NZ625468B2 - Methods and materials for assessing loss of heterozygosity - Google Patents

Abstract

method for assessing loss of heterozygosity (LOH) in a cancer cell or genomic DNA thereof, comprises (a) detecting, in a cancer cell or genomic DNA derived therefrom, LOH regions in at least one pair of human chromosomes of the cancer cell with at least one pair of human chromosomes not being a human X or Y sex chromosome pair; and (b) determining the total number of LOH regions, in the at least one pair of human chromosomes, that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, with the first length being about 1.5 or more megabases. A method of predicting the status of BRCA1, BRCA2 or HDR genes in a cancer cell, comprises determining, in the cancer cell, the total number of LOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, with at least one pair of human chromosomes not being a human X or Y sex chromosome pair with the first length being about 1.5 or more megabases; and correlating the total number that is greater than a reference number with an increased likelihood of a deficiency in the HDR, BRCA1 or BRCA2 gene. man X or Y sex chromosome pair; and (b) determining the total number of LOH regions, in the at least one pair of human chromosomes, that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, with the first length being about 1.5 or more megabases. A method of predicting the status of BRCA1, BRCA2 or HDR genes in a cancer cell, comprises determining, in the cancer cell, the total number of LOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, with at least one pair of human chromosomes not being a human X or Y sex chromosome pair with the first length being about 1.5 or more megabases; and correlating the total number that is greater than a reference number with an increased likelihood of a deficiency in the HDR, BRCA1 or BRCA2 gene.

Description

W0 2013/096843 PCT/U52012/071380 METHODS AND MATERIALS FOR ASSESSING LOSS OF HETEROZYGOSITY CROSS-REFERENCE TO RELATED APPLICATIONS This application is claims priority to US. Provisional Patent Application Serial No. 61/578,713 filed December 21, 2011 and US. Provisional Patent Application Serial No. 61/654,402 filed June l, 2012, the entire contents ofwhich are hereby incorporated by reference.

BACKGROUND 1. Technical Fieid This nt relates to methods and materials involved in assessing samples (e.g., cancer cells) for the ce of a loss of heterozygosity (LOH) ure. For example, this document provides methods and materials for determining whether or not a cell (e.g., a cancer cell) contains an LOH signature. This document also provides materials and methods for identifying cells (e.g., cancer cells) having a deficiency in homology ed repair (HDR) as Well as materials and methods for identifying cancer patients likely to respond to a particular cancer treatment regimen.

Throughout this document, unless indicated otherwise, HDR deficiency and HRD (homologous repair deficiency) are used synonymously. 2. Background Information Cancer is a serious public health m, with 562,340 people in the United States of America dying ofcancer in 2009 alone.

American Cancer Society, Cancer Facts & Figures 2009 (available at American Cancer Society website). One of the primary challenges in cancer treatment is discovering relevant, clinically usefiil characteristics of a patient’s own cancer and then, based these characteristics, administering a treatment plan best suited to the t’s . While strides have been made in this field of personalized ne, there is still a significant need for better molecular diagnostic tools to characterize ts’ cancers.

SUMMARY In l, one aspect of this invention features a method for assessing LOH in cancer cell or genomic DNA thereof. In some ments, the method comprises, or consists essentially of, (a) detecting, in a cancer cell or genomic DNA derived therefrom, LOH regions in at least one pair of human chromosomes of the cancer cell (e. g., any pair of human chromosomes other than a human X/Y sex some pair); and (b) determining the number and size (e.g., length) of said LOH regions. In some embodiments, LOH s are analyzed in a number of chromosome pairs that are representative of the entire genome (e. g., enough chromosomes are analyzed such that the number and size ofLOH regions are expected to be representative of the number and size of LOH regions across the genome). In some ments, the method further comprises ining the total number ofLOH regions that are longer than about 1.5, 5, 12, I3, 14, 15, 16, 17 or more (preferably 14, 15, 16 or more, more preferably 15 or more) megabases but shorter than the entire length of the respective chromosome which the LOH region is d within (Indicator LOH Regions). Alternatively or additionally, the total combined length of such Indicator LOH Regions is determined. In some specific embodiments, if that total number of Indicator LOH s or total combined length of Indicator LOH Regions is equal to or greater than a predetermined reference number, then said cancer cell or genomic DNA or a patient having said cancer cell or genomic DNA is identified as having an HDR—deficiency LOH signature.

An alternative method for assessing LOH in a cancer cell or genomic DNA thereof is also provided which comprises, or consists essentially of, (a) detecting, in a cancer cell or genomic DNA derived therefrom, LOH s in at least one pair of human chromosomes of the cancer cell, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair; and (b) determining the total number and/or combined length of LOH regions, in the at least one pair of human chromosomes, that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the first length is about 1.5 or more (or 5, 10, I3, 14, 15, 16 or more, preferably 15 or more) megabases. In some Specific embodiments, if that total number or ed length is equal to or greater than a predetermined reference , then said cancer cell or genomic DNA or a patient having said cancer cell or genomic DNA is identified having an HDR-deficiency LOH signature.

In another aspect, the present invention provides a method of predicting the status of BRCA1 and BRCA2 genes in a cancer cell. The method comprises, or consists essentially of, ining, in the cancer cell, the total number and/or combined length ofLOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length ofthe whole chromosome containing the LOH region, wherein the at least one pair of human chromosomes is not a human X/Y sex some pair, wherein the first length is about 1.5 PCT/U82012/07l380 deficiency in the BRCA1 or BRCA2 gene.

In another aspect, this invention es a method of predicting the status of HDR in a cancer cell. The method comprises, or consists essentially of, determining, in the cancer cell, the total number and/or combined length of LOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair of human chromosomes is not a human X/Y sex some pair, n the first length is about 1.5 or more (or 5, 10 or more, preferably about 15 or more) megabases; and correlating the total number or combined length that is greater than a reference number with an increased likelihood of a deficiency in HDR.

In another aspect, this ion provides a method of predicting a cancer patient’s se to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor.

The method comprises, or consists essentially of, determining, in a cancer cell from the cancer patient, the number and/or combined length ofLOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, n the at least one pair of human chromosomes is not a human X/Y sex some pair, wherein the first length is about 1.5 or more (or 5, 10 or more, ably about 15 or more) megabases; and correlating the total number or combined length that is greater than a reference number with an increased likelihood that the cancer patient will respond to the cancer ent regimen. In some embodiments, the patients are treatment naive patients.

In another aspect, present invention relates to a method ofpredicting a cancer patient’s reSponse to a treatment regimen. The method comprises, or consists essentially of, determining, in a cancer cell from the cancer patient, the total number and/or ed length of LOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair of human somes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more (or 5, 10 or more, preferably about 15 or more) megabases; and correlating the total number or combined length that is greater than PCTfU82012/071380 reference number with an increased likelihood that the cancer patient will not respond to a treatment regimen including paclitaxel or docetaxel.

In another aspect, this ion is directed to a method of ng cancer. The method comprises, or consists essentially of, (a) determining, in a cancer cell from a cancer t or c DNA obtained therefrom, the total number and/or combined length ofLOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair of human chromosomes is not a human X/Y sex some pair, wherein the first length is about or more (or 5, 10 or more, preferably about 15 or more) megabases; and (b) administering to the cancer patient a cancer treatment regimen comprising one or more drugs chosen from the group consisting ofDNA damaging , anthracyelines, topoisomerase 1 inhibitors, and PARP inhibitors, if the total number or combined length of LOH regions is greater than a reference number.

In some embodiments, the patients are treatment naive patients.

In some embodiments of any one or more of the methods described in the preceding six paragraphs, any one or more of the following can be applied as appropriate. The LOH regions can be determined in at least two, five, ten, or 21 pairs of human chromosomes. The cancer cell can be an ovarian, breast, or esophageal cancer cell. The first length can be about 6, 12, or about 15 or more megabases. The reference number can be 6, 7, 8, 9, 10, ll, 12, 13, 14, 15, 16, 17, 18 or 20 or greater. The at least one pair of human chromosomes can exclude human chromosome 17. The DNA damaging agent can be eisplatin, latin, latin, or atin, the anthracycline can be epirubincin or doxorubiein, the topoisomerase I inhibitor can be campothecin, topoteean, or irinoteean, or the PARP inhibitor can be iniparib, olaparib or velapirib.

In another aSpect, this invention features the use ofone or more drugs ed from the group consisting ofDNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors, in the manufacture of a medicament usefiJl for treating a cancer in a t identified as having a cancer cell determined to have a total of 5, 8, 9, 10, 12, 15, 17, 20 or more tor LOH Regions. The Indicator LOH Regions can be determined in at least two, five, ten, 2] pairs of human chromosomes. The cancer cell can be an ovarian, breast, or esophageal cancer cell. The Indicator LOH Regions can have a length of about 6, 12, or 15 or more megabases. The Indicator LOH Regions can be t on a chromosome other than human chromosome 17. The DNA damaging agent can be a platinum—based chemotherapy drug, the anthracycline can be PCT/U52012/071380 epirubincin or doxorubicin, the topoisomerase I inhibitor can be campothecin, topotecan, or irinotecan, or the PARP inhibitor can be inipan'b, olaparib 0r velapirib. In some embodiments, the patients are treatment naive patients. , ve, ten, or 21 pairs of human chromosomes. The cancer cell can be an n, breast, or geal cancer cell. The Indicator LOH Regions can have a length of about 6, 12, or 15 or more megabases. The Indicator Regions can be t on a chromosome other than human chromosome 17.

In another aspect, this i ' cancer cell of a cancer patient. The system comprises, or consists essentially of, (a) a sample analyzer configured to produce a plurality of signals about c DNA of at least one pair of human chromosomes ofthe cancer cell, and (b) a computer stem programmed to calculate, based on the plurality of signals, the number or combined length of Indicator LOH Regions in the at least one pair of human chromosomes.

The computer sub-system can be programmed to compare the number or combined length of Indicator LOH Regions to a reference number to determine (a) a likelihood of a deficiency in BRCAI and/0r BRCA2 genes in the cancer cell, (b) a likelihood of deficiency in HDR in the cancer cell, or (c) a likelihood that the cancer patient will respond to cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, , or 21 pairs ofhuman chromosomes. Thc cancer cell can be an n, breast, or esophageal cancer cell. The Indicator LOI-I Regions can have a length of about 6 12, or 15 or more megabases. The Indicator LOH Regions can be present on chromosomes other PCT/U82012/071380 than a human chromosome 17. The DNA damaging agent can be a platinum-based chemotherapy drug, the anthracycline can be epirubincin or doxorubicin, the topoisomerase I inhibitor can be campothecin, tOpotecan, or irinotecan, or the PARP inhibitor can be iniparib, olaparib or rib.

In another aspect, the invention provides a er program product embodied in computer readable medium that, when executing on a computer, provides ctions for detecting the presence or absence of any LOH region along one or more of human chromosomes other than the human X and Y sex chromosomes, and the LOH region having a length of about 1.5 or more (or , 10 or more, preferably 15 or more) megabases but shorter than the length of the whole chromosome containing the LOH region; and determining the total number or combined length of the LOH regions in the one or more chromosome pairs. The computer program product can include other instructions. The Indicator LOH Regions can be determined in at least two, five, ten or 21 pairs of human chromosomes. The cancer cell can be an ovarian, breast, or esophageal cancer cell.

The Indicator LOH Regions can have a length of about 6, 12, or 15 or more megabases. The Indicator LOH Regions can be present on somes other than a human chromosome 17. The DNA damaging agent can be a platinum-based chemotherapy drug, the anthracyeline can be incin or doxorubicin, the topoisomerase 1 tor can be campothecin, topotecan, eean, or the PARP inhibitor can be ib, olaparib or velapirib.

In another aspect, the present invention provides a diagnostic kit. The kit comprises, or consists essentially of, at least 500 oligonucleotides e of hybridizing to a plurality of rphic regions of human c DNA; and a computer program product provided herein.

The computer program product can be embodied in a computer readable medium that, when executing on a computer, provides instructions for detecting the presence or absence of any LOH region along one or more ofhuman chromosomes other than the human X and Y sex chromosomes, and the LOH region having a length of about 1.5 or more (or 5 or 10 or more, preferably about 15 more) megabases but shorter than the length of the whole chromosome containing the LOH region; and determining the total number and/or combined length of the LOH region in the one or more chromosome pairs.

In another aspect, this document features a method for assessing cancer cells of patient for the presence of an LOH signature. The method comprises, or ts essentially of, detecting the presence ofmore than a nce number ofLOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than PCT/U82012/071380 In another aspect, this document features a method for assessing cancer cells of a patient for the presence of an HDR deficient . The method comprises, or consists essentially of, (a) detecting the presence of more than a reference number ofLOH regions in at 1 east one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first pair, wherein the first length is about 1.5 or more megabases, and (b) identifying the patient as having cancer cells with the HDR deficient status.

LOH regions in at least one pair ofhuman chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair ofhuman chromosomes is not a human X/Y chromosome pair, wherein the first length is about 1.5 or more mcgabases, and (b) identifying the patient as having cancer cells with the genetic on.

LOH region, wherein the at least one pair of human chromosomes is not a human X/Y sex some pair, n the first length is about 1.5 or more megabases, and (b) identifying the patient as being likely to d to the cancer treatment PCT/U82012/071380 having an LOH signature, wherein the presence of more than a nce number ofLOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than first length but shorter than the length ofthe whole chromosome containing the LOH region indicates that the cancer cells have the LOH signature, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or s in at least one pair of human somes of a cancer cell of the cancer patient that longer than a first length but shorter than the length of the whole chromosome containing the LOH region indicates that the cancer cells have the HDR deficiency status, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) diagnosing the patient as having cancer cells with the HDR deficient status.

