Copy-number signatures and mutational processes in ovarian carcinoma

Abstract

Genomic complexity from profound copynumber aberration has prevented effective molecular stratification of ovarian and other cancers. Here we present a method for copynumber signature identification that decodes this complexity. We derived eight signatures using 117 shallow wholegenome sequenced highgrade serous ovarian cancer cases, which were validated on a further 497 cases. Mutational processes underlying the copynumber signatures were identified, including breakagefusionbridge cycles, homologous recombination deficiency and wholegenome duplication. We show that most tumours are heterogeneous and harbour multiple signature exposures. We also demonstrate that copy number signatures predict overall survival and changes in signature exposure observed in response to chemotherapy suggest potential treatment strategies. 2. CC-BY-NC-ND 4.0 International license not peer-reviewed) is the author/funder. It is made available under aThe copyright holder for this preprint (which was . http://dx.doi.org/10.1101/174201 doi: bioRxiv preprint first posted online Aug. 9, 2017;The discrete mutational processes that drive copynumber change in human cancers are not readily identifiable from genomewide sequence data. This presents a major challenge for the development of precision medicine for cancers that are strongly dominated by copynumber changes, including highgrade serous ovarian (HGSOC), oesophageal, nonsmallcell lung and triple negative breast cancers 1 . These tumours have low frequency of recurrent oncogenic mutations, few recurrent copy number alterations and highly complex genomic profiles 2 .HGSOCs are poor prognosis carcinomas with ubiquitous TP53 mutation 3 . Despite efforts to discover new molecular subtypes and targeted therapies, overall survival has not improved over two decades 4 . Current genomic stratification is limited to defining homologous recombinationdeficient (HRD) tumours 57 , and classification using gene expression does not currently have clinical utility 8,9 . Detailed genomic analysis using whole genome sequencing has shown frequent loss of RB1, NF1 and PTEN by gene breakage events 10 and enrichment of amplification associated foldback inversions in nonHRD tumours 11 . However, none of these approaches has provided a broad mechanistic understanding of HGSOC, reflecting the challenges of detecting classifiers in extreme genomic complexity.Recent algorithmic advances have enabled interpretation of complex genomic changes by identifying mutational signatures genomic patterns that are the imprint of mutagenic processes accumulated over the lifetime of a cancer cell . Importantly, these studies show that tumours typically harbour multiple mutational processes requiring computational approaches that can robustly identify coexistent mutational signatures. Quantification of the exposure of a tumour to specific mutational signatures provides a rational framework to personalise therapy 14 but currently is not readily applicable to copynumber driven tumours. We hypothesized that specific feat...

Multisite analysis of high‐grade serous epithelial ovarian cancers identifies genomic regions of focal and recurrent copy number alteration in 3q26.2 and 8q24.3

“…No data were available for region γ. As TCGA only includes ovary biopsies taken at the time of diagnosis, we used another independent data set including both spatial and temporal biopsies 6 to confirm our results. Analysis of recurrent alterations on this dataset with GISTIC (Supporting Information Methods, Section 5) confirmed that regions α and β show gain respectively in almost 74 and 85% of the samples (Supporting Information Results Section 6, Fig.…”

“…2 From the 591 samples of serous ovarian cancer in the repository, were selected samples from patients with HGS-EOC stage III and IV tumors and with copy number alteration data available (Supporting Information Methods, Section 5). The second cohort was a selection of 91 temporal and spatial biopsies with high data quality from a previously published data set 6 where segmented copy number data were available (see Supporting Information Methods, Section 5).…”

“…The few concordant data so far suggest that: (i) the branching pattern of HGS-EOC evolution sustains the marked genomic heterogeneity of single nucleotide variants (SNVs) among synchronous lesions. [3][4][5][6] Most of these genomic abnormalities arise in a random fashion and are lesion-specific; (ii) relapsed-molecular features are largely derived from selective pressure of already distinct genetic events present in the primary tumor rather than acquired de novo. 4,7 These findings imply that a single tumor lesion-the ovary for example-cannot any longer be considered representative of the entire disease, nor it is informative on the genomic features of relapsed disease.…”

Critical questions in ovarian cancer research and treatment: Report of an American Association for Cancer Research Special Conference

“…Copy number signatures can also be used to combine agents that are likely to be more effective. 20 HGSOC cell lines appear to depend on "quality control" pathways that sense and destroy damaged transcripts and proteins. Nonsense-mediated decay (NMD) is highly active in HGSOC cell lines.…”

Targeting DNA repair: the genome as a potential biomarker

“…Recently, we have used low-cost shallow WGS (sWGS) to identify CN signatures in the genomes of HGSC [88]. We used biopsy samples from the BriTROC-1 study (the first study of the UK Translational Research in Ovarian Cancer Consortium -detailed in [89]) of relapsed HGSC, and combined relative CN data from sWGS (mean ×0.1) and TP53 mutant allele frequency from deep tagged-amplicon sequencing [90] to generate absolute CN profiles.…”