0547] Modified M phase

[0548] Modified staging

[0549] KRAS mutation status

[0550] EGFR amplification status

[0551] TCGA subtype

[0552] Microsatellite status

[0553] Human papillomavirus infection status

[0554] Clinical outcome analysis. The correlation of each event with clinical outcome was tested using the log-rank test. Only patient groups with available clinical data were used for calculation, so the number of patients included in the test may be less than the number of patients measured for driver gene events. At least 4 patients with abnormal events were required to perform the test. Survival time was presented in years, and individual survival / death events were clearly marked on the Kaplan-Meier curve. Correlations were calculated within cancer types and clinically relevant cancer subtypes but not pan-cancer.

[0555] DNA-RNA correlation analysis. For each gene, RNA expression and DNA copy number values ​​were tested for correlation among all patients with these data types of disease measured using Pearson's correlation. Correlations were calculated within cancer types and clinically relevant cancer subtypes but not pan-cancer.

[0556] Differential Expression Analysis. For each event, each gene associated with the event was tested for differential expression in event-positive patients versus event-negative patients using Student's T-Test. For events involving several genes, such as fusions, each gene was tested. Differential expression was calculated within cancer types and clinically relevant cancer subtypes but not pan-cancer.

[0557]

[0558]

[0559]

[0560]

[0561]

[0562]

[0563]

[0564]

[0565]

[0566]

[0567]

[0568]

[0569]

[0570] Example 5: Other fusion methods

[0571] Clinical data sources. All RNASeq data used for gene fusion analysis were obtained from The Cancer Genomics Hub (CGHub, a repository for current TCGA genomic data - https: / / cghub.ucsc.edu / ) is obtained.

[0572] Cell line data sources. All CCLE RNASeq data used for gene fusion analysis were obtained from The Cancer Genomics Hub (CGHub, a repository for current CCLE NGS data - https: / / cghub.ucsc.edu / ) is obtained.

[0573] BAM to FASTQ conversion. The input into the fusion caller consisted of RNASeq reads in FASTQ format, which required converting the BAM files provided by TCGA into one or two FASTQ files for single-end or paired-end data (respectively).

[0574] BAM files vary in provenance and processing, and many require special handling. For example, older BAM files provided by UNC are aligned using BWA (Burrows-Wheeler Aligner), while newer BAMs containing reads are aligned via MapSplice. TCGA recently updated the RNASeq pipeline to support alternative gene expression reporting. (The previous pipeline relied on RPKM measurements of gene expression, while the latter uses RSEM.) These different RNASeq analysis pipelines are referred to by UNC as V1 and V2 ( https: / / wiki.nci.nih.gov / display / TCGA / RNASeq+ Version+2 ). When both formats were available for the same TCGA sample, we used the following BAM prioritization pipeline to select a single “original BAM”: 1) select the V2 BAM relative to the V1 BAM; and 2) when multiple files of the same situation were presented, select the BAM with the more recent upload date.

[0575] The custom SamToFastq converter described above was used to generate FASTQ files from TCGA BAM files.

[0576] There are 2 cancer types (COADREAD and UCEC) that are only available as single-end RNASeq data. For single-end BAM file conversion, the program BamTools (Secure Hypertext Transfer Protocol :: / / github.com / pezmaster31 / bamtools ) to generate FASTQ files.

[0577] With the goal of supporting both single-ended and dual-ended data, we use TopHat to process all single-ended data and deFuse to process all dual-ended data.

[0578] Broadly, our analysis pipeline consists of 5 main steps:

[0579] Preprocessing of raw data to obtain FASTQ files

[0580] Run the fusion caller

[0581] Filter breakpoints to relevant gene regions

[0582] Annotation of breakpoints with Oncomine transcriptome

[0583] Overview and prioritization of novel fusions of potential interest

[0584] Steps 1 and 2 were performed in parallel for all samples based on a high-performance cloud computing cluster. The aggregated data were filtered and annotated as a post-processing step to enable exploratory analysis of the effects of various filtering and annotation processes. After completing the filtering criteria to minimize false positive fusions (step 5), the list of Oncomine prioritized fusions was validated with RNASeq exon expression data.

[0585] TopHat.TopHat-Fusion from author HTTP : / / tophat.cbcb.umd.edu Software and reference data dependencies are configured as specified by the TopHat documentation:

[0586] software:

[0587] TopHat: 2.0.4, including TopHat-Fusion Post (04 / 09 / 2012 issued )

[0588] bowtie: 0.12.8 (05 / 06 / 2012 issued )

[0589] samtools: 0.1.18 (09 / 02 / 2011 issued )

[0590] blast(2.2.26)(03 / 03 / 2012 issued )

[0591] blast+(2.2.26)(10 / 21 / 11 issued )

[0592] References and Notes:

[0593] Reference genome: UCSC hg19 (Downloaded in May 2012)

[0594] Gene model: refGene , ensGene (Downloaded in May 2012)

[0595] BLAST DB: nt , Human , other (Downloaded in May 2012)

[0596] parameter:

[0597] We ran TopHat on single-end and paired-end TCGA Illumina data using essentially default parameters as specified in the TopHat documentation. Below is a list of the parameters used.

[0598] Table 25

[0599]

[0600]

[0601] The --mate-inner-dist and --mate-std-dev parameters do not have default values. The first parameter specifies the expected insert size of RNASeq paired-end reads, while the second parameter specifies the expected standard deviation of the values. The TopHat authors recommend values ​​of 0 and 80 for most datasets.

[0602] Table 26

[0603]

[0604] TopHat-Fusion was performed simultaneously on one sample, followed by TopHat-Fusion Post. We retained both the unfiltered TopHat-Fusion output and the filtered TopHat-Fusion Post output to enable deeper analysis.

[0605] deFuse.deFuse from the author: Hypertext Transfer Protocol : / / defuse.sf.net Get. Software and reference data dependencies are configured as specified by the deFuse documentation:

[0606] software:

[0607] deFuse: 0.5.0 (04 / 07 / 2012 issued )

[0608] bowtie: 0.12.8 (05 / 06 / 2012 issued )

[0609] R 2.15.0 (03 / 30 / 2012 issued )

[0610] blat , faToTwoBit (Obtained on 05 / 01 / 2012)

[0611] References and Notes:

[0612] Reference genome: Ensembl GRCh37.62 fa (Downloaded in May 2012)

[0613] Gene model: Ensembl gtf (Downloaded in May 2012)

[0614] Genomic data:

[0615] UCSC EST fasta, EST Compare and Repeating sequence (Downloaded in May 2012)

[0616] NCBI UniGene (Downloaded in May 2012)

[0617] parameter:

[0618] We ran deFuse with default parameters as specified in the deFuse program documentation.