In r aspect, this document features a method for assessing a patient. The method comprises, or consists essentially of, (a) determining that the t comprises cancer cells having a genetic mutation within a gene from an HDR pathway, wherein the presence ofmore than a reference number of LOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region indicates that the cancer cells have the genetic mutation, wherein the at least one pair of human somes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) diagnosing the patient as having cancer cells with the genetic mutation. tors, and PARP inhibitors. The method comprises, or consists essentially of, (a) determining that the patient comprises cancer cells having an LOH signature, wherein the ce of more than a reference number of LOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than the length of the whole PCT/U82012/071380 chromosome containing the LOH region indicates that the cancer cells have the LOH signature, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) diagnosing, based at least in part on the presence of the LOH signature, the patient as being likely to respond to the cancer ent regimen. analysis of a cancer cell of a patient. The method comprises, or consists essentially of, (a) detecting the ce of more than a nce number ofLOH regions in at least one pair of human somes ofthe cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH , wherein the at least one pair ofhuman chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) identifying the patient as having cancer cells with an LOH signature. analysis of a cancer cell of a patient. The method comprises, or consists essentially of, (a) detecting the presence of more than a reference number of LOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) identifying the patient as having cancer cells with a HDR deficient status.

In another aspect, this document features a method for performing a diagnostic analysis of a cancer cell of a patient. The method ses, or consists essentially of, (a) detecting the presence of more than a reference number ofLOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair ofhuman chromosomes is not a human X/Y sex chromosome pair, n the first length is about 1.5 or more megabases, and (b) identifying the patient as having cancer cells with a genetic on within a gene from an HDR pathway.

PCT/U82012/07] 380 method comprises, or consists essentially of, (a) detecting the presence of more than a reference number ofLOH regions in at least one pair of human chromosomes of the cancer cell that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) identifying the patient as being likely to respond to the cancer treatment regimen.

In another aspect, this document features a method for sing a patient as having cancer cells having an LOH signature. The method comprises, or consists ially of, (a) determining that the patient comprises cancer cells having the LOH signature, wherein the presence of more than a reference number ofLOH regions in at least one pair of human chromosomes of a cancer cell of the cancer t that are longer than a first length but shorter than the length of the whole chromosome containing the LOH region indicates that the cancer cells have the LOH signature, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, n the first length is about 1.5 or more megabases, and (b) diagnosing the t as having cancer cells with the LOH signature.

In another aspect, this document features a method for diagnosing a patient as having cancer cells with an HDR deficient status. The method comprises, or consists essentially of, (a) determining that the patient comprises cancer cells having the HDR deficiency status, wherein the presence of more than a reference number of LOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but r than the length of the whole chromosome containing the LOH region indicates that the cancer cells have the HDR deficiency status, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) diagnosing the patient as having cancer cells with the HDR deficient status.

In another , this document features a method for diagnosing a patient as having cancer cells with a genetic mutation within a gene from an HDR pathway. The method comprises, or consists essentially of, (a) ining that the patient comprises cancer cells having the genetic on, wherein the presence of more than a reference number ofLOH regions in at least one pair of human chromosomes of a cancer cell of the cancer patient that are longer than a first length but shorter than the length of the whole some containing the LOH region tes that the cancer cells have the c mutation, wherein the at least one pair of human chromosomes is not a WO 96843 PCT/U82012/07I380 human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) sing the patient as having cancer cells with the genetic mutation, In another aspect, this document features a method for diagnosing a t as being a candidate for a cancer treatment regimen comprising administering radiation or a drug selected from the group consisting ofDNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors. The method comprises, or consists essentially of, (a) ining that the patient comprises cancer cells having an LOH signature, wherein the presence ofmore than a reference containing the LOH region indicates that the cancer cells have the LOH signature, wherein the at least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases, and (b) diagnosing, based at least in part on the ce of the LOH ure, the patient as being likely to respond to the cancer treatment regimen.

Unless otherwise , all technical and scientific terms used herein have the same g as ly understood by one of ordinary skill in the art to which this invention pertains.

Although methods and materials similar or equivalent to those described herein can be used to practice the invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

The details of one or more embodiments of the invention are set forth in the description and accompanying drawings below. The als, methods, and examples are illustrative only and not intended to be limiting.

Other features, objects, and advantages of invention will be apparent from the description and drawings, and from the claims.

DESCRIPTION OF THE DRAWINGS Figure 1 is a graph plotting allele dosages of breast cancer cells from a breast cancer patient along chromosome 1 as determined using a SNP array. The arrow indicates a transition between a region of heterozygosity and an LOH region.

Figure 2 is a graph plotting allele dosages of breast cancer cells for the same breast cancer patient as on Figure 1 along chromosome 1 as determined using high-throughput sequencing.

Figure 3 is a flow chart of an example process for ing the genome of a cell (e.g, a cancer cell) for an LOH signature. with intact or deficient BRCA1 and BRCA2 genes. The size of the circles is proportional to the number of samples with such number ofLOH regions.

Figure 9 is a table showing the percent of ovarian cancer samples that are BRCA deficient, HDR deficient/BRCA intact, and HDR intact.

Figure 10 is a graph plotting the number ofLOH regions longer than Mb and shorter than the entire chromosome for cancer cell lines for the indicated cancers. The size of the s is proportional to the number of s with such number ofLOH regions PCT/U82012/071380 Figure 11 is a graph plotting the number of LOH regions longer than 15 Mb and shorter than the entire Chromosome for lung cancer samples.

Figure 12 is a graph plotting the percentage of the indicated cancers or cancer cell lines having an HDR deficiency.

Figure 13 contains graphs plotting the 1050 values (Logmflcso) of camptothecin, well as averaged L0g10(IC50) values for platinum nds (oxaliplatin, cisplatin, and carboplatin). or anthracyclines (doxorubicin and epirubicin) when exposed to 29 breast cancer cell lines having the indicated number of LOH regions longer than 15 Mb and shorter than the entire chromosome or the ICSO values (Logm(IC50)) of axel when exposed to 27 n cancer cell lines having the indicated number ofLOH regions longer than 15 Mb and r than the entire chromosome. The dashed lines place a threshold number at nine.

Figure 14 is a d n of a graph from Figure 13 that plots the averaged Log10(IC50) values of platinum compounds (oxaliplatin, cisplatin, and carboplatin) when exposed to 29 breast cancer cell lines having the indicated number of LOH regions longer than 15 Mb and shorter than the entire chromosome.

Figure 15 is a flow chart of an example computational process for identifying LOH loci and regions.

Figure 17 shows HRD score in tumor samples. Blue circles: BRCAI or BRCA2 deficient samples. Red circles: BRCA1 and BRCA2 intact samples. Combined area under blue and red circles is the same. The area of each individual circle is proportional to the number of s with the corresponding number of LOH regions.

Figure 17a. HRD score for the first cohort (46 of 152 samples were BRCAl or BRCA2 deficient).

Figure 17b. HRD score for the second cohort (19 of 53 samples were BRCAl or BRCA2 deficient).

PCT/U52012/071380 Figure 17c. HRD score for the third cohort (146 of 435 samples were BRCAl or BRCA2 deficient).

BRCA2 mutants; D: 82 samples with low expression or methylation of BRCA1; E: 13 samples with methylation of RADSlC. Red circles: 416 samples with BRCAI, BRCA2, and RADS 1 C intact genes.

Figure 18a. Comparison ofHRD scores in cancer cell lines. Red circles: cell lines with intact BRCA] or BRCA2. A: intact arian cell lines; B: 22 intact ovarian cell lines.

Green circles: 6 carriers of heterozygous ons in either BRCAI or BRCA2. Violet s: 2 carriers of homozygous ons with reversion in either BRCAl or BRCA2.

Blue s: 7 carriers ofhomozygous mutations in either BRCAl or BRCA2 or with methylated BRCAl. The combined area under the green, red, blue, and violet s is the same. The area under each individual circle is proportional to the number of s with the corresponding number ofLOH regions.

Figure 18b. Kaplan—Meier plot of OS post-surgery for HRD score split at its median.

These data were generated using 507 samples from the TCGA dataset for which COpy number data and survival information were available. Median OS for samples with high and low HRD score were 1499 (95% Cl=(1355-l 769)) and 1163 (95% CI=(l 081-1354» days, respectively.

Figure 19 shows the ation between LOH scores and HR deficiency calculated for different LOH region length cut-offs for the first cohort. Corresponding log10(p-value) are on the y—axis. The relationship between the f of the size of LOH regions and the significance correlation of the LOH score with HR deficiency was investigated. This figure shows that LOH length cut—offs may readily range from I] to 21 Mb. The cut-off of 15 Mb, approximately in the middle of the interval, may be used in some preferred embodiments since it was found to be more sensitive to statistical noise present in the data.

Figure 20 shows comparison ofLOH scores in three groups ofBRCA1 and BRCA2 deficient samples for the combined data from all three cohorts. Row A: 49 rs of germline mutations in BRCAl; B: 25 carriers of somatic mutations in BRCA]; C: 82 samples with either methylation or low expression of BRCA l; D: 27 carriers of germline mutations in BRCAZ; E: 9 carriers of somatic mutations in BRCA2.

PCT/USZOIZ/071380 Figure 21 shows a comparison of LOH scores of BRCA1, BRCA2, and RADS l C deficient samples. Blue circles correspond to BRCAl deficient samples, red circles correspond to BRCA2 deficient samples, and green s correspond to RADSIC deficient samples. The combined area under red, blue, and green circles is the same. The area under each individual circle proportional to the number of samples with the correSponding number ofLOH regions.

Figure 22 shows a comparison of LOH ) scores in ts who responded versus patients who did not respond to treatment comprising platinum therapy. The area under each individual circle is proportional to the number of samples with the corresponding number ofLOH regions.

Figure 23 shows a comparison ofLOH (“HRD”) scores in BRCAl or BRCA2 deficient samples. The area under each individual circle is proportional to the number of samples with the corresponding number of LOH regions. One r sample with significant contamination is highlighted.

Figure 24 shows the on of non-reSponders in each group of patients with a given LOH (“HRD”) score.

DETAILED DESCRIPTION This document provides methods and materials involved in assessing samples (e.g., cancer cells) for the presence of an LOH signature. For example, this document provides methods and als for determining whether or not a cell (e.g., a human cancer cell) contains an LOH signature (e.g., a HDR—defieiency LOH ure).

In general, a comparison of sequences t at the same locus on each some (each autosomal chromosome for males) can reveal whether that particular locus is homozygous heterozygous within the genome of a cell. Polymorphic loci within the human genome are generally heterozygous Within an individual since that individual typically receives one copy from the biological father and one copy from the ical mother. In some cases, a polymorphic locus or a string ofpolymorphic loci within an individual are homozygous as a result of inheriting identical copies from both biological parents.

Loss of heterozygosity (LOH) may result from several mechanisms. For example, in some cases, a region of one chromosome can be d in a somatic cell. The region that remains W0 2013/096843 PCT/U82012/071380 present on the other chromosome (the other non-sex chromosome for males) is an LOH region as there is only one copy (instead of two copies) of that region present within the genome of the affected cells. This LOH region can be any length (e.g., from a length less than about 1.5 Mb up to a length equal to the entire length of the chromosome). This type of LOH event results in a copy number reduction. In other cases, a region of one chromosome (one non-sex chromosome for males) in a somatic cell can be ed with a copy of that region fi'om the other chromosome, thereby eliminating any zygosity that may have been present within the replaced region. In such cases, the region that remains present on each some is an LOH region and can be referred to as a copy neutral LOH region. Copy neutral LOH regions can be any length (e.g., from a length less than about 1.5 Mb up to a length equal to the entire length of the some).

As described herein, a cellular sample (e.g., cancer cell sample) can be identified having a “positive LOH signature status” (or alternatively called “HDR-deficiency LOH ure”) if the genome of the cells being assessed contains five or more (e. g, six or more, seven or more, eight or more, nine or more, ten or more, eleven or more, 12 or more, 13 or more, 14 or more, 15 or more, 16 or more, 17 or more, 18 or more, 19 or more, or 20 or more) LOH regions that are (a) longer than about 1.5 megabases (e.g., longer than about 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10, ll, 12, 13, 14, , 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, 50, 75, or 100 megabases (Mb), preferably longer than about 14 or 15 or 16, more preferably longer than about 15 megabases) and (b) less than the length of the entire some that contains that LOH region. In some cases, a cancer cell sample can be identified as having a positive LOH signature status if the genome of the cells being assessed contains nine or more LOH regions that are (a) longer than about 15 Mb and (b) less than the length of the entire chromosome that contains that LOH . Unless otherwise defined, the term “Indicator LOH Region” refers to an LOH region that is in a pair of human chromosomes other than the human X/Y sex chromosome pair, and that is characterized by loss of heterozygosity with a length of about 1.5 or more megabases but shorter than the length of the Whole chromosome is any LOH region about 2, 2.5, 3, 4, , 6, 7, 8, 9,10, 11,12, 13,14 15,16,17,18, 19, 20, 25, 30, , 40, 45, 50, 75, or 100 megabases (Mb) or more (preferably longer than about 14 or 15 megabases) and less than the length of the whole chromosome that contains that LOH region.

PCT/U82012/071380 Cells (6g, cancer cells) identified as having a positive LOH signature (also termed herein “HDR—deficiency LOH ure”) can be classified as having an increased likelihood of having an HDR deficiency and/or as having an increased likelihood ofhaving a deficient status in one or more genes in the HDR pathway.

For example, cancer cells identified as having a positive LOH signature status can be classified as having an increased likelihood of having an HDR deficient . In some cases, cancer cells fied as having a positive LOH signature status can be classified as having an increased likelihood of having a deficient status for one or more genes in the HDR pathway. As used herein, deficient status for a gene means the sequence, structure, expression and/or activity of the gene or its product is/are deficient as compared to . Examples include, but are not limited to, low or no mRNA or protein expression, deleterious mutations, hypermethylation, attenuated activity (e.g., enzymatic activity, y to bind to another biomolecule), etc. As used , deficient status for a pathway (e.g., HDR pathway) means at least one gene in that pathway (e.g., BRCAI) is deficient. Examples ofhighly deleterious mutations e hift ons, stop codon mutations, and mutations that lead to altered RNA splicing.

Deficient status in a gene in the HDR pathway may result in deficiency or reduced activity in homology directed repair in the cancer cells.

Examples ofgenes in the HDR pathway include, without limitation, the genes listed in Table 1.

Table 1. Selected HDR Pathway Genes l Entrez Gene Gene Entrez Gene Entrez Gene Symbol (if I Entrez Name Symbol (if Gene ld Name assigned) Gene Id assigned) BLM l BLM 64l—l_RAD50 RAD50 101111 POLD4 57804 RAD54B 12412543 25788 DNA + l— ‘1 polymerase POLH 5429 RM]! RM]! 80010 DNAZ JDUN/12 I763 r7111412 ] C16orf75 116028 EME] EMEI 146956 RPA RPAI 6117 ERCCI ERCC] 2067 RTELZ RTELI 51750 EXO] Jr59101 ‘l 9156 J—SLXI FANCM FANCM _J 57697 SLXZ l— GEN] GEN] 348654 _J SLX4 84464 MREII Jr 11—81214 MREIIA J 4361 TOPZA —_l TOPZA 7153 MUS8J MUS81 80198TX’PF ERCC4 2072'] NBS] NBN 4683—1371062 L 172002 7516—] PALB2 PALB2 79728_| XRCC3 XRCC3 7517 PCNA PCNA 517'— —_l Examples ofgenetic mutations that can be t within a gene of the HDR pathway include, t limitation, those listed in Table PCT/USZ012/07I380 l RAD51C L138F 5889 Bone Y75stX0 5889 In some cases, a cellular sample (e.g., cancer cell sample) can be fied as having an increased number of LOH regions (e.g., at least 7, 8, 9, 10, or more LOH regions) that cover the whole chromosome. Cells (e.g., cancer cells) identified as having an increased number ofLOH regions that cover the whole chromosome can be classified as having an increased likelihood of having HDR proficiency, that is, intact HDR pathway. For example, cancer cells identified having an increased number ofLOH regions that cover the whole chromosome can be classified as being more likely to have intact BRCA1 and BRCA2 genes.

As described , fying LOH loci (as well as the size and number ofLOH regions) can include, first, determining the genotype of a sample at various genomic loci (e.g., SNP loci, individual bases in large sequencing) and, second, determining whether homozygous loci due to LOH events. Any appropriate technique can be used to determine genotypes at loci of interest within the genome of a cell. For e, single nucleotide polymorphisms (SNP) arrays (e.g., human genome-wide SNP arrays), targeted sequencing of loci of interest (e. g, sequencing SNP loci and their surrounding sequences), and even untargeted sequencing (e.g., whole exome, transcriptome, or genome sequencing) can be used to identify loci as being homozygous 0r heterozygous. In some cases, an is ofthe homozygous or heterozygous nature of loci over a length of a chromosome can be med to determine the length of regions of homozygosity heterozygosity. For example, a stretch of SNP locations that are spaced apart (e. g., spaced about 25 kb to about 100 kb apart) along a chromosome can be ted using SNP array results to determine not only the presence of a region of gosity along a chromosome but also the length of that region. s from a SNP array can be used to generate a graph that plots allclc dosages along a chromosome. Allele dosage di for SNP i can be calculated from adjusted signal ities of two alleles (A, and Bi): di = Ai/(Ai + Bi). An example of such a graph is presented in Figure 1. Numerous variations on nucleic acid arrays usefiil in the invention are known in the art.

These include the arrays used in the various examples below (cg, Affymetrix 500K GeneChip array in Example 3; Affymetrix OneoScanTM FFPE Express 2.0 es (Formerly MIP CN Services) in Example 4).

WO 96843 PCTfU82012/071380 Once a ’s pe has been determined for a plurality of loci (e.g., SNPs), common techniques can be used to identify loci and regions ofLOH. One way to determine whether homozygosity is due to LOH is to compare the somatic genotype to the germline. For e, the genotype for a plurality of loci (e.g., SNPs) can be ined in both a germline (e.g., blood) sample and a somatic (e.g, tumor) sample. The genotypes for each sample can be compared (typically computationally) to determine where the genome ofthe gennline cell was heterozygous and the genome of the somatic cell is homozygous.

Such loci are LOH loci and regions of such loci are LOH regions.

Computational techniques can also be used to determine whether homozygosity is due to LOH. Such techniques are particularly useful when a germline sample is not available for analysis and comparison. For example, thms such as those described elsewhere can be used to detect LOH regions using information from SNP arrays (Nannya et a/., Cancer Res. (2005) 65:6071- 6079 (2005)). Typically these algorithms do not explicitly take into account ination of tumor samples with benign tissue. Cf International Application No.

PCT/U32011/026098 to Abkevich et (1].; Goransson et (21., PLoS One (2009) 4(6):e6057. This contamination is often high enough make the detection ofLOH regions challenging.

Improved analytical methods according to the present invention for identifying LOH, even in spite of contamination, include those embodied in computer software products as described below.

The following is one e. lfthe ed ratio ofthe signals of two alleles, and B, is two to one, there are two possibilities.

The first possibility is that cancer cells have with deletion of allele B in a sample with 50% contamination with normal cells. The second possibility is that there is no LOH but allele A is duplicated in a sample with no contamination with normal cells. An thm can be implemented as a computer program as described herein to reconstruct LOH regions based on genotype (e.g., SNP genotype) data. One point of the algorithm is to first reconstruct allele specific copy numbers (ASCN) at each locus (e. g, SNP). ASCNS are the numbers of copies ofboth paternal and al alleles. An LOH region is then determined as a algorithm designed to reconstruct total c0py number (rather than ASCN) at each locus (e.g., SNP).

See lntemational Application No. PCT/U820] 1/026098 to Abkevich et a]. The hood function can be maximized over ASCN of all loci, level of contamination with benign tissue, total copy PCT/U52012/071380 number ed over the whole genome, and sample specific noise level. The input data for the algorithm can include or t of (I) sample-specific normalized signal intensities for both allele of each locus and (2) assay-specific approach) set ofparameters defined based on analysis of large number of samples with known ASCN profiles.

In some cases, nucleic acid sequencing techniques can be used to identify loci as being homozygous or heterozygous. For e cell sample) can be ted and nted. Any riate method can be used to extract and fragment genomic nucleic acid including, without limitation, commercial kits such as QIAampTM DNA Mini Kit (QiagenTM), MagNATM Pure DNA Isolation Kit (Roche Applied ScienceTM) and GenEluteTM Mannnalian Genomic DNA Miniprep Kit (Sigma-AldrichTM). Once extracted and fragmented, either targeted or untargeted sequencing can be done to determine the sample’ genotypes at loci. For example, whole genome “loci” to be ted).

In some cases, ed sequencing ofknown polymorphic loci (e.g., SNPs and surrounding sequences) can be done as an alternative to microarray analysis. For example, the genomic DNA can be enriched for those nts containing a locus (e.g., SNP location) to be analyzed using kits designed for this purpose (e.g., Agilent SureSelectTM , Illumina TruSeq Capturem, and Nimblegen SeqCap EZ ChoiceTM).

For example, genomic DNA containing the loci to be analyzed can be hybridized to biotinylated capture RNA fragments to form biotinylated RNA/genomic DNA complexes. Alternatively, DNA capture probes may be utilized resulting in the formation of biotinylated DNA/genomic DNA hybrids. Streptavidin coated magnetic beads and a magnetic force can be used to separate the biotinylated RNA/genomic DNA xes from those genomic DNA fragments not present within a ylated RNA/genomic DNA complex. obtained biotinylated RNA/genomic DNA complexes can be treated to remove the captured RNA from the magnetic beads, thereby leaving intact genomic DNA fragments containing a locus to be analyzed. These intact genomic DNA fragments containing the loci to be analyzed can be amplified using, for e, PCR techniques. The amplified genomic DNA fragments can be sequenced using a high-throughput sequencing technology or a next-generation sequencing technology such as 2012/071380 nmmmanemnflmmmminsmmfimarmmmmgasmnnm or Ion Torrentm, or Roche 454“% The sequencing s from the genomic DNA fragments can be used to identify loci as being gous or heterozygous, analogous to the microarray analysis described herein. some cases, an analysis of the homozygous or heterozygous nature of loci over a length of a chromosome can be performed to determine the length of s of homozygosity or heterozygosity. For example, a stretch of SNP ons that are spaced apart (e. g, spaced about 25 kb to about 100 kb apart) along a chromosome can be evaluated by sequencing, and the sequencing results used to determine not only the presence of a region of gosity along a chromosome also the length ofthat LOH region.

Obtained sequencing results can be used to generate a graph that plots allele dosages along a chromosome. Allele dosage di for SNP i can be calculated from adjusted number of captured probes for two alleles (A and Bi): di = Ai/(Ai + Bi). An example ofsuch a graph is presented in Figure 2. ining whether homozygosity is due to LOH (as Opposed to homozygosity in the germline) can be performed as bed herein.

In some cases, a selection process can be used to select loci (e.g., SNP loci) to be evaluated using an assay configured to identify loci as being homozygous or heterozygous (eg, SNP array-based assays and cing-based assays). For example, any human SNP location be selected for inclusion in a SNP array—based assay or a cing-based assay configured identify loci as being homozygous or heterozygous within the genome of cells. In some cases, 0.5, 1.0, 1.5, 2.0, 2.5 million or more SNP locations present within the human genome can be evaluated to identify those SNPs that (a) are not present on the Y chromosome, (b) are not mitochondrial SNPs, (e) have a minor allele frequency of at least about five percent in Caucasians, (d) have a minor allele frequency of at least about one percent in three races other than Caucasians (e.g, Chinese, Japanese, and Yoruba), and/or (e) do not have a significant deviation from Hardy Weinberg equilibrium in any of the four races. In some cases, more than 100,000, 0, or 200,000 human SNPs can be selected that meet criteria (a) through (c). Of the human SNPs meeting criteria through (c), a group of SNPs (ag. , top 110,000 SNPs) can be selected such that the SNPs have high degree of allele frequency in Caucasians, cover the human genome in a somewhat evenly spaced manner (e.g._, at least one SNP cvcry about 25 kb to about 500 kb), and are not in linkage disequilibrium with another selected SNP for in any of the four races. In some cases, about 40, 50, 60, 70, 80, 90, 100, 110, 120, 130 thousand or more SNPs can be selected as meeting each of these 2012/07l380 criteria and included in an assay configured to identify LOH regions across a human genome. For example, between about 70,000 and about 90,000 (e.g., about 80,000) SNPs can be selected for analysis with a SNP array-based assay, and between about 45,000 and about 55,000 (e.g., about ) SNPs can be selected for analysis with a sequencing—based assay.

As described herein, a cell sample can be assessed to determine if the genome of cells ofthe sample contains an LOH ure, lacks an LOH signature, has an increased number of LOH regions that cover the whole chromosome, or lacks an sed number ofLOH regions that cover the Whole chromosome. Any appropriate type of sample can be assessed. For example, a sample containing cancer cells can be assessed to determine if the genome ofthe cancer cells contains an LOH signature, lacks an LOH signature, has an increased number ofLOH regions that cover the whole chromosome, or lacks an increased number ofLOH regions that cover the whole include, without limitation, tumor biopsy samples (e.g., breast tumor biOpsy samples), formalin- fixed, paraffin-embedded tissue samples containing cancer cells, core needle biopsies, fine needle aspirates, and samples ning cancer cells shed from a tumor (e.g., blood, urine or other bodily fluids). For formalin-fixed, paraffin-embedded tissue samples, the sample can be prepared by DNA extraction using a genomic DNA extraction kit optimized for FFPE tissue, including but not d to those described above (e.g., QuickExtractTM FFPE DNA Extraction Kit (EpicentreTM), QlAampTM DNA FFPE Tissue Kit (QiagenTMD.

In some cases, laser dissection techniques can be performed on a tissue sample minimize the number ofnon-cancer cells within a cancer cell sample to be ed. In some cases, antibody based purification methods can be used to enrich for cancer cells and/or deplete non—cancer cells. Examples of dies that could be used for cancer cell ment include, without limitation, anti-EpCAM, anti-TROP-Z, anti-c-Met, anti-Folate binding protein, anti—N-Cadherin, anti-CD318, anti-antimesencymal stem cell antigen, anti-Her2, anti-MUCl, anti-EGFR, anti- eytokeratins (e.g., cytokeratin 7, cytokeratin , etc), aveolin-l, anti-PSA, anti-CAl25, and anti—surfactant protein antibodies.

. For example, breast cancer cells, ovarian cancer cells, liver cancer cells, esophageal cancer cells, lung cancer cells, head and neck cancer cells, te cancer cells, colon, rectal, or colorectal cancer cells, and pancreatic cancer cells can be assessed to determine if the genome of the cancer cells contains an LOH signature, lacks an LOH signature, has an increased number ofLOH regions that cover the whole chromosome, or lacks an sed number of LOH regions that cover the whole chromosome. In some embodiments, the cancer cells are primary or metastatic cancer cells of n cancer, breast cancer, lung cancer or esophageal cancer.

When assessing the genome of cancer cells for the presence or absence of an LOH signature, one or more (e.g, one, two, three, four, five, six, seven, eight, nine, ten, eleven, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, or 23) pairs ofchromosomes can be ed. In some cases, the genome of cancer cells is assessed for the presence or absence of an LOH signature using one or more (e.g., one, two, three, four, five, six, seven, eight, nine, ten, eleven, , 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23) pairs ofchromosomes.