[0619] Table 27

[0620] deFuse Parameters The values ​​used -bowtie_quals phred33-quals -max_insert_size 500 -discord_read_trim 50 -clustering_precision 0.95 -span_count_threshold 5 -split_count_threshold 3 -percent_identity_threshold 0.90 -max_dist_pos 600 -num_dist_genes 500 -split_min_anchor 4 -max_concordant_ratio 0.1 -splice_bias 10 -denovo_assembly No -probability_threshold 0.5 -covariance_sampling_density 0.01 -reads_per_job 1,000,000 -regions_per_job 20 -p 4

[0621] Execute deFuse on one sample at a time. We keep both the filtered and unfiltered results of deFuse output to enable deeper analysis.

[0622] Integration. We integrated the "Level I" data, the output of the potential_fusion.txt file from TopHat-Fusion Post and the output of the results.classify.tsv file from deFuse. deFuse reports many more potential calls at this level than TopHat, and therefore may also report more false positive predictions. Level I data was chosen to strike a balance between taking advantage of the built-in filtering of the calling program and allowing enough results to be run through to identify possible true fusions with slightly weaker evidence.

[0623] Since each caller provides different levels of annotation and supporting evidence for fusion calls, the breakpoints of the predicted fusions from both callers were extracted and integrated into a common format for filtering and annotation. The integration step consisted of converting the reported breakpoints into genomic coordinate-based breakpoints and merging into a common file format.

[0624] Breakpoint filtering. The predicted fusions from the "Level I" output of the caller were filtered, retaining only those calls where each breakpoint was in the 5'UTR or CDS region of a RefSeq transcript (refGene obtained from UCSC on ca. 18 Jul. 2012). This was done to enrich the predicted fusions for those containing functional gene regions, filtering out fusion calls where, for example, the 3'UTR of one gene was predicted to be fused to the 3'UTR of another gene. Although at the genomic DNA level, breakpoints may occur in introns, in RNASeq data, they will be observed at the closest exon-intron boundary. Therefore, breakpoints predicted to occur in intronic sequences were also excluded.

[0625] Breakpoint Annotation. After excluding fusions outside the 5'UTR or CDS regions of RefSeq transcripts, annotations from RefSeq transcripts were transferred to some of the predicted remaining breakpoints with annotations relative to multiple Entrez IDs.

[0626] For each pair of breakpoints, only one transcript was retained per Entrez ID. In the case of multiple transcripts, the transcript with the shortest transcript accession number was selected; further ties were broken by sorting the accession numbers alphanumerically and retaining the first accession number. This process ensured consistency in annotated breakpoints at the same position. However, the predicted breakpoints at different positions of the same gene pair can still produce multiple transcripts representing a pair of genes, which is evidence of the possibility of alternative transcripts.

[0627] The basic annotation from the caller itself is discarded, as it is the default annotation source based on each corresponding caller. However, certain output fields from both TopHat and deFuse are retained to help prioritize predicted fusions. Additionally, certain annotation features not explicitly reported by the caller are inferred from other caller features.

[0628] Inferred properties. Supporting and spanning read counts were obtained from each caller and summarized in the columns Read span and Read span support. The latter column is the sum of reads spanning the fusion and reads supporting the fusion (not to be confused with TopHat's counts for "spanning mate pairs where one end spans the fusion", which are sometimes referred to as 'spanning and supporting reads').

[0629] The breakpoint sequence reported by the calling program was trimmed to include 50 bases on each side of the fusion and merged into the breakpoint sequence column. The fusion breakpoint is depicted by "|". It should be noted that this is the breakpoint sequence as inferred by the calling program and is not simply obtained from the reference genome. Because the inferred sequence may reflect the actual sequence observed by the spanning reads, this sequence may represent the complement of the reference genome sequence.

[0630] Since neither of the callers provided a definitive '5'' or '3'' label, we inferred the relative 5'-3' orientation of the fusion partners by combining the caller parameters with the gene strand annotation. For deFuse, the orientation of each partner was inferred based on the following combination of gene strand and the deFuse output property 'genome_strand':

[0631] Table 28

[0632]

[0633] TopHat reports a different metric, namely, the relative orientation of reads mapped to gene pairs, and thus requires a different set of rules to infer the 5'-3' order of gene pairs:

[0634] Table 29

[0635]

[0636] If there are putative 5' and 3' partners for a given gene fusion call, then the valid targeting field is marked as "Y".

[0637] RepeatMasker annotation. Each predicted breakpoint position was also annotated with RepeatMasker features near the breakpoint. This was done to identify breakpoints in highly repetitive genomic regions where alignment errors are likely to affect the prediction of chimeric transcripts.

[0638] Specifically, the 25 bp sequences upstream or downstream of the 5' and 3' partner breakpoints were selected as 'breakpoint flanking sequences', respectively. These flanking sequences were relative to the RepeatMasker element set (Hypertext Transfer Protocol 504) downloaded from the UCSC table browser on August 24, 2012. : / / www.repeatmasker.org / ) intersect. We report the element name, element length, and amount of overlap with the 26-base breakpoint flanking region for each breakpoint. Currently, RepeatMasker elements are not filtered for specific element types (LINES, SINES, simple repeats, etc.).

[0639] For each fusion prediction, we set the RepeatMasker overlap field to equal the number of bases that the breakpoint flanking sequence overlaps with the RepeatMasker element, and considered overlaps of 12 or more bases to be significant. The frequency of significant overlapping fusion calls was used for Oncomine prioritization described below, so that gene fusions with lower overlap frequencies were considered higher quality.

[0640] Fusion exon expression imbalance. Fusions were observed using RNASeq exon expression data to provide secondary evidence for true positive fusion events by looking for imbalance in exon expression before and after the breakpoint call. Specifically, if expression of the 3' partner is affected by the promoter region of the 5' partner, then exon expression should increase after the predicted breakpoint. This effect is particularly visible when looking at fusion vs. non-fusion patient samples.

[0641] TCGA exon expression data. TCGA exon expression data were downloaded from the Broad GDAC Firehose location. The RPKM RNASeq values ​​for each patient were listed as Gene Annotation Format (GAF) features corresponding to a composite of UCSC exons from several different gene definitions (including RefSeq). After downloading data for 21 diseases, we found 4 different sets of GAF features for annotating RPKM expression. Finally, the availability of patient expression data varied by disease in the V1 and V2 RNASeq analysis pipelines described above.

[0642] To address these challenges, we first mapped UCSC RefSeq exons to available GAF features and calculated the percent overlap between each RefSeq exon and the GAF feature. This step is critical because all CBI-processed fusion breakpoints are mapped to the UCSC Refgene definition downloaded on 7 / 18 / 2012, and these breakpoints must in turn be mapped to GAF features. 80.8% of the 396,298 RefSeq exons map perfectly to the GAF features in the curves shown below. We select and report the RefSeq exon and GAF ​​feature pairs that yield the greatest overlap.

[0643] The value called rg_pct provides a measure of the quality of mapping of a given RefSeq exon to a GAF feature based on the following formula:

[0644] rg_pct = overlap / length refseq *Overlap / Length GAF特征

[0645] Mappings with an rg_pct value of 1 overlap perfectly, while values ​​less than 1 indicate that the RefSeq exons or GAF features do not map to the exact same genomic region and the RPKM value may be questionable.