In some cases, it can be helpful to exclude certain chromosomes from this analysis.

For example, in the case of females, a pair to be assessed can include the pair of X sex chromosomes; whereas, in the case of males, a pair of any mal chromosomes (i.e., any pair other than the pair ofX and Y sex chromosomes) can be assessed. As r example, in some cases the chromosome number 17 pair may be excluded from the analysis. It has been determined that certain chromosomes carry unusually high levels of LOH in certain cancers and, thus, it can be helpful to exclude such chromosomes when analyzing samples as described herein from patients having these cancers. In some cases, the sample is from a patient having ovarian cancer, and the chromosome to be excluded is chromosome cases, the genome of cancer cells is assessed for the presence or absence of an increased number of LOH regions that cover the whole chromosome using 10 or more (e.g., 13, 16, 19, or 23) pairs of chromosomes.

Thus, a ned number of chromosomes may be analyzed to determine the total number ofIndicator LOH Regions, preferably the total number ofLOH regions of a length of greater WO 96843 PCT/U82012/071380 than 9 megabases, 10 megabases, 12 megabases, I4 ses, more preferably greater than 15 megabases. Alternatively or in addition, the sizes of all identified Indicator LOH Regions may be summed up to obtain a total length of Indicator LOH Regions.

For classification of positive LOH signature , the reference number discussed above for the total number oflndicator LOH Regions may be 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 19, 20 or greater, preferably 5, preferably 8, more preferably 9 or 10, most preferably 10. The reference number for the total (e. g., combined) length of Indicator LOH Regions may be about 75, 90, 105, 120, 130, 135, 150, 175, 200, 225, 250, 275, 300, 325 350, 375, 400, 425, 450, 475, 500 megabases or greater, preferably about 75 megabases or greater, preferably about 90 or 105 megabases or greater, more preferably about 120 or 130 megabases or r, and more preferably about 135 megabases or greater, and most preferably about 150 megabases or greater.

In some specific embodiments, the total number of LOH s of a length of greater than about 14 or 15 megabases is determined and compared to a reference number of about , 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 19, or 20. Alternatively or in addition, the total length of LOH regions of a length of greater than about 14 or 15 megabases is determined and ed to a reference number ofabout 75, 90, 105, 120, 130, 135, 150, 175, 200, 225, 250, 275, 300, 325 350, 375, 400, 425, 450, 475, or 500 megabases. it is considered “greater” if it is at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater than the reference. sely, in some embodiments the number ofLOH regions (or the combined length, or a test value or score derived from either) in a patient sample is considered “not greater” than a reference if it is not more than 2-, 3—, 4-, 5 -, 6-, 7-, 8-, 9-, or IO-fold greater than the reference while in some embodiments, it is considered “not greater” if it is not more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater than the reference.

In some ments the reference number (or length, value or score) is derived from a relevant reference tion.

Such reference populations may include patients (a) with the same cancer as the patient being tested, (b) with the same cancer sub-type, (c) with cancer having r genetic or other clinical or moleCular features, (d) who responded to a particular ent, (e) who did not respond to a particular treatment, (1) who are ntly healthy (e.g., do not have of response, prognosis (including time to cancer—specific death), etc). other synthetically, e.g, by administration of a pathway drug). Synthetic ity ches to cancer therapy are described in, e.g., O’Brien er al., Converting cancer mutations into therapeutic opportunities, EMBO MOL. MED. (2009) 1:297—299.

Examples of synthetic lethality agents include, without limitation, PARP inhibitors or double strand break repair inhibitors in homologous repair- deficient tumor cells, PARP inhibitors in PTEN—deficient tumor cells, methotrexate in MSI—I2- deficient tumor cells, etc. Examples ofPARP inhibitors include, without tion, olaparib, iniparib, and veliparib. Examples of double strand break repair tors include, without PCT/U52012/071380 limitation, KU55933 (ATM inhibitor) and NU7441 (DNA-PKcs inhibitor). Examples of information that can be used in addition to a ve LOH signature status to base a classification of situ (non-invasive), etc), disease stage, tumor or cancer grade (e.g., well, moderately, or poorly differentiated (e.g., Gleason, modified Bloom Richardson), etc), number ofprevious courses of treatment, etc.

Once classified as being likely to respond to a ular cancer treatment regimen (e.g,, a cancer treatment regimen that includes the use of a DNA damaging agent, a PARP inhibitor, radiation, or a combination thereof), the cancer patient can be treated with such a cancer treatment regimen. In some embodiments, the patients are treatment naive patients. Any appropriate method for treating the cancer at issue can be used to treat a cancer patient identified as having cancer cells having a positive LOH signature status. For example, platinum-based chemotherapy drugs or a combination of platinum-based chemotherapy drugs can be used to treat cancer as described elsewhere (see, e.g., US. Patent Nos. 3,892,790, 3,904,663, 7,759,510, 7,759,488 and 7,754,684. In some cases, anthracyclincs or a combination of anthracyclinos can be used to treat cancer described elsewhere (see, e.g., US. Patent Nos. 3,590,028, 480, 4,950,738, 6,087,340, 7,868,040, and 7,485,707. In some cases, topoisomerase I inhibitors or a ation of omerase I inhibitors can be used to treat cancer as described elsewhere (see, e.g, US. Patent Nos. 5,633,016 and 6,403,563. In some cases, PARP inhibitors or a combination of PARP inhibitors can be used to treat cancer as described elsewhere (see, e.g, US. Patent Nos. 5,177,075, 7,915,280, and 7,351,701. In some cases, radiation can be used to treat cancer as described ere (see, e.g., US. Patent No. 5,295,944). In some cases, a ation comprising different agents (e.g., a combination comprising any ofplatinum-based herapy drugs, anthracyclines, topoisomerase I PCTfU82012/071380 In some cases, patients fied as having cancer cells with a genome lacking LOH signature can be classified, based at least in part on a negative LOH signature status, as being less likely to respond to a treatment regimen that es a DNA ng agent, a PARP inhibitor, radiation, or a combination thereof. In turn, such a patient can be classified as likely to respond to a cancer treatment regimen that includes the use ofone or more cancer treatment agents not associated with HDR, such as a taxane agent (e.g, xel, paelitaxel, abraxane), a growth factor or growth factor receptor inhibitor (e. g, erlotinib, ib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab), and/or an antimetabolite agent (e.g., 5-flourouracil, methotrexate). In some embodiments, the patients are treatment naive patients. Once classified as being likely to d to a particular cancer treatment n (e.g., a cancer treatment regimen that includes the use of a cancer treatment agent not associated with HDR), the cancer t can be treated with such a cancer treatment regimen. Any appropriate method for the cancer being treated can be used to Examples of ation that can be used in addition to a negative LOH signature status to base a classification of being likely to respond to a particular cancer treatment regimen include, without limitation, previous treatment results, germline or somatic DNA mutations, gene or protein expression profiling (e.g., HER2 status, PSA levels), tumor histology (e. us cell carcinoma, papillary serous carcinoma of previous courses of treatment, etc.

Once treated for a particular period of time (e.g., between one to six months), the patient can be assessed to determine whether or not the treatment regimen has an effect. If a beneficial effect is detected, the patient can continue with the same or a similar cancer treatment n. If a minimal or no beneficial effect is ed, then adjustments to the cancer treatment regimen can be made. For example, the dose, frequency of administration, or duration of treatment can be increased. In some cases, additional anti— or a particular anti-cancer agent can be replaced with one or more different anti-cancer , The patient being treated can continue to be monitored as appropriate, and changes can be made to the cancer treatment regimen as appropriate. 2012/07l380 least one pair of human chromosomes is not a human X/Y sex chromosome pair, wherein the first length is about 1.5 or more megabases), and (b)(1) ining, based at least in part on the presence ofthe LOH signature, that the patient has a relatively good sis, or (b)(2) determining, based at least in part on the absence of the LOH signature, that the patient has a relatively poor prognosis.

Prognosis may include the t’s likelihood of survival (e.g., progression-free survival, overall survival), wherein a relatively good prognosis would e an increased likelihood of survival compared to some reference population (e. g., average patient with this patient’s cancer type/subtype, average patient not having an LOH signature, etc.) Conversely, a relatively poor prognosis in terms of survival would include a decreased likelihood of survival as compared to some reference population (e.g., average patient with this patient’s cancer type/subtype, average patient having LOH signature, etc.).

As bed , this document provides methods for assessing patients for cells (e.g., cancer cells) having a genome containing an LOH signature. In some embodiments, the patients are treatment naive patients. For example, one or more clinicians or medical professionals can determine if a patient contains cancer cells having a genome ning an LOH signature. In some cases, one or more ians or l professionals can determine if a patient contains cancer cells having a genome containing an LOH signature by obtaining a cancer cell sample from the patient and assessing the genome of cancer cells of the cancer cell sample to determine the presence or absence of an LOH signature as described herein.

VVO 2013/096843 PCT/U52012/071380 In some cases, one or more clinicians or medical professionals can obtain a cancer cell sample from a t and provide that sample to a testing laboratory having the ability to assess the genome of cancer cells of the cancer cell sample to provide an indication about the presence or absence of an LOH signature as described herein. In some embodiments, the ts are treatment For example, a testing laboratory, after assessing the genome of cancer cells for presence or absence of an LOH signature as described herein, can provide a clinician or medical professional with, or access to, a written, onic, or oral report or medical record that provides an indication about the presence or absence of an LOH signature for a particular patient being assessed. Such a written, electronic, or oral report or medical record can allow the one or more clinicians or medical or can be based at least in part on a determination that a ular patient being assessed contains cancer cells having a genome containing an LOH signature. For example, a patient ined have cancer cells whose genome contains the presence of an LOH signature can be diagnosed as likely to be deficient in HDR based on the combination ofa positive LOH signature status and deficient status in one or more tumor suppressor genes (e.g., BRCAl/2, RADS 1C), a family history ofcancer, or the presence of oral risk factors (6. g. , g).

In some cases, a clinician or medical professional or group of clinicians or medical professionals can diagnose a patient ined to have cancer cells whose genome contains the PCT/U82012/071380 presence ofan LOH signature as having cancer cells likely to contain genetic mutations in one or more genes in the HDR pathway. In some embodiments, the patients are treatment na'ive patients.

Such a diagnosis can be based solely on a determination that a particular patient being assessed ns cancer cells having a genome containing an LOH signature or can be based at least in part on a determination that a particular patient being assessed contains cancer cells having a genome containing an LOH signature. For example, a patient ined to have cancer cells whose genome contains the presence of an LOH signature can be diagnosed as having cancer cells likely to contain genetic ons in one or more genes in the HDR pathway based on the combination of a positive LOH positive status and a family history of cancer, or the presence of behavioral risk factors (e.g., smoking). cancer cells having a genome containing an LOH signature or can be based at least in part on a determination that a particular t being assessed ns cancer cells having a genome ning an LOH signature. For example, a patient determined to have cancer cells whose genome contains the presence of an LOH signature can be diagnosed as being likely to respond to a particular cancer treatment regimen based on the combination of a ve LOH signature status and nt status in one or more tumor suppressor genes (e.g., BRCAl/Z, RADSI), a family history of cancer, or the presence of behavioral risk factors (e.g, smoking). As bed herein, a determined to have cancer cells whose genome ns the presence of an LOH signature can be diagnosed as likely to respond to a cancer treatment regimen that includes the use of a platinum- based chemotherapy drug such as eisplatin, carboplatin, oxaliplatin, or picoplatin, an anthracycline such as epirubicin or doxorubicin, a topoisomerase I inhibitor such as hecin, topotecan, irinotecan, a PARP inhibitor, ion, a combination thereof, or a combination of any of the preceding with another anti~caneer agent. In some embodiments, the patients are treatment naive patients.

Once a clinician or medical professional or group of clinicians or medical professionals determines that a particular patient being assessed contains cancer cells having a 2012/071380 . the clinician or medical professional (or group) can classify that patient as having cancer cells whose genome contains an absence of an LOH signature.

In some embodiments, the patients are treatment naive patients. In some cases, a clinician or medical professional or group of clinicians or medical professionals can diagnose a patient determined have cancer cells containing a genome that lacks the ce of an LOH signature as having cancer cells likely to have functional HDR.

In some cases, a clinician or medical professional or group of clinicians or medical professionals can diagnose a patient determined to have cancer cells containing a genome that lacks the presence ofan LOH ure as having cancer cells that do not likely contain genetic mutations in one or more genes in the HDR pathway. In some cases, a ian medical sional or group of clinicians or medical professionals can diagnose a patient determined to have cancer cells containing a genome that lacks the ce of an LOH signature or contains an sed number of LOH regions that cover the whole chromosome as having cancer cells that are less likely to respond to a um-based chemotherapy drug such as cisplatin, latin, oxalaplatin, or picoplatin, an anthraeycline such as epirubineiu or doxorubicin, t0poisomerase 1 inhibitor such as campothecin, topotecan, or irinotecan, a PARP inhibitor, or radiation and/0r more likely to respond to a cancer treatment regimen that includes the use of a cancer treatment agent not associated with HDR such as one or more taxane agents, growth factor growth factor receptor inhibitors, anti-metabolite agents, etc. In some embodiments, the patients treatment naive patients.

ZOIZ/07l380 genomic nucleic acid sample obtained from cancer cells ed from the patient and (b) performing an analysis (e.g., a SNP array-based assay or a sequencing-based assay) using the ed material to detect the presence or absence of an LOH signature or the presence or absence of an increased number of LOH regions that cover the whole chromosome as described herein. In some cases, one or more laboratory cians or laboratory professionals can receive a sample to be analyzed (e. g., a cancer cell sample ed from the patient, a genomic nucleic acid sample obtained from cancer cells obtained from the patient, or an enriched and/or amplified genomic nucleic acid sample obtained from cancer cells obtained from the patient) directly or indirectly from a clinician or medical professional. In some embodiments, the patients are treatment naive patients.

Once a laboratory technician or tory professional or group of laboratory technicians or laboratory professionals detects the presence of an LOH signature as described herein, the laboratory technician or laboratory professional (or group) can identify the t whose cancer cells were detected as having an LOH signature as having cancer cells with a positive LOH signature status. For example, one or more laboratory technicians or laboratory professionals can identify a patient having cancer cells that were detected to have an LOH signature as having cancer cells with a positive LOH signature status by associating that positive LOH ure status or the result (or results or a summary of s) of the performed diagnostic analysis with the corresponding patient’s name, medical record, symbolic/numerical identifier, or a combination f. In some cases, a laboratory technician or tory professional or group of laboratory technicians laboratory professionals can identify a patient having cancer cells that were detected to have an LOH signature as having cancer cells potentially deficient in HDR by associating the positive signature status, the ially deficient in HDR status, or the result (or results or a summary of s) of the performed diagnostic analysis with the correSponding patient’s name, medical record, 2012/07l380 In some cases, a laboratory technician or laboratory professional or group of laboratory technicians or laboratory professionals can identify a patient having cancer cells that were detected to have an LOH signature as having cancer cells potentially containing a genetic mutation in one or more genes in the HDR pathway by associating the positive LOH signature status, the potential presence of a genetic mutation in one or more genes in the HDR pathway, or the result (or results or a summary of results) of the performed stic is with the ponding patient’s name, medical record, symbolic/numerical identifier, or a combination thereof. Such identification can be based solely on detecting the presence of an LOH signature or can be based at least in part on detecting the ce of an LOH signature. For example, a laboratory technician tory professional can identify a patient having cancer cells that were detected to have an LOH signature as having cancer cells potentially containing a genetic mutation in one or more genes in the HDR y based on a combination of a positive LOH signature status and the results of other genetic and biochemical tests performed at the testing laboratory. In some embodiments, the patients are ent naive patients.

In some cases, a laboratory technician or laboratory professional or group of laboratory cians or laboratory sionals can identify a patient having cancer cells that were detected to have an LOH signature as having cancer cells likely to respond to a particular cancer treatment regimen by associating the positive LOH signature status, a potentially deficient HDR status, a potential presence of a nt status in one or more genes in the HDR pathway, or the result (or s or a summary ofresults) of the performed diagnostic analysis with the patients are treatment naive patients.

Once a laboratory technician or laboratory professional or group of laboratory technicians or laboratory professionals detects the absence of an LOH signature, the laboratory technician or laboratory professional (or group) can identify the patient whose cancer cells were PCT/U52012/071380 detected as lacking an LOH signature as having cancer cells with a negative LOH signature status.

For example, one or more laboratory technicians or laboratory sionals can identify a patient having cancer cells that were detected to lack an LOH ure as having cancer cells with a negative LOH signature status by associating that negative LOH ure status or the result results or a summary of results) of the performed diagnostic analysis with the corresponding patient’s name, medical record, symbolic/numerical identifier, or a combination thereof. In some cases, a laboratory technician or laboratory professional or group of laboratory technicians laboratory professionals can identify a patient having cancer cells that were detected to lack an LOH signature as having cancer cells with potentially intact HDR by ating the negative LOH signature status, the potentially intact HDR status, or the result (or s or a summary of results) of the performed diagnostic analysis with the corresponding patient’s name, medical record, symbolic/numerical identifier, or a combination thereof. In some embodiments, the patients treatment naive patients.

In some cases, a laboratory technician or laboratory professional or group of laboratory technicians or laboratory professionals can identify a patient having cancer cells that were detected to lack an LOH signature as having cancer cells with ially intact genes of the HDR pathway by associating the negative LOH ure status, the potential absence of c mutations in genes of the HDR pathway, or the result (or results or a summary of results) of the performed stic analysis with the corresponding patient’s name, medical record, symbolic/numerical identifier, or a combination thereof. In some embodiments, the ts are treatment naive patients.

] In some cases, a laboratory technician or laboratory sional or group of laboratory technicians or laboratory professionals can identify a t having cancer cells that were detected to lack an LOH signature as having cancer cells as less likely to reSpond to one particular treatment (e.g., a platinum-based chemotherapy drug such as cisplatin, carboplatin, oxalaplatin, picoplatin, an anthracyeline such as epirubincin 01' bicin, a topoisomerase I inhibitor such as PCT/U52012/071380 symbolic/numerical identifier, or a combination f In some embodiments, the patients treatment na'ive patients.

For example, one or more laboratory technicians or laboratory professionals can identify a patient having cancer cells that were detected to have an sed number ofLOH regions that cover the whole chromosome as likely having cancer cells with an intact BRCA1 and BRCA2 status by aS‘ symbolic/numerical identifier, or a combination thereof. In some ments, the patients treatment naive patients.

The results of any analyses according to the invention will ofien researchers) in a transmittablc form that can be communicated or transmitted to any of the above parties. Such a form can vary and can be le or intangible. The results can be embodied in descriptive statements, diagrams, photographs, charts, images or any other Visual forms. For example, graphs or diagrams showing genotype or LOH (or HRD status) information can be used in explaining the results. The ents and Visual forms can be recorded on a tangible medium such [001 10] Thus, the information and data on a test result can be produced anywhere in the world and transmitted to a different location. As an rative example, when an assay is conducted outside the United States, the information and data on a test result may be generated, east in a transmittable form as bed above, and then imported into the United States. Accordingly, PCT/U52012/071380 the present invention also encompasses a method for producing a ittable form of information determining an LOH signature according to methods of the present ion; and (2) embodying the result of the determining step in a transmittable form. The ittable form is a product of such a method.

Several ments of the invention described herein involve a step of correlating an LOH ure according to the present invention (e.g., the total number of LOH length but shorter than the length of the whole chromosome containing the LOH region, wherein said at least one pair ofhuman chromosomes is not a human X/Y sex chromosome pair, wherein said first length is about 1.5 or more megabases) to a particular clinical feature (e.g, an increased likelihood of a deficiency in the BRCA1 or BRCAZ gene; an increased likelihood of deficiency; an increased likelihood of response to a treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor; etc.) if the number is greater than some reference (or ally to another feature if the number is less than some reference). Throughout this document, wherever such an embodiment is described, another embodiment ofthe invention may involve, in addition to or instead of a correlating step, one or both ofthe following steps: (a) concluding that the patient has the clinical feature based at least in part on the presence or absence of the LOH signature; or (b) communicating that the patient has the clinical e based at least in part on the presence or absence of the LOH signature.

By way of illustration, but not limitation, one ment described in this document is a method of predicting a cancer patient’s response to a cancer treatment regimen sing a DNA damaging agent, an anthraeycline, a topoisomerase I inhibitor, radiation, and/or PARP inhibitor, said method comprising: (1) determining, in a cancer cell from said cancer patient, the number ofLOH regions in at least one pair of human chromosomes of a cancer cell of said cancer patient that are longer than a first length but shorter than the length ofthe whole chromosome ning the LOH region, wherein said at least one pair of human chromosomes is not a human X’Y sex chromosome pair, wherein said first length is about 1.5 or more megabases; and (2) correlating said total number that is r than a reference number with an increased hood that said cancer patient will respond to said cancer treatment n.

According to the ing paragraph, this description of this embodiment is understood to include a description of two related ments, z'. ., comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor, said method sing: (1) determining, in a cancer cell from said cancer patient, the number of LOH regions in at least one pair of human chromosomes of a cancer cell of said cancer t that are longer than a first length but shorter than the length ofthe Whole chromosome containing the LOH region, wherein said at least one pair ofhuman chromosomes is not a human X/Y sex chromosome pair, wherein said first length is about 1.5 or more megabases; and (2)(a) concluding that said patient has an increased likelihood that said cancer patient will respond to said cancer ent regimen based at least in part on a total number that is greater than a nce number; or (2)(b) communicating that said patient has an increased likelihood that said cancer patient will reSpond to said cancer treatment regimen based at least in part on a total number that greater than a reference number.

In each ment described in this document involving ating a particular assay or analysis output (e.g., total number of LOH regions greater than a reference number, etc.) to some likelihood (e. g, increased, not increased, decreased, etc.) of some clinical feature (e. g., response to a particular treatment, cancer-specific death, etc), or additionally alternatively concluding or icating such clinical feature based at least in part on such particular assay or analysis output, such correlating, below 5%, 10%, 15%, 20%, 25%, %, 35%, 40%, 45%, or 50%. In some embodiments “intermediate risk” is any percentage ility above 5%, 10%, 15%, 20%, %, 30%, 35%, 40%, 45%, or 50% and below 15%, 20%, %, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%. In some embodiments “high risk” is any percentage probability above 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or 99%.

As used herein, “communicating” a particular piece of information means to make such information known to another person or transfer such ation to a thing (eg, a computer). In some methods ofthe invention, a patient’s prognosis or likelihood ofresponse to particular treatment is communicated. In some embodiments, the information used to arrive at such PCT/U52012/07l380 a prognosis or response prediction (e.g, LOH signature ing to the present ion, etc.) is communicated. This communication may be auditory (e.g., verbal), visual (e.g., n), electronic (e.g., data transferred from one computer system to another), etc. In some embodiments, communicating a cancer classification (e.g., prognosis, likelihood of response, appropriate treatment, etc.) comprises generating a report that communicates the cancer classification.

In some embodiments the report is a paper report, an auditory report, or an electronic record. In some embodiments the report is displayed and/or stored on a ing device (e.g., handheld device, desktop er, smart device, website, etc). In some embodiments the cancer classification is communicated to a physician (cg, a report communicating the classification is provided to the ian). In user to access such information (e.g, by viewing the information as displayed from the server, by downloading the information in the form of one or more files transferred from the server to the intermediary or end- user’s device, etc). length but shorter than the length of the whole chromosome containing the LOH region indicates that the cancer cells have the LOH signature, wherein the at least one pair of human chromosomes PCT/USZOIZ/07l380 one or more reference values derived from the number of said LOH regions in a reference population (e.g., mean, median, tereiles, quartiles, quintiles, etc.); and (4)(a) administering to said patient an ancer drug, or recommending or prescribing or initiating a treatment regimen comprising chemotherapy and/or a synthetic lethality agent based at least in part on said comparing step revealing that the test value is greater (e. g, at least 2-, 3—, 4-, 5—, 6-, 7—, 8-, 9-, or 10-fold greater; at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater) than at least one said reference value; or (4)(b) recommending or prescribing or initiating a treatment regimen not comprising chemotherapy and/or a synthetic lethality agent based at least in part on said ing step revealing that the test value is not greater (e.g., not more than 2-, 3—, 4, 5—, 6-, 7-, 8—, 9-, or 10—fold greater; not more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater) than at least one said reference value. The invention encompasses, s mutandis, ponding embodiments where the test value or score is used to determine the patient’s prognosis, the patient’s hood of Figure 15 shows an exemplary process by which a computing system (or er m (e.g., software) containing computer-executable instructions) can identify LOH loci or regions from genotype data as described herein. If the ed ratio of the signals of two alleles, A and B, is two to one, there are two possibilities. The first possibility is that cancer cells have LOH with deletion of allele B in a sample with 50% contamination with normal cells. The second possibility is that there is no LOH but allele A is ated in a sample with no contamination with normal cells. The process begins at box 1500, where the following data collected by the computing system; (1) sample-specific normalized signal intensities for both alleles of each locus and (2) assay-specific (specific for different SNP arrays and for sequence based approach) set of parameters defined based on analysis of large number of s with known ASCN s. As described , any apprOpriate assay such as a SNP array-based assay or sequencing-based assay can be used to assess loci along a chromosome for homozygosity or heterozygosity. In some cases, a system including a signal detector and a computer can be used to collect data (e.g., fluorescent signals or sequencing results) regarding the homozygous heterozygous nature of the plurality of loci (cg. , sample—specific normalized signal intensities for 2012/071380 each locus (e.g., each SNP). ASCNs are the numbers of copies of both paternal and maternal alleles.

At box 153 0, a likelihood function is used to determine whether a homozygous locus or region of homozygous loci is due to LOH. This can be conceptually analogous to a usly described algorithm designed to reconstruct total copy number (rather than ASCN) at each locus (e. g., SNP).

See International Application No.

PCT/US$201 1/026098 to Abkevich et a]. The likelihood function can be maximized over ASCN of all loci, level of contamination with benign , total copy number averaged over the whole genome, and sample specific noise level. At box 1540, an LOH region is determined as a stretch of SNPs with one of the ASCNs (paternal or al) being zero.

In some embodiments, the computer process fiirther comprises a step of inquiring or determining whether a patient is treatment naive.

Figure 3 shows an exemplary process by which a ing system can regarding the homozygous or heterozygous nature of a plurality of loci along a chromosome is collected by the computing system. As described , any appropriate assay such as a SNP array- based assay or cing-based assay can be used to assess loci along a chromosome for gosity or heterozygosity. In some cases, a system including a signal detector and a computer can be used to collect data (e. g., fluorescent s or sequencing results) regarding the homozygous or heterozygous nature of the ity of loci.

At box 310, data regarding the homozygous heterozygous nature of a plurality of loci as well as the location or spatial relationship of each locus is assessed by the computing system to determine the length of any LOH regions present along a chromosome. At box 320, data regarding the number ofLOH s detected and the length each detected LOH region is assessed by the computing system to determine the number of LOH regions that have a length (a) greater than or equal to a preset number of Mb (e.g., 15 Mb) and less than the entire length of the chromosome containing that LOH region.