[0646] We selected RNASeq V2 data for all diseases except STAD (due to unavailability of V2 data).

[0647] Cell line exonic expression data. Exonic expression data for cell line samples were generated from CCLEBAM files obtained from CGHub. The approach used was similar to step 18, as described in the “TCGA mRNA-seq pipeline for UNC data” methods available here: http: / / www.cghub.org / : / / webshare.bioinf.unc.edu / public / mRNAseq_ TCGA / UNC_mRNAseq_summary.pdf .

[0648] The difference between the UNC method and ours is that the RefSeq exon BED is used in our method instead of the composite exon BED used by TCGA.

[0649] Exonic expression imbalance calculation. Each sample was systematically analyzed for evidence of potential 5' promoter-induced imbalance in 3' partner expression. First, the expression level of each gene was converted to a logarithmic scale, and then the z-scores were normalized across the sample set for each disease. This normalization was performed at the exon level to account for broad population trends such as 3' bias or poor RefSeq exon / GAF signature matching (see below).

[0650] Crude RPKM expression values ​​of PLXNB21 and COL7A1 in serous ovarian cancer patients (top) versus z-score normalized values ​​(see Fig. 8A -D). Extensive population drops in PLXNB1 expression at exons 12, 17, and 23 are smoothed in the normalized data. Samples predicted to have fusions between these genes are highlighted in red; wild-type patients are shown in blue. Red diamonds indicate breakpoint exons predicted by the caller.

[0651] Prior to normalization, samples that were considered wild-type for the fusion in question but were predicted to have other fusions involving one of the gene partners were removed from the wild-type population so as not to contaminate the z-score calculation.

[0652] After normalization, each sample was assigned a p-value calculated via a one-sided Student's t-test (H0: μ A ≤μ B). The breakpoints predicted by the calling programs were used to separate the expression populations of samples identified by either fusion calling program.

[0653] The p-value for each wild-type sample was also calculated to facilitate analysis of the p-values ​​for fusion-positive samples in the context of the entire population. This allowed us to discard fusions involving genes that exhibited a broad population exon imbalance trend that was not induced by the fusion. Any sample that did not rank within the top portion of the wild-type sample p-value was discarded. The breakpoint that maximized the difference between expression levels before and after the breakpoint was used for the wild-type sample p-value calculation.

[0654] Fusion Summary. Fusions within the disease are summarized based on the occurrence of unique gene pairs as well as based on the occurrence of individual genes (possibly with multiple partners).

[0655] For unique fusion pairs (unique according to Entrez ID pairs), the number of samples within a disease that have at least one fusion prediction according to any caller is the sample count of the fusion. Since multiple breakpoints for the same gene pair can be reported in one sample and across multiple samples, the number of unique fusion pairs within each disease is much smaller than the total number of fusion calls. In order to filter and prioritize fusions at the gene pair level rather than the fusion call level, several fusion caller properties are outlined. The following table shows the properties outlined for a given fusion partner pair across individual predictions:

[0656] Table 30

[0657] characteristic Overview Method DEFUSE Flip Total fusion calls % = 'Y' DEFUSE effective orientation Total fusion calls % = 'Y' DEFUSE multi-mapping count Total fusion calls %>0 TOPHAT effective orientation Total fusion calls % = 'Y' 3P / 5P_REPEATMASKER overlap Total fusion calls % ≥ 12

[0658] If genes were <1 Mb apart on the genome, a proximity flag was set for the fusion, and a defuse flip flag was set in individual fusion predictions for ≤75% of these fusion partners.

[0659] Gene-level overview. Sample counts for fusions are also summarized at the gene level (uniquely by Entrez gene ID) within each disease type and across diseases (pan-cancer). This summary method is independent of the orientation inferred within the fusion. In addition, sample counts are only calculated for fusions of Oncomine priority fusions (described below).

[0660] Individual unique fusions were cross-referenced to the Mitelman database of genomic aberrations (HT : / / cgap.nci.nih.gov / Chromosomes / Mitelman ). Matching is based on gene name rather than disease type. Therefore, a gene fusion reported in Mitelman for one disease may have appeared in a different disease type in the TCGA dataset.

[0661] Gene fusions summarized at the gene level were cross-referenced to the Mitelman database based on gene name. Thus, genes reported in Mitelman as having different histology or completely different aberration types (e.g., larger chromosome level deletions rather than fusions) than predicted unique fusion pairs have more potential.

[0662] Normal sample fusion blacklist. Assuming that all fusions called in TCGA normal samples are false positives, we ask the following questions: 1) Are fusion calls in tumor samples identified in normal samples? 2) Are Oncomine prioritized fusions identified in tumor samples also identified in normal samples? Answering the first question provides a baseline sense of the technical false positive rate in tumor gene fusion calls. The second question is a sanity check on how well Oncomine priority fusion filtering overcomes this problem. 344 double-end normal samples within 10 diseases were downloaded and processed using the same deFuse pipeline described above. A total of 56,579 total fusion calls consisting of 6,024 unique fusions were observed. These normal sample fusion calls are used to generate a blacklist, and these false positives are removed from the Oncomine priority gene fusions.

[0663] Paralogous fusion partners blacklist. The blacklist of fusions between members of the paralogous gene family is compiled using two strategies: 1) manually detecting high frequency fusion partners gene names and 2) comparing the first 3 characters of all priority fusion partners gene names. In the latter strategy, the fusion partners are verified as "paralogous" using HomoloGene, Ensembl, SIMAP and GeneDecks V3, which are then included in the final blacklist. The following table shows the first 10 most frequently observed gene fusion calls between paralogous fusion partners. The entire table is composed of more than 400 unique paralogous gene fusions and is used to remove these false positives from our Oncomine priority gene fusions.

[0664] Table 31

[0665]

[0666]

[0667] Fusion Prioritization - Oncomine Prioritization Pipeline. The Oncomine Prioritization Pipeline outlined below is designed by iteratively exploring the top results in Level I fusion predictions and systematically eliminating suspected false-positive fusions while retaining previously discovered 'true positive' (Mitelman) fusions. This pipeline is intended to highlight fusions that meet certain characteristics expected of 'true positive' fusions and conversely lack characteristics observed in many 'false positive' fusions.

[0668] A fusion is an Oncomine priority fusion if:

[0669] Table 32

[0670]

[0671] Example 6: Oncomine NGS mutation method

[0672] Mutation integration. The goal of data integration is to form the most complete NGS mutation data set currently available. We consider the following sources:

[0673] Original data source

[0674] COSMIC Cell Line Project

[0675] TCGA data from Bode GDAC Mutation_Packager (stddata build)

[0676] TCGA data from DCC level 2

[0677] Compendia mutation calling based on TCGA data

[0678] Publications containing NGS mutation data

[0679] COSMIC Cell Line Project

[0680] The Cancer Genome Project has characterized the exomes of more than 1,000 cancer cell lines for mutations. The database provides mutation data filtered for quality in a flat file format. Cell line data undergo the same Oncomine curation and annotation methods used for clinical mutation data. Cell line names are reviewed against the Oncomine ontology, and cancer types are standardized to be comparable to clinical mutation data.