Alternatively the computing system can determine the total or combined LOH length as described above. At box 330, the computing system formats an output ing an indication of the presence or absence of an LOH signature. Once formatted, the computing system can present the output to a user (eg, tory technician, clinician, or medical professional). As described herein, the presence or absence of an LOH signature can be used to provide an indication about a patient’s likely HDR status, an indication about the likely presence or absence of c mutations in genes of the HDR pathway, and/or an indication about possible cancer treatment regimens. “’0 2013/096843 PCT/U52012/071380 Figure 4 is a diagram of an example of a computer device 1400 and a mobile computer device 1450, which may be used with the techniques bed herein. Computing device 1400 is intended to represent various forms of digital computers, such as laptops, desktOps, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Computing device 1450 is intended to represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smart phones, and other similar ing devices.

The ents shown here, their connections and relationships, and their fiinctions, are meant to be exemplary only. and are not meant to limit implementations of the ions described and/or claimed in this document.

Computing device 1400 includes a processor 1402, memory 1404, a storage device 1406, a high-speed interface 1408 connecting to memory 1404 and high-speed expansion ports 1410, and a low speed interface 1415 connecting to low speed bus 1414 and e device 1406. Each of the components 1402, 1404, 1406, 1408, 1410, and 1415, are interconnected using various busses, and may be mounted on a common motherboard or in other manners as appropriate. used, as appropriate, along with multiple memories and types ofmemory. Also, le computing devices 1400 may be connected, with each device providing portions of the necessary operations (e.g., as a server bank, a group of blade servers, or a multi-processor system).

The memory 1404 stores ation within the computing device 1400. In one implementation, the memory 1404 is a volatile memory unit or units. In another implementation, the memory 1404 is a non-volatile memory unit or units. The memory 1404 may also be another form of computer—readable medium, such as a magnetic or optical disk.

The storage device 1406 is capable of providing mass storage for the computing device 1400. In one implementation, the e device 1406 may be or contain a computer-readable medium, such as a floppy disk device, a hard disk device, an optical disk , or a tape , a flash memory or other similar solid state memory , or an array of dcviccs, including devices in a storage area network or other configurations. A computer program product can be tangibly embodied in an ation carrier. The computer program product may also PCT/U52012/071380 contain instructions that, when executed, perform one or more methods, such as those bed herein. The information r is a computer— or machine-readable medium, such as the memory 1404, the storage device 1406, memory on processor 1402, or a propagated signal.

] The high speed controller 1408 manages bandwidth-intensive operations for the computing device 1400, while the low speed controller 1415 manages lower bandwidth-intensive operations. Such allocation of functions is exemplary only. In one implementation, the high-speed controller 1408 is coupled to memory 1404, y 1416 (e.g., through a cs processor or accelerator), and to peed expansion ports 1410, which may accept various expansion cards (not shown). In the implementation, low-speed controller 1415 is coupled to storage device 1406 and low-speed expansion port 1414. The low-speed expansion port, which may include various communication ports (e.g., USB, Bluetooth, Ethernet, or wireless Ethernet) may be d to one or The computing device 1400 may be implemented in a number of different forms, as shown in the figure. For example, it may be implemented as a standard server 1420, multiple times in a group of such servers. It may also be ented as part of a rack server system 1424. In addition, it may be implemented in a personal computer such as a laptop computer 1422. Alternatively, components from computing device 1400 may be combined with other components in a mobile device (not , such as device 1450. Each of such devices may contain one or more of computing device 1400, 1450, and an entire system may be made up ofmultiple computing devices 1400, 1450 communicating with each other.

Computing device 1450 includes a processor 1452, memory 1464, an input/output device such as a display 1454, a communication interface 1466, and a eiver 1468, among other components (e.g., a scanner, an Optical reader, a fluorescent signal detector). The device 1450 may also be ed with a storage device, such as a microdrive or other device, to provide additional storage. Each of the components 1450, 1452, 1464, 1454, 1466, and 1468, interconnected using various buses, and several of the components may be mounted on a common motherboard or in other manners as appropriate.

] The sor 1452 can execute instructions within the computing device 1450. including instructions stored in the memory 1464. The processor may be implemented as a WO 96843 PCT/U52012/071380 chipset of chips that include separate and multiple analog and digital processors. The processor may provide, for example, for coordination of the other components of the device 1450, such as control of user interfaces, applications run by device 1450, and ss communication by device 145 0.

Processor 1452 may communicate with a user through control interface 1458 and display interface 1456 coupled to a display 1454. The display 1454 may be, for example, 21 TFT LCD (Thin-Film-Transistor Liquid Crystal Display) or an OLED (Organic Light Emitting Diode) y, or other appropriate display teclmology. The display ace 1456 may comprise appropriate circuitry for driving the display 1454 to present graphical and other information to a user. The control interface 1458 may receive commands from a user and convert them submission to the processor 1452. In addition, an external interface 1462 may be provide in ication with sor 1452, so as to enable near area communication of device 1450 with other devices. External interface 1462 may provide, for example, for wired communication in some implementations, or for ss communication in other implementations, and multiple interfaces may also be used.

The memory 1464 stores information within the ing device 1450. memory 1464 can be implemented as one or more of a computer-readable medium or media, a volatile memory unit or units, or a latile memory unit or units. Expansion memory 1474 may also be provided and connected to device 1450 through expansion interface 1472, which may include, for example, a SIMM (Single In Line Memory Module) card interface. Such expansion memory 1474 may provide extra storage space for device 1450, or may also store applications other information for device 1450.

For e, expansion memory 1474 may include instructions applications may be provided via the SIMM cards, along with additional information, such placing identifying ation on the SIMM card in a non-hackable manner.

PCT/U82012/071380 memory on processor 1452, or a propagated signal that may be received, for example, over transceiver 1468 or external interface 1462. transceiver 1468. In addition, short-range communication may occur, such as using a Bluetooth, WiFi, or other such transceiver (not shown). In addition, GPS (Global Positioning System) receiver module 1470 may provide additional navigation- and location-related wireless data to device 1450, which may be used as appropriate by applications running on device 1450.

Device 1450 may also communicate audibly using audio codec 1460, which codec 1460 may se generate audible sound for a user, such as through a speaker, e.g., in a handset of device 1450. Such sound may include sound from voice telephone calls, may include recorded sound (e.g., voice messages, music files, etc.) and may also include sound generated by applications operating on device 1450.

The computing device 1450 may be implemented in a number of different forms, as shown in the figure. For example, it may be implemented as a cellular telephone 1480. may also be implemented as part of a hone 1482, personal l assistant, or other similar mobile .

Various implementations of the systems and techniques described herein be realized in digital electronic try, integrated circuitry, specially ed ASICs cation specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof.

These various implementations can include implementation in one or more computer programs that are executable and/or interpretable on a mmable system including at least one programmable processor, which may be special or general purpose, d to e data and instructions from, and to it data and instructions to, a storage system, at least one input device, and at least one output device.

PCTfUS2012/071380 assembly/machine language. As used herein, the terms “machine-readable medium” and “computer- readable medium” refer to any er program product, apparatus and/or device (e.g., magnetic discs, optical disks, memory, and Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable sor, ing a machine-readable medium that feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described herein can be implemented in computing system that includes a back end component (e.g, as a data server), or that includes middlcwarc component (e. g., an application ), or that includes a front end component (e. g, a client computer having a graphical user interface or a Web r through which a user can interact with an implementation of the systems and techniques described ), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (“LAN”), a wide area network (“WAN”), and the Internet.

The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by Virtue of computer programs running on the respective ers and having a client—server relationship to each other.

WO 96843 PCT/U52012/07l380 e, a sample analyzer can produce signals that are capable of being interpreted in a manner that identifies the homozygous or heterozygous nature of loci along a some. In some cases, a sample analyzer can be configured to carry out one or more steps of a SNP array-based assay or sequencing—based assay and can be configured to produce and/or capture signals from such assays.

In some cases, a computing system provided herein can be configured to include a ing device. In such cases, the computing device can be configured to receive signals from a sample analyzer. The computing device can include computer—executable instructions or a computer program (e.g., software) containing computer-executable instructions for carrying out one or more of the methods or steps described herein.

In some cases, such er-executable instructions can instruct a computing device to analyze signals from a sample analyzer, from r computing device, from a SNP array-based assay, or from a sequencing-based assay. The analysis of such signals can be carried out to determine genotypes, homozygosity at certain loci, regions of homozygosity, the number ofLOH regions, to ine the size of LOH regions, to ine number ofLOH regions having a particular size or range of sizes, to determine whether or not a sample is ve for an LOH signature, to determine the number of tor LOH Regions in at least one pair of human somes, to determine a likelihood of a deficiency in BRCAI and/or BRCA2 genes, to determine a likelihood of a deficiency in HDR, to determine a likelihood that cancer patient will respond to a ular cancer treatment regimen (e. g. , a regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, a PARP inhibitor, or a combination thereof), or to determine a combination of these items.

In some cases, a computing system provided herein can e computer- executable instructions or a computer program (e.g., software) containing computer—executable instructions for formatting an output providing an indication about the number of LOH regions, the size ofLOH regions, the number of LOH regions having a particular size or range of sizes, whether or not a sample is positive for an LOH signature, the number of Indicator LOH Regions in at least one pair of human chromosomes, a likelihood of a deficiency in BRCAl and/or BRCA2 genes, a likelihood of a deficiency in HDR, a likelihood that a cancer t will respond to a particular cancer treatment regimen (e.g., a regimen that includes a DNA damaging agent, an anthracycline, a topoisomcrasc I inhibitor, radiation, a PARP inhibitor, or a combination thereof), or a combination ofthese items. In some cases, a computing system ed herein can include computer-executable instructions or a computer program (6.g. , software) containing er—executable instructions for wo 2013/096843 PCT/U82012/071380 determining a d cancer treatment regimen for a particular patient based at least in part on the presence or absence of an LOH signature or on the number of Indicator LOH Regions.

In some cases, a computing system provided herein can include a pre- processing device configured to process a sample (e.g, cancer cells) such that a SNP array-based assay or sequencing-based assay can be performed.

Examples of pre-processing devices include, without tion, devices configured to enrich cell populations for cancer cells as opposed to non- cancer cells, devices configured to lyse cells and/or extract genomic nucleic acid, and devices configured to enrich a sample for particular genomic DNA fragments.

This document also provides kits for assessing s (e.g., cancer cells) as described herein. For example, this document provides kits for assessing cancer cells for the presence of an LOH signature or to determine the number of Indicator LOH Regions in at least one pair of human chromosomes. A kit provided herein can e either SNP probes (eg, an array of SNP probes for ng out a SNP array—based assay described herein) or primers (e.g., primers designed for sequencing SNP regions via a sequencing-based assay) in combination with a er program product ning computer—executable instructions for g out one or more of the methods or steps described herein (e. g., computer-executable instructions for determining number ofLOH regions having a particular size or range of . In some cases, a kit provided ingredients for performing a SNP array—based assay or a sequencing-based assay. Examples of such other ingredients include, without limitation, s, sequencing nucleotides, enzymes (e.g, polymerases), etc. This document also provides the use of any apprOpriate number of the materials provided herein in the manufacture of a kit for carrying out one or more of the methods or steps described herein. For example, this nt provides the use of a collection of SNP probes (e. g., a collection of 10,000 to 0 SNP probes) and a computer program product provided herein in the manufacture of a kit for assessing cancer cells for the presence of an LOH signature. As another PCT/U82012/071380 ] The invention will be fiirther described in the following examples, which not limit the scope of the invention described in the claims.

EXAMPLES Example 1 ~ Assessing LOH regions and HDR Two sets of tumors were used from advanced ovarian cancer ts. The first set of 94 tumors (training set) was used to derive a candidate signature, and the second set of 40 tumors (validation set) was used to validate the signature. All coding s of BRCA1 BRCA2 genes were sequenced to detect germ line and somatic mutations. Levels ofBRCA1 and BRCA2 mRNA expression were measured, and Affymetrix SNP microarrays were performed.

] A computer program was used to reconstruct LOH signature status based on allele intensities derived from the microarray data. An algorithm was developed and implemented One point of the algorithm was to first reconstruct allele c copy numbers (ASCN) at each locus (eg, SNP). ASCNs are the numbers of copies of both paternal and al alleles. An LOH region was then determined as a stretch of SNPs with one of the ASCNs (paternal or maternal) being zero. The algorithm was based on zing a likelihood fimction number (rather than ASCN) at each locus (e.g., SNP). See International Application No.

PCT/U820] 1/026098 to Abkevich er al. The likelihood function was maximized over ASCN of all loci, level of contamination with benign tissue, total copy number averaged over the whole genome, and sample specific noise level.

The input data for the algorithm included (1) sample-specific normalized signal intensities for both allele of each locus and (2) assay—specific fic for different SNP arrays and for ce based approach) set ofparameters defined based on analysis of large number of s with known ASCN profiles.

Tumors were defined as being HDR deficient for the purpose of this analysis if they either had one or more deleterious mutations in BRCAl and/or BRCA2 genes or if they had low expression ofBRCA1 mRNA.

The rest of the tumors were defined as likely HDR non-deficient for the purpose of this analysis.

PCTfU52012/071380 The distn'bution of the lengths of LOH regions was investigated (Figure Three categories ofLOH regions were used: (1) LOH affecting a whole chromosome; (2) large LOH regions (greater than about 15 Mb), which typically affect a part of a chromosomal arm or the whole chromosomal arm; and (3) multiple short LOH regions (less than about lSMb).

Second, using the training set only, the number ofLOH regions of one ofthese three categories was assessed for possible correlations with HDR ncy. It was discovered that (I) the number of short LOH regions did not significantly correlate with HDR deficiency (p>0.05); (2) LOH covering an entire chromosome correlated weakly with HDR deficiency (p=0.0011); and (3) the number of large LOH regions correlated significantly with HDR deficiency (p=l.9e-8). More specifically, it was discovered that all HDR deficient tumors had a high number of large LOH s (e. g., nine or more), while the majority of tumors likely to be HDR non-deficient had a small number of large LOH regions (Figures 6-8). It was le that tumors likely to be HDR non-deficient were in fact HDR deficient due to other genetic tions, excluding BRCA1 and BRCA2 mutations and low mRNA expression. In addition to the number of large LOH regions, the total length of these regions also correlated significantly with HDR deficiency.