[0681] The dataset was obtained from the Wellcome Trust Sanger Institute Cell Lines Project website when it was presented in November 2013: http: / / www.htpl.org / : / / cancer.sanger.ac.uk / cancergenome / projects / cell_lines / get.

[0682] Bode GDAC Mutation_Packager. Bode has been working on collecting and integrating mutation data from multiple sources since the third quarter of 2011.

[0683] https: / / docs.google.com / document / d / 18XlWv- a9xLBOflNikOa9rCXOyiravMM8_-PVJxAQ PPo / edit

[0684] The above document details the source of the MAF file that Bode is integrated into the Mutation_Packager standard data run. Bode has integrated many MAF files, and the MAF files are maintained outside the central TCGA DCC system, usually maintained by the members of the analysis working group itself. We have performed an in-depth comparison between all MAF files available to us. We believe that Bode has the most complete available mutation data.

[0685] For this release, we integrate data from the 2013_08_09stddata build.

[0686] TCGA DCC Level 2. This is controlled access mutation data available from the DCC. TCGA has a page on its wiki that provides additional details about the available MAF files:

[0687] https: / / wiki.nci.nih.gov / display / TCGA / TCGA+MAF+Files

[0688] For this version we took into account all MAF files available until September 15, 2013.

[0689] KenPendia NGSDNASeq mutation calls. We felt that the PRAD mutation calls available from TCGA were of low quality and produced false positive 'gain of function' predictions. Therefore, all calls for this disease were derived from KenPendia's own mutation calling pipeline. KenPendia mutation calls were made to conform to the MAF file format for integration. For more details, see Appendix: Compendia NGS DNASeq mutation calling Mutation Calling . 170 patients with prostate adenocarcinoma were included in this version.

[0690] Manual curation of all NGS data.TCGA and non-TCGA NGS datasets were obtained directly from their original sources (primarily peer-reviewed cancer publications and the publicly accessible databases mentioned above) by the Oncomine curation team. The mutation data generally available in the supplementary materials were brought up to the standards required for mutation re-annotation and classification as part of the entire NGS mutation processing pipeline. Cancer types are managed using the Oncomine cancer type ontology, and appropriate Oncomine cancer types are assigned based on the best available clinical metadata present in the publications. Since all published experiments require whole genome ('NGS') coverage, the empty genome for each dataset is assumed to include all human RefSeq genes. Non-TCGA data are processed in exactly the same way as TCGA MAF file data used for the rest of the mutation analysis pipeline.

[0691] Removing duplicate mutations. We performed some simple cleanup operations to remove duplicate mutation records present in the source data. We also performed some file-column name remapping, since many sources do not adhere to the MAF file standard. Remove duplicate mutations from various tumor / normal aliquot pairs from the same patient sample.

[0692] Mutation Annotation. The data obtained from TCGA and non-TCGA sources contain mutations generated by data sets processed and annotated by different genome sequencing centers or authors over several years. This produces mutation calls annotated using different gene models and using different variant classification conventions. Since the method used by KenPendia to define mutations relies on accurate variant annotations, we re-annotate mutations relative to a single transcript group and consistent variant classification rules. The standard annotation pipeline ensures that mutations across disease types are consistently evaluated and undergo a common interpretation during the nomination of potential oncogenes or tumor suppressor genes. Important annotations that are not consistently available from the original source are also provided, such as HGVS style mutation nomenclature (e.g., V600E).

[0693] Mutations obtained from primary sources were processed by Compendia according to the following general procedures (details provided below).

[0694] We first re-annotated each mutation using the Oncomine transcriptome from Compendia. Successfully annotated mutations received annotations from Compendia, while the rest retained annotations obtained from the original source. Annotations included:

[0695] Variant classification

[0696] Variant location

[0697] Variant changes

[0698] Several filtering steps were performed to remove redundant annotations across multiple transcripts as well as mutations located outside of the relevant gene regions.

[0699] Excluding "hypermutation gene" samples. In some diseases (such as endometrial cancer), several highly mutated samples may dominate the overall mutation count. We also observed the "hypermutation gene" samples in lung adenocarcinoma, gastric cancer, melanoma, and colorectal cancer. Based on the cutoff value determined by analyzing the hypermutation gene outliers in several cancer types, we decided to use <5,000 non-silent exon mutations as the threshold for including samples in our reproducible analysis. We therefore exclude many hypermutation gene samples in this data set from our downstream analysis pipeline.

[0700] In the mutation annotation step, we attempted to re-annotate mutations obtained from the original sources relative to a standard set of transcripts compiled by Compendia, Inc. This set of transcripts includes RefGene transcripts from the hg18 and hg19 genome builds obtained from UCSC on February 19, 2012.

[0701] Each mutation is mapped individually to the contigs in the Oncomine transcriptome within the specified genome build. SNP mutations are mapped directly to their starting position, while the relevant position is selected for small insertion (INS) and deletion (DEL) mutations for mapping. For insertions, the relevant position is the base where the insertion occurs. Depending on the direction of the transcript, this can be the starting or ending coordinate of the mutation, depending on whether the gene is on the positive strand or on the negative strand, respectively. For deletions, the relevant position is the base of the deletion when the transcript is on the positive strand or the last base of the deletion when the transcript is on the negative strand. This adjustment ensures that the mutation position is defined as the first base affected by the insertion / deletion relative to the direction of transcript translation (i.e., 5'→3').

[0702] For mutations that were successfully mapped to transcripts, the Compendia mutation annotations were inferred relative to the transcripts. For mutations that failed to map, the annotations from the original data source were retained, and variant positions for hotspot calculations were constructed based on genomic coordinates (more details below). Since only the standard set of 23 chromosomes were included in our transcriptome, mutations located on mitochondrial or other non-standard contigs were not mapped.

[0703] Below is a description of the criteria used to annotate mutations mapped to the Oncomine transcriptome.

[0704] Variant classification. For each mutation successfully mapped to a transcript, a combination of mutation and annotation features was used to infer the variant classification. Our method identified six major mutation variant classifications, all located within transcripts. Variant classifications for mutations outside of gene regions (e.g., intergenic) are not currently considered (see filtering section below). The following are the guidelines for inferring variant classifications:

[0705] Table 33

[0706]

[0707] This list of variant classifications is a subset of the allowed variant classifications specified by TCGA for the MAF file format.

[0708] https: / / wiki.nci.nih.gov / display / TCGA / Mutation+Annotation+Format+% 28MAF%29+Specification

[0709] This subset covers relevant mutation classes for reproducible analysis and identification of potential gain- or loss-of-function genes, and is therefore sufficient for the vast majority of mutations mapped to the Oncomine transcriptome. The following table describes the possible variant classifications that would be assigned compared to the original author classifications (assuming the mutations map to the same transcripts as those used to define the classifications) and the relative abundance of said mutation types in the source datasets:

[0710] Table 34

[0711]

[0712] Variant position. One of the initial goals of the current analysis is to identify genes with hotspot mutations, which are mutations of a certain classification observed at the same position in multiple tumor samples. In order to effectively identify the hotspots that reproduce and define each mutation, we must construct mutation point identifiers that cover mutation positions, affected amino acids or base identities, and variant classifications. We aggregate mutations that occur at the same position and are unrelated to the specific base changes they produce. Therefore, we only use reference bases or amino acids to define variant positions. This ensures that mutations that affect the same codon or genomic position will be counted for possible hotspots, even if the alternative alleles they produce are different. For example, for a given gene, missense mutations V600E, V600F, and V600G will all have a variant position of V600, and will thus be clustered together when identifying hotspot mutations. Therefore, our variant positions are defined as follows:

[0713] Variant position = mutation point {base|codon} + reference {base|AA} + [variant classification]

[0714] If the mutation is in the coding region, the codon number and the corresponding amino acid at the relevant base are used to identify the mutation point, such as p.L116_in-frame deletion. If the mutation is in a non-coding region (such as UTR), the position and identity of the reference nucleotide at the relevant base are used to identify the mutation point, such as c.*110C.

[0715] For splice site mutations outside the coding region, the variant position is specified relative to the splice boundary. The relative position is identified using +{1|2} or -{1|2} (splice site mutations are those within 2 bases of the splice junction). As with insertions and deletions, a suffix of "_splice site" is added for splice site mutations. For insertion and deletion mutations, a suffix indicating in-frame ("_in-frame insertion" or "_in-frame deletion") or frame shift ("_frame shift insertion" or "_frame shift deletion") is added to the variant position.

[0716] In summary, the following are examples of different possible variant position formats:

[0717] Table 35

[0718]

[0719] For mutations that do not map to the Oncomine transcriptome and therefore do not have a transcript-based position, the genomic position (start position) and reference nucleotide (reference allele) are used as the variant position independent of coding region or splice site proximity. The variant classification provided by the raw data is then added as a suffix. For example, a SNP missense mutation will have a variant position such as "chr19_c.C22952756_missense mutation", and a splice site SNP will have a variant position "chr1_c.A155025094_splice site". The variant changes (see below) for these mutations are not defined.

[0720] Although the suffix of the variant position is usually implicitly associated with the variant classification, when calculating hot spots, both the variant position and the variant classification are explicitly used to cluster mutations. Therefore, mutations that may generate the same variant position but have different variant classifications (such as missense and nonsense SNPs) are calculated separately.

[0721] Variant changes. Variant changes provide HGVS-like information about the alternative allele changes of the mutation. For SNP mutations in the coding region, the variant changes are complete HGVS protein-level sequence variant descriptions, indicating the alternative amino acid. For SNPs outside the coding region, the alternative allele nucleotide base is provided.

[0722] For insertion and deletion variant types, the variant position (see above) was used as the variant change. In these cases, the consequences of the changes at the amino acid level were not inferred. Thus, the variant changes of insertions / deletions do not absolutely follow the HGVS specifications.

[0723] Below are illustrative examples of variant changes derived from Compendia's mutation annotation:

[0724] Table 36

[0725] mutation Variant classification Variant location Variant changes SNP in CDS, E>K, residue 137 Missense mutation p.E137 p.E137K The SNP is located in intron C>2, two bp away from the splice site Splice site c.4913-1_splicing site c.4913-1 Inserted in CDS at residue Gly 264 Frame shift and insert p.G264_Frame shift insertion p.G264_Frame shift insertion One base deletion in the UTR 3'UTR c.*1007A_frameshift deletion c.*1007A_frameshift deletion

[0726] For mutations that did not map to the Oncomine transcriptome, the variant classification from the original data source was retained.

[0727] Transcript filtering. To avoid retrieving multiple transcripts, and therefore multiple annotations for a single mutation within a gene, we kept only one transcript per mutation per gene (unique Entrez ID). If a mutation maps to several transcripts of a gene, only one transcript is selected. However, if a mutation maps to several genes, only one transcript is selected per gene. Thus, a mutation can receive two different annotations, but only if they are from transcripts with different Entrez IDs.

[0728] We selected representative transcripts for mutations based on the following prioritization process:

[0729] Transcripts with the most influential variant classifications:

[0730] High impact in coding: missense, nonsense, no stop, frameshift

[0731] Lower impact in coding: Box, Silence

[0732] Outside the coding region: splice sites, 3' or 5' UTR, noncoding exons

[0733] Outside the exons: introns

[0734] If there is a priority-based connection, the transcript with the shortest (in terms of length) RefSeq transcript accession number is selected, followed by the alphanumeric smallest transcript accession number among further connected events. For example, among the transcripts NM_003319, NM_133378, and NM_00125685 of the TTN gene, we would select NM_003319 as the representative transcript.

[0735] These steps allow us to repeatedly select consistent transcripts for the same type of mutation at a position. One consequence of selecting the most influential transcripts is that multiple transcripts are available for mutations at multiple positions in a single gene. However, a benefit of this process is that any mutation of the same variant classification at the same position is always assigned to the same transcript, and therefore will be in the same reference frame when calculating the repetitions used for hotspot identification.

[0736] Filtering by mutation category and type. All mutations were further filtered by variant type and category. To avoid including less relevant mutations into gene function analysis, we filtered out mutations that did not resolve to gene regions, either because they fell significantly outside the transcript, or because they were in locations not associated with RefSeq genes. These mutations were evident by their lack of a gene identifier (Entrez ID=0 or blank) or membership in the following variant categories: introns, 5' flanking sequences, IGRs, and miRNAs.

[0737] We also filtered out mutations with variant types of DNP, TNP, ONP, complex substitutions, and Indels, as their annotations were not supported by our pipeline.

[0738] Since some data sources include extensive amounts of intronic and intergenic mutations, this filtering step significantly reduces the size of the dataset, as many NGS datasets do not apply these filters pre-publication.

[0739] Classification of mutations as hotspot, deleterious, or other. The next step in our analysis pipeline identifies mutations that recur across multiple samples based on their variant position and classifies them into hotspot, deleterious, or other variant categories. For this step and subsequent frequency calculations, mutations for each disease type are processed independently. Only mutations of the same variant classification are counted together, so for example, missense mutations and silent mutations at the same position are counted separately.

[0740] To identify driver genetic events, each mutation for a given Entrez Gene ID was classified as either "deleterious" or "hotspot" depending on the following criteria:

[0741] A mutation was considered 'recurrent' if it was observed in the same variant position in 3 or more tumor samples.

[0742] A mutation belongs to the "hotspot" variant category if it is:

[0743] Reproducible and

[0744] Annotated with one of the following variant classifications:

[0745] In-frame insertions / deletions

[0746] No termination

[0747] Misunderstanding

[0748] Non-coding exons

[0749] A mutation is classified as "deleterious" if it:

[0750] Non-recurring and

[0751] Annotated with one of the following variant classifications:

[0752] Frameshift insertion / deletion

[0753] No Meaning

[0754] If a mutation did not meet the above criteria, it was considered to be in the "other" variant category.