These results were confirmed with the validation set: (1) the number of short LOH regions did not significantly ate with HDR deficiency 5); (2) LOH covering an entire chromosome correlated weakly with HDR deficiency (p=0.05); and (3) the number of large LOH regions correlated significantly with HDR deficiency e-6).

The 134 tumors were d from combined training and validation data sets into three groups: (1) BRCA deficient if they either had one or more deleterious mutations in BRCAI and/or BRCA2 genes or if they had low sion of BRCA1 mRNA; (2) HDR deficient / BRCA intact if they have 9 or more large LOH regions (greater than Mb but less than the length ofthe entire chromosome); (3) HDR intact if they have less than 9 large LOH regions (greater than Mb but less than the length of the entire some). Results of this analysis are presented in Figure 9. It shows a high frequency of BRCA deficiency as well as HDR deficiency that is not due to BRCA deficiency among ovarian tumors.

PCT/U82012/071380 among breast and esophagus cancer cell lines. No HDR deficiency was observed among colon cancer cell lines. Validating the previous findings for ovarian tumors, all BRCA deficient cell lines were found to be HDR nt as well.

] Figure 11 shows the distribution of large LOH regions (greater than 15 Mb but less than the length of the entire chromosome) for publicly available lung tumor data set regions (greater than 15 Mb but less than the length of the entire chromosome) is presented for several tumors and cell lines. This frequency is as high as 50% among ovarian tumors and was not observed at all among brain and colon cell lines. Thus it appears that HDR deficiency plays an important role for the majority of cancers.

Example 2 — Chemo Toxicity ses In preparation of chemo toxicity response experiments, all cell lines were grown at 37°C plus 5% C ‘ trypsinized (Invitrogen ation Cat # 25200—056), counted, and seeded in Advanced RPMI 1640 rogen Corporation Cat # 020), 3% PBS, 1 % peni cillin/streptomycin (Invitrogen Corporation Cat # 15140—122) at 2500 cells or 5000 cells in 100 uL media per well from columns 2— 12 of l polystyrene mieroplates with clear bottom (Perkin Ehner Cat #6005181), leaving column 1 with 100 uL per well a only. The cell-seeded plates were then incubated at 37°C plus 5% C01 ovemight.

Two different final drug concentration working stocks were prepared. In cases where 100% DMSO was required for drug lity, Advanced RPM] 1640 was used as the diluent for the highest concentration.

Advanced RPMI 1640 plus a predetermined amount ofDMSO equal to the total DMSO in the high concentration working stock was used for the low concentration, with a maximum of 60% DMSO used for the lowest concentration.

This was done to keep the DMSO concentrations equal in every well and prevent non-Specific cell death as a result of DMSO.

The lower ofthe two drug concentrations was placed in a 96-well, thin-wall PCR cycle plate (Robbins Scientific Cat # 1055—00—0) in rows A—D, column 12, while the higher concentration was PCT/U82012/071380 placed in rows E-H, column 12, of the same plate. Serial dilutions of 1:2 or 1:3 were performed in a descending manner from column 12 to 3, leaving columns 1 and 2 to be used for no cell/no drug and no drug controls. This allowed for quadruplet data points for each drug concentration. Once dilutions were complete, 5 uL was transferred from the dilution plate to the ponding well of the seeded cell plate. Plates receiving drugs were then incubated at 37°C plus 5% C02 for either 3 days or 6 days.

Following a 3-day or 6-day dose regimen, ATPlite assays (Perkin Elmer cat # 6016941) were run on each well of each plate according to the ATPLite Assay ol. The luminescence was then read on a FUSION machine and saved as a .CSV file. For each cell-line and drug combination, the four replicates of the no-drug l were ed and divided by 100 to create a lization factor” used to calculate a normalized percent survival. The normalized percent survival for the no-drug controls was 100%. The four replicates of the cell-plus-drug wells were averaged and divided by the normalization factor for each drug concentration. The percent al for each chug concentration, starting with a tration equal to 0, was used to calculate an IC50 using proprietary software.

Figure 13 shows se to herapy for breast and ovarian cancer cell lines. On y-axis are indicated values ofLog10(IC50) for different chemotherapy drugs (camptothecin, as well as averaged results for platinum nds (oxaliplatin, cisplatin, and carboplatin) anthracyclines (doxorubicin and epirubicin)) when exposed to 29 breast cancer cell lines as well as Log10(IC50) ofpaclitaxel when exposed to 27 ovarian cancer cell lines. On the x—axis the number of large LOH regions longer than 15 Mb and r than the entire chromosome are indicated for these cell lines. The dashed lines place a threshold number at nine.

Figure 14 is a version of a graph from Figure 13 that indicates specificity and sensitivity among responders and non-reSponders to treatment with platinum compounds (oxaliplatin, tin, and carboplatin) when exposed to 29 breast cancer cell lines. The dashed lines place a threshold number of large LOH regions longer than 15 Mb and shorter than the entire chromosome at nine. The solid line divides cell lines into responders and non—responders.

Example 3 — Further Validation of HR Deficiency Assay Materials and Methods Ovarian tumor Samples PCT/U52012/071380 unselected ovarian cancer s. 2: 53 high grade serous ovarian tumors. 3: Publicly available data from 435 serous n cancer samples for which complete information was available were downloaded from The Cancer Genome Atlas (TCGA) Network web site on October 31, 2011. All l— First cohort Second cohort Third cohort Total Number of Patients 152 53 435 Age at diagnosis J— Range 37 - 88 38 - 77 30 - 89 Median 59 56 59 Unknown 4 (2.6%) 0 0 Follow—up time T 4— Range 20 - 5570 213 - 3294 8 - 5480 Median 1127 701 874 Unknown L _L 5 (3.2%) 0 2 ) m 1— 9 (5.9%) 0 6 (1.38%) 14 (9.2%) 0 21 (4.83%) 107 (70.4%) 46 (86.8%) 338 (77.70%) 21 (13.8%) 7 (13.2%) 69 (15.86%) ’ 1 (0.7%) 0 1 (0.23%) LliMQlLLL'J Serous 133 (87.5%) 40 (75.5%) 435 (100.00%) Non—serous 8 (5.3%) 4 (7.6%) 0 Mixed (6.6%) 1 (1.9%) 0 WO 96843 PCTfU52012/07l380 Unknown [— 1 (0.7%) [—8 (15.1%) 0 F m l 5 R J 8 (5.3%) 1 (1.9%) 2 (0.46%) 18 (11.8%) 12 (22.6%) 50 (11.49%) 126 ) 40 (75.5%) 373 (85.75%) 0 0 1 (0.23%) Unknown 0 0 8 ( 1 34%) Residual disease afier surgefl 1h— T— —_l —_l 9 (5.9%) 0 84 (19.31%) <= 1 cm 95 (62.5%) 44 (83%) 200 (45.98%) > 1 cm 40 (26.3%) 9 (17%) 102 %) '— 8 (5.3%) 0 49 (11.26%) .Smggrx J— j T _' 152 (100%) 53 (100%) 386 (88.74%) 0 0 0 Unknown 0 0 49 (11.26%) [— Chemotherapy 4% 139 (91.5%) 52 (98.1%) 399 (91.72%) Platinum (cis or carboplatin)-based (no taxane) 12 (7.9%) 1 (1.9%) NA Platinum plus Taxanc (paclitaxel or xel)-based 128 (83.6%) 51 (96.2%) NA 7 (4.6%) 0 23 (5.29%) Unknown 6 (4%) 1 (1.9%) 13 (2.99%) Table 3. Number of samples used in each assay.

Cohort 1 Cohort 2 Number of Reason assay was not Number of Reason assay was not Assay samples applied to all samples samples applied to all samples Affymetrix 500K SNP 152 not applicable not applicable ] 2/071380 l—arrays _ BRCA1 and BRCA2 tumor l— l sequencing 150 sequencing failed 52 sequencing failed _I j normal tissue not l normal tissue not ble or no available or no BRCA 1 and BRCA2 mutation detected in mutation detected in gennline sequencing 19 tumor 11 tumor cient tissue for l CCP and BRCA1 qPCR l3 7 RNA extraction l—BRCAJ and BRCA2 53 not applicable insufficient DNA for T— ] insufficient DNA for methylation is 126 analysis 34 analysis liOther HR gene Tinsufficient DNA for methylation analysis cient DNA for 92 analysis 0 analysis _. L _l % penicillin/streptomycin media at 37° C in T75 flasks until ~5x106 cell density. Exceptions were cell lines that required non-standard media, amine, or insulin. Cells grown in suspension were centrifuged for 5 minutes at 1700 rpm in a 1.5 mL centrifuge tube and the supernatant discarded. Cells grown in a monolayer had medium removed by aspiration, were washed with PBS, and trypsin solution added. Afier the cells detached they were collected in , transferred to a 1.5 mL microccntrifugc tube and centrifuged at 1700 rpm for 5 minutes. The supernatant was discarded.

Isolated cells were resuspended in 200 uL PBS Extraction ofGenomz'c DNA and Total RNAfiom Frozen Tumors and Cell Lines lOum frozen sections were cut and macrodissected.

The tissue was homogenized (TissueRuptor (Qiagen)) after addition of QlAzol lysis reagent, following by RNA isolation using a Qiagen miRNAeasy Mini Kit per the manufacturers protocol. A QIAamp DNA PCT/U52012/071380 lysis incubation at 56°C and RNase A treatment.

BRCA1 and BRCA2 Sequencing BRCA1 and BRCA2 sequencing was performed as described in Hennessy at, 2010. Mutations identified were only included in the analyses if fied as deleterious or suspected deleterious based on previously described criteria et and Tsui, 1993).

Promoter Met/zjzlation qPCR Assays 50 - 300 ng ofDNA was incubated for ~5 hours at 60°C with brief elevations to 95°C under acidic conditions in the presence of bisulfite. After tion, the reaction bound to a spin column and washed under basic conditions to remove bisulfite, converted DNA then eluted in 15 uL. Lower case region of primers is specific to the genomic region being amplified. Upper case region of s corresponds to the 454 Titanium chemistry tails and a 4 bp bareode (last 4 bases before the region c bases). By combining the forward and reverse primers in multiple combinations, it is possible to multiplex up to 100 s in a single sequence reaction.

BRCA1 and Cell Cycle Progression Signature Expression Assays PCT/U52012/071380 preamplification replicates were run at 8 and 18 cycles respectively for cell cycle gene . Three preamplification replicates were run at 18 cycles only for BRCAI assays. The post-amplification products were d 1:5 in low-EDTA Tris -EDTA (TE). Quantitative Polymerase Chain Reaction (qPCR) was then performed and assessed on Gene Expression M48 Dynamic Arrays (Fluidigm, South San Francisco, CA) per manufacturer’s protocol. The comparative cycle threshold (CT) method was used to calculate relative gene expression. CTs from preamplification of different numbers of cycles were centered by the average of the genes on the replicate that were in common between all replicates. The resulting values were normalized first by the average CTs of the housekeeper genes then by the average of the normalized CTs of each assay on all samples from the first cohort to yield AACT. CCP score and relative BRCAI expression were calculated as the average of the negative of the AACTs of the cell—cycle genes and BRCA1 assays, respectively.

Identification ofSamples with Loss 0fBRC41 Expression: Samples in which CCP expression and BRCA1 expression are orrelated were defined as BRCA] deficient. The threshold for identifying patients with abnormal BRCAI 300). BRCAI expression was regressed on CCP score using iteratively ghted least squares (IWLS). Points outside of the 99% prediction interval on the low end were considered abnormal.

This method is described in r detail in International ation No.

PCT/U$201 1/0543 69 to Timms et a].

Afifi/metrix 500K GeneChip arrays generated an amplified product of average size between 200 and 1,100 bp. Amplification products were purified using a Clontech DNA amplification cleanup kit. 90 ug fied DNA fragmented using trix Fragmentation Reagent. —labeling of the fragmented sample was accomplished using the GencChip DNA Labeling t. Biotin-labeled DNA was ized NSpl or Sty] Affymetrix microarrays at 490C for 16 to 18 hours in the Affymetrix rotation oven.

After hybridization, probe array wash and stain procedures were carried out on the automatic PCT/U82012/071380 Affymetrix Fluidics Stations as per manufacturer’s manual and microarrays Were scanned and raw data was collected by Affymetrix GeneChip Scanner 3000.

CNandLOH analysis ofSNP microarray data Abkevich er al. (publication no. WO/201 1/106541). The algorithm used in this paper was implemented in two versions, one for analysis of Affymetrix 500K ip array data generated ally, and the other for analysis of GenomeWideSNP6 Affymetrix array data downloaded from the TCGA web site (http://tcga—data.nci.nih.gov/tcga/dataAccessMatriX.htm? diseaseType=OV). The latter array, in on to SNP probes, contains a number ofprobes for non-polymorphic ons across the human genome. These probes are informative for CN analysis but are not ly informative for LOH is.

Statistical Analysis p-values in this paper were calculated using Kolmogorov-Smimov test unless otherwise specified.

Results HR Deficient Tumors samples from the first cohort were carriers of mutations in BRCAl and/or BRCA2, along with 14/53 from the second cohort and 83/435 from the third cohort (two of which were excluded from the further analysis, see below). Mutations are summarized in Table 4.

Table 4. BRCAI, BRCA2, and RADSIC defects detected in the study cohorts. $312+ BRCAI RADSIC BRCA2 TM 1 RAD51C mutation mutation N methylation mutation a methylation + BRCA] mutation PCT/USZOl2/071380 2 53 0 11 3 14 ND ND ND 3 435 0 51 85‘ 435 11 0 T J34] — Two of these mutations were excluded from the analysis because one copy of BRCA2 remained intact.