[0755] Oncomine mutation classification and variant classification can be used to summarize the relative frequency of various mutations at the gene level.

[0756] Naming "Gain of Function" and "Loss of Function" Genes. Individual genes were classified into predicted functional categories, namely "Gain of Function", "Recurrent Other", and "Loss of Function", to reflect their relative enrichment in potentially activating or deleterious mutations. Details of the process used to perform the classification are provided below.

[0757] Sample frequency calculation of mutations. The mutation frequency of each gene was calculated relative to a given variant classification and variant class in all samples within a disease type. The overall mutation frequency of genes within a disease was calculated by combining mutations from all variant classifications.

[0758] Overall mutation frequency. The overall mutation frequency of a gene was obtained by dividing the total number of samples with at least one mutation of any variant classification in the gene (mutated sample count) by the total number of samples in a given cancer type (sample count).

[0759] Hotspot frequency. The hotspot frequency of a gene is obtained by dividing the total number of samples with at least one mutation belonging to the "hotspot" Oncomine mutation classification by the mutated sample count (the total number of samples with at least one mutation for a given gene). For example, if a sample has both a hotspot missense and a hotspot in-frame deletion, it will only be counted once.

[0760] Hotspot missense frequency. To obtain the hotspot missense frequency for a gene, the number of samples containing at least one missense mutation with the Oncomine mutation classification of "hotspot" was divided by the sample count for mutations (the number of samples with at least one mutation of any type in this gene). Samples with more than one mutation of the stated type were counted only once.

[0761] Deleterious frequency. To obtain the deleterious frequency of a gene, the number of samples containing at least one mutation with the Oncomine mutation classification of "deleterious" was divided by the sample count for mutation (the number of samples with at least one mutation for a given gene). Samples with more than one mutation of the stated type were counted only once.

[0762] Other frequencies. To obtain the Other frequency for a gene, the total number of samples with at least one mutation with the Oncomine mutation classification of "Other" was divided by the Mutated Sample Count (the total number of samples with at least one mutation for a given gene). For example, if a sample contained both a splice site and a UTR mutation, it would only be counted once.

[0763] Hot Spot, Other, and Offensive Frequency Considerations. Hot Spot, Other, and Offensive Frequency should not be expected to total 100% because a sample may have been counted in more than one of these categories.

[0764] Assessing the significance of hotspot and deleterious mutations. Hotspot and deleterious p-values ​​were calculated for each gene within the disease by two independent methods.

[0765] Significance of Enrichment for Deleterious Mutations. To assess whether a gene is significantly enriched for deleterious mutations compared to other genes, given the background mutation rate, we performed Fisher's exact test using the following contingency table:

[0766] Table 37

[0767] harmful other Related genes A B All other genes C D

[0768] A, B, C, and D are mutation counts across the disease. Nonsense mutations, frameshift insertions, and frameshift deletions are classified as deleterious mutations, while any other type of mutation (UTR, silent, missense, etc., but not intergenic) is counted as others.

[0769] The Q value within each disease is calculated by counting the number of genes (N) that harbor deleterious mutations and calculating the rank for each. The q value for a given p value is then Q = p*N / rank.

[0770] Significance of hotspot mutations that are reproduced. In order to calculate the p-value specific to a gene, the significance of the hotspots that are reproduced the most on the gene is assessed. In view of the assumption that each sequence position is likely to mutate equally, each gene can be tested, whether the most reproduced is significantly greater than that expected using a polynomial test. This is an exact test of the sampling algorithm that has been implemented in the previous version. One of the advantages of this test is that the accuracy of the p-value is increased to 1E-16, so there is no bottoming. In order to obtain hotspot mutations, we filter mutations to remove any that do not affect the coding sequence (i.e., by removing silence, UTR, stop codons, and splice site mutations), and then remove the mutation data of genes that we may not annotate with RefSeq transcript identifiers. We then count the mutations observed for each transcript in each disease. We calculate the amino acid sequence length by dividing the CDS length by three and subtracting 1.

[0771] The exact calculation of the p-value is formulated as follows. Given an amino acid sequence of length x, the number of observed hotspot mutations n, what is the probability of observing r or more mutations by chance at the most recurring point. For each gene, the p-value is calculated by the following formula:

[0772]

[0773] where y (x) is the mutation count at the most reproduced hotspot, and y1,…,y X represents the mutation count at each point 1,…,x.

[0774] When n and x are large, the above formula can be very slow, using an approximation with the Bonferroni-Mallows (BM) bound:

[0775]

[0776] where y i is a Poisson random variable with mean n / x, and where Y is a truncated Poisson. P(W=n) is estimated by Edgeworth Expansion. The Bonferroni-Maros upper and lower bounds for the p-value are:

[0777]

[0778] If the approximation falls outside the BM bounds, then the lower or upper bound is used. It rarely occurs in our data, and it occurs mainly for small p-values ​​(p<1e-16) or large p-values ​​(p approximately equal to 1).

[0779] Q values ​​were calculated using the Benjamini-Hochberg method, which is Q = p*N / rank, where N is the number of transcripts and rank is the rank for each p-value.

[0780] Silent hotspot mutations. Recurrent silent mutations - Silent hotspots appear to be indicative of sequencing errors, appearing in regions of low sequence quality and acting as 'canaries in the coal mine' for false positive missense mutation peaks in adjacent nucleotides. Based on examination of genes with silent hotspots and evaluation of adjacent silent peaks, we believe that these genes are subject to systematic sequencing errors and that hotspot mutations in these genes should not contribute to gene classification.

[0781] Oncomine Gene Classification Rules. Once mutations have been classified, individual genes are named into one of three categories: "Gain of Function", "Loss of Function" and "Recurrent Other". The classification is based on a combination of the relative frequency and significance of the mutations observed in the gene. The significance of each gene mutation is assessed by a p-value.

[0782] Recurrent silent mutations. "Gain of function" genes will have a relatively high frequency of hotspot missense mutations and a low frequency of deleterious mutations, while "loss of function" genes contain a larger fraction of deleterious mutations. "Recurrent other" tends to contain recurrent indel mutations, some of which (e.g., recurrent frameshift indels of 1 base) exhibit signs of potential false positive calls that may be caused by local alignment errors. In general, we are more confident about the functional importance of genes classified as gain / loss of function.

[0783] Pan-cancer analysis. To summarize mutations across diseases, we performed the same calculations as we did for the intra-disease analysis, but without stratifying mutation records by disease. All mutation records were aggregated, and frequencies, variant classes, and gene classes were calculated in this pan-cancer scenario. For the pan-disease overview, genes in all cancer types (unique according to Entrez ID) were summarized, and one row per disease. However, an overview of genes within a disease but in a pan-cancer scenario is also provided. This means, for example, totaling samples with hotspot mutations within a disease, but only for mutations that are considered hotspots in a pan-cancer scenario. Cancer types with <20 samples are included in the pan-cancer analysis, although they are not eligible for intra-disease analysis due to low sample counts.