] Low mRNA sion ofBRCA1 or BRCA2 might also lead to HR deficiency, and be the result ofmechanisms other than promoter methylation. BRCA1 and BRCA2 second cohort. Expression of BRCA1 in 20 samples was abnormally low. Only five s with abnormally low expression ofBRCA1 were not flagged as HR deficient due to BRCA] promoter methylation. No abnormally low expression was observed for BRCAZ.

A single intact copy ofBRCA1 or BRCA2 is required for functionality. all BRCA] deficient samples, the BRCA1 gene is contained within a region of LOH. In addition, for all but two BRCA2 deficient samples, the BRCA2 gene is observed within an LOH region.

These two BRCA2 deficient samples were not considered HR deficient in our analysis.

Distribution oflengths QfLOH regions The initial hypothesis was that regions with LOH of ent length might appear in the cancer genome h different pathways, thus association between LOH and HR deficiency might depend on the length of LOH regions.

The distribution of lengths ofLOH adjusted on the length of chromosome arm on which these LOH s have been obsewed is shown in Figure 16. Chromosomes 13, 14, 15, and 22 were excluded because SNPs are not l380 to LOH over the whole chromosome.

The observed distribution is quite different from the similar distribution obtained for CN variations (Beroukhim et al. 2010), this suggests that CN variations LOH regions might arise via ent mechanisms.

Correlation between samples with HR deficiency and LOH ] The first cohort of samples was used as the “discovery” cohort. LOH s on chromosome 17 were excluded from the analysis because in almost all s LOH observed over this chromosome, probably because genes important for progression of ovarian cancer are on this chromosome. We checked for correlation between HR deficiency and the number of short LOH regions (<15 Mb), the number of long LOH regions (>15 Mb but less than the whole some), and the number ofLOH regions covering whole chromosomes. Various ent LOH region length cut-offs can be used and the influence of this cut-off on detecting HR deficiency is explored in Figure 19 and its accompanying discussion, though 15 Mb was found to be preferred. There was no significant correlation between the number of short LOH regions and HR ncy. The number ofLOH regions covering the whole chromosome was significantly larger in tumors with intact BRCA] or BRCA2 (p=4x10’5).

The number of long LOH regions (termed hereafter in this Example 3 and throughout this document as “HRD score”) was significantly higher in tumors with deficient BRCAl or BRCA2 (p=9x10‘”) (Figure 17a).

The second and third cohorts were used to validate the results obtained for the first cohort. The correlation between HR deficiency and number of LOH regions covering whole chromosomes did not validate in the second , possibly due to low sample number, but significantly larger (p=3x10’l 1) among tumors with intact BRCA1 and BRCA2 in the third cohort.

A highly cant correlation was observed between HRD score and HR deficiency for both s (p=2x10‘7 and p=9x10‘30 respectively) with HRD score being distinctly reduced among n tumors with intact BRCA1 and BRCA2 (Figures 17b and l7e).

Alterations in RAD51C and other HR pathway genes Available data suggest that BRCA1 and BRCA2 are the y genes responsible for HR deficiency in ovarian cancer. However, many other genes may also be important with, for example, both RADS 1C (Meindl et al., 2010) and RADS 1D (Loveday er al, 2011) recently being implicated as predisposition genes for ovarian cancer. The degree of ation measured for promoter CpG islands of eight additional genes involved in the HR pathway (Table 5) in the first cohort. Only RADSIC had high levels of promoter methylation (3 of 89 samples). In the WO 96843 PCTfU82012/071380 third cohort ll of435 samples had methylation of the RADS 1C promoter. All samples positive for RADSlC methylation from both cohorts were homozygous at the RADSlC locus due to LOH. To test whether the HRD score is elevated in samples with RADSlC promoter methylation these samples from both cohorts were ed with BRCA intact samples without RADS lC methylation. tent with our observations for BRCA1 and BRCA2 genes, HRD score was significantly higher (p=0.0003) among samples with RADS lC methylation.

Table 5. Promoter methylation assays used (SABiosciences).

Gene Symbol Description _'- Assay catalog ID Mediator or DNA damage checkpoint l—MEPH08721—2A J PARP] Poly(ADP-ribose) polymerase 1 _i—MePH02379-2A —_i BRCA] Breast Cancer 1, early onset _| MePH28472-1A l—BRCA2 Breast Cancer 2, early onset WCPH2 8473— l A _| i313” 0 homolog MePH2 8350-1A RAD51C RADSI homolog C |?ALB2 Partner and localizer ofBRCA2 —_iLN1613H22389-1A\J_‘ MePH28516-1A I-CHEIQ CHKZ checkpoint g LMePH28264-1A —| l—ATM Ataxia telangiectasia mutated +MePH28470-1A —_1 RAD5/ homolog —_i MePHl907l-2A defects with a high likelihood of being deleterious (e. g., nonsense and frameshifi mutations), resulting in a total of 8 deleterious mutations in 6 genes (ATM, ATR, FANCA, FANCDZ, FANCM, and PALBZ). An additional 5 samples had methylation of HR y genes. Loss of the second allele was both alleles is needed to loose function of a tumor ssor, most of these 13 samples are expected to have intact HR. Not surprisingly, HRD score was not elevated in the majority ofthese samples. is ofCombined data PCTfU52012/071380 Correlation between HRD score and HR deficienc y (defined as deficiency of BRCAI, BRCA2, or RADS 1C) for all three cohorts is presented in the Figure 17d.

A highly significant association is seen (p=2x10‘54).

An important question is whether the distribution of HRD scores is the same for HR deficiency due to different genomic loci. To answer this, the butions ofHRD scores for BRCA], BRCA2, and RADSIC deficient tumors were analyzed separately (Figure 21). A significant difference was observed (p=7x10‘5) with BRCAl deficient samples having higher average HRD score (16.1; SD: ions in BRCA1 and BRCA2 were gennline or somatic.

There is no cant difference for somatic vs. germline in the distributions ofHRD scores for either BRCA] or BRCA2 deficiency (Figure 20).

HRD score in BRCA1 and BRCA2 deficient cell lines cted breast (n=34) and ovarian (11:29) cell lines were obtained from multiple s; in addition 3 colon and one atic cell line from NCI60 with published BRCA1 and BRCA2 status were analyzed. Ofthese 67 cell lines, seven either carried homozygous deleterious mutations or had methylation of the BRCA1 er, two had homozygous mutations with apparent Functional reversion, and six carried heterozygous mutations. Figure 18a shows the distributions ofHRD scores for these three groups ofmutants, as well as for wild PCT/U52012/07I380 is significant correlation between HRD score and BRCA1 and BRCA2 deficiency after excluding ovarian cancer cell lines from the dataset (p=0. 01), suggesting that association ofHRD score with HR deficiency18 not cted to n cancer.

Correlation n HR deficiency and overall survival (OS) and progressionfree survival (PFS) A cant ation was observed between PFS (p==0.03) and OS (p=6x105) for the third cohort with improved survival for patients with higher HRD scores (Figure 18b). P-values were calculated using Cox model The s are in agreement with, and extend previously reported data showing that germline mutations 1n BRCA1 and BRCA2 are associated with improved outcomes for ovarian cancer (Rubin et al, 1996, Boyd et al., 2000; Cass et al 2003; Tan et al., 2008, Hennessy et al, 2010,). sion The HRD score was validated in two independent ovarian cancer datasets, and also reflected mutations resulting 1n HR ncy1n breast and pancreatic cell lines.

Table 6. Average of HRD score for BRCA1 and BRCA2 deficient and intact tumors and corresponding p values.

HR deficient . HR deficient HR intact (BRCAT HR Intact (BRCAJ (BRCA 1 and (BRCA1,BRCA",- BRCA ° and BRCA2) 2, an(1 BRCA2) and RAD51 C) RAD5IC) .9 (SD=4.6) 8.3 (SD=6 1) 16.2 (SD=4.9) 8.0 (SD=5.8) First cohort _I . _J p=9x10‘“ |__ p=7x10” _,_ _i_ .6 (SD=4.4) 5.6 (SD=4.9) 15.6 (SD=-4 4.) 56 (SD=4 .9) Second cohort p=2x10'7 p=2x107 .3 (SD=4.3) 8.8 (SD=5.0) 151 (SD=4.3) 8.6 (SD=5. 0) Third cohort p=9x10‘3° l —| p=2x103 Combined data for 15.5 (SD=4.4) ‘ 84 (SD==53) 154 (SD=44)p22, —| 8.2 (SD=5.2) three cohorts p=1045 "l— L )(1‘054 PCT/USZOl2/071380 19.7 ('SD=4.6) 8.2 4) 19.7 (SD=4.6) 8.2 (SD=5,4) Cancer cell lines p=l 0‘5 p=1 0'5 of this type ofLOH class exists e it incorporates double strand DNA breaks as part of its genesis and requires repair by HR. In contrast, LOH at the whole chromosome level is significantly less frequent in HR deficient tumors. One possible ation is that LOH at the whole chromosome level originates through an alternative competing mechanism that does not involve double strand DNA breaks.

In addition to BRCA1 and BRCA2 defects, RADS l C er methylation is observed in ovarian tumors. High HRD score was significantly associated with RADSlC deficiency in two ts. Only one additional HR deficient tumor was confirmed in the 3 datasets, a nonsense mutation in FANCM with LOH resulting in loss of the second .

The HRD score associated with the FANCM mutation (8) is within the range of the normal distribution for samples with elevated HRD score.

Among tumors with apparently intact BRCAl , BRCA2, and RADS l C, a. substantial fraction of the samples have an elevated HRD score. One possible ation is that there is a substantial rate of defects in other genes in the HR pathway in many of these samples. An alternative explanation is that contamination of the tumor with normal tissue complicates detection of defects. Data suggest that the HRD score is less sensitive to contamination than other assays, and that undetected defects may explain a significant fraction of those samples with elevated HRD score (see Supplementary Results).

Published s have demonstrated that secondary reversion mutations 2012/071380 associated with HR deficiency because they are functionally linked to it. Consequently, the HRD score is a very robust measure of HR deficiency.

However. its permanence means the score would tumors used in this study, however data obtained from cell lines is consistent with this hypothesis.

Failure to detect reversion mutations will result in false ves. This is likely to affect very few tumors in the neoadjuvant or adjuvant setting (Norquist er al., 2011) and is less of a concern than false negatives which would ectly identify duals as likely non—responders.

High HRD score is highly correlated with HR deficiency, and this score can be utilized to identify patients with high likelihood of responding to DNA damaging agents and PARP inhibitors (among other agents).

Such a test has clear al utility in breast and ovarian cancer, and can be used to expand the use ofPARPi and platinum salts to other cancers where HR deficiency is less well characterized.

Example 4 — Further Validation of HR ncy Assay Materials and Methods The patient cohort analyzed in this e included 56 breast cancer patients, all ofwhom are either BRCA mutation positive or have triple negative breast cancer (most are TNBC). Stages 1 - III were included (most are 11 or III). The patients received 6 cycles of neoadjuvant gcmcitabine + iniparib + carboplatin.

Response was measured as relatively lower residual cancer burden following treatment. 56 fresh frozen breast tumors were analyzed. Median degree of contamination is 60%. Nine samples had ination of at least 90%. 11 of these tumors were carriers of BRCA1 deleterious mutations and three were carriers ofBRCA2 deleterious mutations.

In all of these tumors there was LOH at the deficient genes. One of the carriers of BRCA1 deleterious mutations also d a deleterious mutation in BRCAZ. However in that sample there was no LOH at BRCA2 gene. 30 s were obtained from patients who ded to treatment (residual cancer burden either 0 or 1). 13 ofthem are BRCAl/Z deficient. 26 samples were obtained from non-responders (residual cancer burden either 2 or 3), One of them is BRCAl deficient.

Gcnotyping analysis was performed by Affymetrix using Affymctrix MIP arrays (as described in US. Pat. No. 412; US. Patent Application Publication No, USZOO60234264; Hardenbol et a/., PCT/U82012/071380 for BRCAI deficient non-responder was 8. According to the Mann-Whitney U test p-value for association n se to treatment and HRD score was 0004. If BRCAl/Z deficient samples are excluded association n se to treatment and HRD score remains si ‘ = 0.02).

The differences in HRD score amongst samples with residual cancer burden 0 and l were not significant. Similarly, the differences in HRD score amongst samples with residual cancer burden 2 and 3 were not significant. Correlations between response to treatment and clinical parameters (stage, grade) were not significant.

OTHER EMBODIMENTS WHAT WE

Claims (2)

CLAIM IS

1. A method for assessing LOH in a cancer cell or genomic DNA thereof, wherein said method comprises: (a) detecting, in a cancer cell or genomic DNA derived therefrom, LOH s in at least one pair of human somes of said cancer cell, wherein said at least one pair of human chromosomes is not a human X/Y sex chromosome pair; (b) determining the total number of LOH regions, in said at least one pair of human chromosomes, that are longerthan a first length but shorter than the length of the whole chromosome containing the LOH region, wherein said first length is about 1.5 or more megabases, and (c) correlating said total number that is greater than a reference number with an ficiency LOH signature of the cancer cell or genomic DNA thereof.

2. A method of predicting the status of BRCA1 and BRCA2 genes in a cancer cell, comprising: determining, in the cancer cell, the total number of LOH regions in at least one pair of human chromosomes of said cancer cell that are longer than first length but shorter than the length of the whole chromosome containing the LOH region, wherein said at least one pair of human chromosomes is not human X/Y sex chromosome pair, wherein said first length is about 1.5 or more megabases; and correlating said total number that is rthan a reference number with an increased likelihood of a deficiency in the BRCA1 or BRCA2 gene.