[0784] Cell line annotation. Cell line mutation data undergo the same Oncomine curation and annotation methods described above, except for mutation and gene classification. Instead, mutations from cell lines are annotated with Oncomine mutation classification and gene classification whenever a mutation in a cell line is also observed in a clinical sample. This annotation is only performed for mutations with hotspot or deleterious or other Oncomine mutation classifications. If a mutation is not observed in a tumor, it will receive a mutation classification of "not observed in tumor".

[0785] Mutations from cell lines and tumor samples were considered equivalent if they belonged to the same gene and had the same variant position and variant classification.

[0786] Cell line names were reviewed against the internal Oncomine ontology, and cell line cancer types were normalized to be comparable to clinical mutation data. Several cell lines whose identities or cancer types may not have been independently verified across databases or publications were removed from our analysis. Mutation annotation from clinical data was performed in both pan-cancer and within-disease contexts.

[0787] Compendia NGSDNASeq mutation calling

[0788] BAM file selection. We queried TCGA's CGHub to identify patients with a single tumor-normal BAM pair. We did this to remove the possibility of differences in mutation calls attributed to different tumor-normal pairs.

[0789] Reference genome builds. We identified the reference genome build used to align reads in the BAM file by parsing the SAM header. We located, downloaded, and indexed all reference genome builds required as input to the mutation calling package.

[0790] Mutation calling. We employed the following somatic mutation calling packages for this analysis:

[0791] MuTect (1.0.27783), Broad Institute, Cancer Genome Analysis Group (CGA) (Cibulskis, 2013)

[0792] SomaticIndelDetector (1.6-13-g91f02df), Broad Institute, Genome Analysis Toolkit (GATK)

[0793] MuTect.MuTect performs initial preprocessing to remove “reads with too many mismatches or very low quality scores” (MuTect data). Next, for candidate mutations, the LOD score describing the mutation present in the tumor sample is calculated. T ) and no mutation in normal samples (LOD N ) are two logarithmic difference (LOD) scores of the likelihood:

[0794]

[0795] MuTect expects somatic mutations to occur at a rate of approximately 1 per Mb and requires LOD T >=6.3. MuTect requires that the mutation is not in dbSNP and has LOD N >=2.3, because non-dbSNPs are expected to appear at a rate of 100 per Mb. Two cutoffs were selected to ensure that the false positive rate was less than half of the expected somatic mutation rate. Finally, other post-processing steps were performed, including testing that alternative alleles were observed in both read directions. For mutations to be considered, MuTect requires at least 14 tumor reads and 8 normal reads.

[0796] SomaticIndelDetector (SID). For a given mutation site, SID considers candidate indels using a count-based threshold and an indel consensus voting process. The indel with the largest number of supporting reads or votes is selected as the putative indel call. This call is reported if the following are present:

[0797] Sufficient coverage (default: normal >= 4 reads, tumor >= 6 reads),

[0798] Most of the reads at that site support the putative call (default: >= 30%)

[0799] This fraction is sufficiently large compared to those of any site to support indels (default: >= 70%)

[0800] Indel calls in tumor samples are annotated as "germline" if there is even weak evidence of the same indel in normal samples; otherwise, they are labeled "somatic". Calls observed only in normal samples are ignored. SID takes BAM files as input and outputs VCF and BED formatted putative calls.

[0801] Mutation filtering. The caller outputs all candidate mutation calls, including germline mutations and other calls with low statistical confidence. We filter the mutation caller output to only somatic mutations, mutations marked "KEEP" by MuTect, and mutations that appear within the CDS of RefSeq genes. The following table details the specific filters applied to MuTect and SomaticIndelDetector outputs:

[0802] Table 38

[0803]

[0804]

[0805] References

[0806] Cibulskis, K. et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnology (2013). doi:10.1038 / nbt.2514

[0807] MuTect: Hypertext Transfer Protocol : / / www.broadinstitute.org / cancer / cga / mutect

[0808] SID: Hypertext Transfer Protocol : / / gatkforums.broadinstitute.org / discussion / 35 / somatic-indel-detection Table 16: Table 2 Druggable status of genes / fusions

[0809]

[0810]

[0811]

[0812]

[0813]

[0814]

[0815]

[0816] Table 17: Table 3 Druggability status of genes / fusions

[0817]

[0818] Table 19: Gene Fusions

[0819]

[0820]

[0821]

[0822]

[0823]

[0824]

[0825]

[0826]

[0827]

[0828]

[0829]

[0830]

[0831]

[0832]

[0833]

[0834]

[0835]

[0836] Example 7 Identification of TP53 status

[0837] Advances in both molecular diagnostics and understanding cancer biology have raised the bar for the clinical trial paradigm with the expectation that more effective patient stratification will improve outcomes and expedite approval of effective cancer drugs.

[0838] The mutation status of TP53 has been identified as a predictive biomarker for treatment response and prognosis. For example, TP53 wild-type (WT) patients have been shown to exhibit significantly prolonged progression-free survival following therapy including adjuvant 5-fluorouracil and cetuximab combination therapy compared to patients harboring TP53 mutations.

[0839] TP53 mutation annotation from ONCOMINETM NGS mutation browser (Compendia Biosciences, MI) was obtained. Of the total 776 patients evaluated for TP53 mutation status; 259 patients contained at least one mutation in TP53 and were annotated as TP53 mutant, while 519 patients lacked the detected TP53 mutation and were annotated as TP53 wild-type. The TP53 wild-type and TP53 mutant annotations were then mapped at the patient level to the corresponding microarray samples from the TCGA breast cancer dataset. When the mutation annotations were mapped to patients with corresponding microarray data, 327 patients were annotated as TP53 wild-type and 188 were annotated as TP53 mutant. TP53 Wild-type and TP53 mutant signatures were generated by differential expression analysis of the TCGA breast cancer data set. The gene list was ranked by p-value according to the two-class t-test of Studen. Genes that were differentially upregulated in TP53 wild-type patients contributed to the TP53 wild-type signature, while genes upregulated in TP53 mutant patients contributed to the TP53 mutant signature. Each signature contained the top 1% of the ranked genes (n=204). All genes in the TP53 wild-type and TP53 mutant signatures were highly significant (Q<0.0001) after correcting for false discoveries. The Q value was calculated as (p-value / p-value ranking)*the number of genes measured.

[0840] Five types of ONCOMINE TM Cancer types contained sufficient TP53 mutation status data to complete the analysis. Among these, significantly elevated marker expression was found in TP53 WT compared to clinical samples with TP53 mutations from breast (p<0.001; n=189 WT, 37 mutants), lung (p=0.0003; n=23 WT, 18 mutations), liver (p=0.0069; n=74 WT, 11 mutations), and ovarian (p=0.05; n=22 WT, 15 mutations) cancer patients, and a trend was found within lymphoma patients (p=0.068; n=65 WT, 16 mutations) (see Figure 5-7 and 9-10). Table 40 contains TP53 WT TCGA breast cancer markers.

[0841] Clinically derived expression signatures effectively distinguished TP53 WT from mutant tumor samples.

[0842] Table 40

[0843]

[0844]

[0845] Unless otherwise indicated, all quantities used in this specification and claims to represent component amounts, characteristics (such as molecular weight), reaction conditions, etc. should be understood to be modified by the term "about" in all cases. Therefore, unless otherwise indicated to the contrary, the numerical parameters set forth in the specification and the appended claims are approximate values ​​that may vary depending on the desired properties attempted to be obtained by the present invention. At a minimum, and without attempting to limit the application of the doctrine of equivalents to the scope of the claims, each numerical parameter should be interpreted at least according to the number of reported significant figures and by applying general rounding techniques. Although the numerical range and parameters setting forth the wide range of the present invention are approximate values, the numerical values ​​set forth in the specific examples are reported as accurately as possible. However, any numerical value inherently contains some errors that are inevitably caused by the standard deviation found in its individual test measurements.

[0846] Unless otherwise indicated herein or obviously contradictory to the context, the terms "a / an", "said" and similar indicators used in describing the present invention (especially in the case of the appended claims) should be interpreted as covering both the singular and the plural. The description of the range of values ​​herein is intended only to serve as a shorthand method for individually referring to each individual value in the range. Unless otherwise indicated herein, each individual value is incorporated into this specification as if it is individually listed herein. Unless otherwise indicated herein or obviously contradictory to the content, all methods described herein can be performed in any suitable order. The use of any and all examples or illustrative language (e.g., "such as") provided herein is intended only to better illustrate the present invention, and does not limit the scope of the present invention otherwise required. Any language in this specification should not be interpreted as indicating that any unrequired element is necessary for practicing the present invention.

[0847] The grouping of alternative elements or embodiments of the invention disclosed herein should not be construed as limiting. Each group member can be referred to and claimed individually or in any combination with other members of the group or other elements found herein. It is expected that one or more members of a group may be included in a group or deleted from a group for reasons of convenience and / or patentability. When any of the inclusions or deletions occur, this specification is deemed to contain the group as modified, thus satisfying the written description of all Markush groups used in the appended claims.

[0848] Certain embodiments of the present invention are described herein, including the best mode known to the inventor for carrying out the present invention. Of course, after reading the above description, the changes of these described embodiments will become apparent to those skilled in the art. The inventor expects that a skilled person will adopt such changes when appropriate, and the inventor intends to practice the present invention in other ways different from those specifically described herein. Therefore, the present invention includes all modifications and equivalents of the subject matter cited in the appended claims as allowed by applicable law. In addition, unless otherwise indicated herein or otherwise obviously contradictory to the content, the present invention covers any combination of the above-mentioned elements with all possible variations thereof.

[0849] In addition, numerous references to patents and printed publications have been made throughout this specification. Each of the references and printed publications cited above is individually incorporated herein by reference in its entirety.

[0850] It should be understood that the embodiments of the invention disclosed herein illustrate the principles of the invention. Other modifications that may be employed are within the scope of the invention. Therefore, as examples and not limitations, alternative configurations of the invention may be utilized according to the teachings herein. Therefore, the invention is not limited to what is precisely shown and described.

[0851] The specific embodiments disclosed herein may be further limited in the claims using consisting of or consisting essentially of language. When used in the claims, whether as filed or added by amendment, the transition term "consisting of" excludes any element, step, or ingredient not specified in the claim. The transition term "consisting essentially of" limits the scope of the claim to the specified materials or steps and those that do not materially affect the basic and novel characteristics. The embodiments of the invention so claimed are inherently or expressly described and enabled herein.

[0852] References to sequence identifiers (such as those beginning with NM_) refer to the database accession number and sequence below as it was found on April 18, 2013.

[0853] Table 7

[0854]

[0855]

[0856]

[0857]

[0858]

Claims

1. A composition comprising a plurality of probes, wherein the plurality of probes comprises one or more probe sets, wherein each probe set specifically hybridizes to a gene fusion of a first gene and a second gene, wherein the gene fusion of the first gene and the second gene comprises one or more of the following: PRKAR1A / ALK, NCOA1 / ALK, LPP / CASR, MDM2 / EGFR, FGFR3 / ELAVL3, B2M / GNAS, DOCK8 / JAK2, HNF1B / N OTCH1, NFASC / NTRK1, SSBP2 / NTRK1, SQSTM1 / NTRK1, TBL1XR1 / PIK3CA, AKAP13 / RET, FKBP15 / RET, TBL1XR1 / RE T, CEP85L / ROS1, CLCN6 / RAF1, TRAK1 / RAF1, PRKACA / AKT1, PRKACA / AKT2, MLL / FYN, ECHD1 / FYN, and TTC13 / JAK2.

2. The composition of claim 1, wherein each gene fusion comprises the corresponding breakpoints provided in Tables 4 and 5.

3. The composition of claim 1, wherein each gene fusion comprises a corresponding breakpoint sequence provided in SEQ ID NO: 2-27. A kit comprising the composition according to claim 1.

5. The kit of claim 4, wherein each probe comprises a detectable moiety covalently bonded to a nucleic acid sequence.

6. The kit of claim 4, wherein each probe comprises a deoxyribonucleic acid sequence comprising at least one deoxyuridine residue in place of one deoxythymidine residue.

7. Use of the composition of claim 1 in the preparation of a kit for detecting one or more gene fusions in a biological sample from a subject, wherein each gene fusion is detected by the presence of an amplicon generated by amplification of the gene fusion using a corresponding probe set that specifically hybridizes to the gene fusion, and wherein amplification of the one or more gene fusions is performed using a single amplification reaction.

8. The method according to claim 7, wherein the detection is performed by a method comprising the following steps: A nucleic acid sample from a subject is contacted with a reaction mixture comprising the one or more probe sets.

9. The use according to claim 7, wherein the detection is performed by a method comprising PCR and / or sequencing, wherein the PCR comprises quantitative real-time PCR and the sequencing comprises next generation sequencing.

10. The use according to claim 7, wherein the sample is a tumor sample.

Citation Information

Patent Citations

  • Method for detecting nucleic acid amplification using self-quenching fluorescence probe

    US5538848A

  • Detectably labeled dual conformation oligonucleotide probes, assays and kits

    US5925517A

  • Detectably labeled, dual conformation oligonucleotide probes, assays and kits

    US6103476A

  • Nucleic acid detection probes having non-FRET fluorescence quenching and kits and assays including such probes

    US6150097A

  • Methods, kits and compositions pertaining to PNA molecular beacons

    US6355421B1

Cited By

  • Biological information processing system and method applied to synthetic biology

    CN120708691A

  • A bioinformatics processing system and method applied to synthetic biology

    CN120708691B