Mark Pratt

Abstract 3576: Creating and accurately interpreting clinical grade cancer exomes: Challenges and solutions

Proceedings: AACR Annual Meeting 2014; April 5-9, 2014; San Diego, CA Exome and genome sequencing are increasingly utilized to diagnose and direct treatment for cancer patients. However, there are many barriers to sequencing cancer exomes in an accurate and comprehensive way that assesses all crucial cancer mutations. Critical sequencing coverage gaps in cancer genes, challenges calling structural variants, difficulty isolating DNA from FFPE and small biopsy samples, tumor heterogeneity and cellularity, and accurate annotation of mutations remain significant hurdles to accurate clinical cancer interpretation. We have developed a comprehensive approach aimed at improving accuracy and completeness of sequencing, structural variant calling, and annotation of cancer samples. Whole exome sequencing in particular is susceptible to missing key parts of the cancer genome. We have designed a unique targeted assay and set of protocols to augment and improve the standard exome sequencing assay in order to assess all major cancer genome mutations, including important structural variations. We have optimized DNA isolation and library creation from FFPE tumor samples and small sample sizes. To improve sequencing coverage of 494 cancer related genes and over 7,000 other genes of medical importance, we include additional optimized targeted enrichments on top of the standard exome. We also utilize altered sequencing protocols to fill in regions containing genomic elements that are typically hard to sequence by standard protocols. We compared coverage over 494 cancer genes with our augmented exome approach to two different standard exome capture kits, all at 12G of sequencing. In this comparison, we defined “finishing” a gene to mean greater than 99% of bases within the gene at greater than 20x average coverage. Our augmented exome approach finished 459 of 494 (93%) cancer genes. The standard exome platforms only finished 276 (56%) and 318 (64%) cancer genes at this level. While finishing cancer genes is paramount to detecting small mutations, structural variants (SVs) are also a very important class of mutations when assessing clinical cancer samples. In addition to our augmented exome, we also perform whole genome sequencing to facilitate detection of structural variants. We utilize a combined algorithm for structural variant calling which leverages multiple lines of genomic evidence to identify and score SVs. By combining orthogonal SV detection algorithms with local reassembly, we achieve greater sensitivity (96.27% compared to 55.6% average for each algorithm independently), and a lower false discovery rate (1.37% compared to 27.55% average) on SV detection in genome sequence than any of the methods used independently. By comprehensively and accurately assessing cancer mutations and leveraging public and privately curated databases for annotation, we are able to sequence and interpret clinical grade cancer exomes to a much higher level than standard exome sequencing. Citation Format: Michael J. Clark, Deanna Church, Mark Pratt, Elena Helman, Gabor Bartha, Stephen Chervitz, Sarah Garcia, Shujun Luo, Jason Harris, Anil Patwardhan, Richard Chen, John West. Creating and accurately interpreting clinical grade cancer exomes: Challenges and solutions. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3576. doi:10.1158/1538-7445.AM2014-3576

Variant priorization and analysis incorporating problematic regions of the genome.

In case-control studies of rare Mendelian disorders and complex diseases, the power to detect variant and gene-level associations of a given effect size is limited by the size of the study sample. Paradoxically, low statistical power may increase the likelihood that a statistically significant finding is also a false positive. The prioritization of variants based on call quality, putative effects on protein function, the predicted degree of deleteriousness, and allele frequency is often used as a mechanism for reducing the occurrence of false positives, while preserving the set of variants most likely to contain true disease associations. We propose that specificity can be further improved by considering errors that are specific to the regions of the genome being sequenced. These problematic regions (PRs) are identified a-priori and are used to down-weight constitutive variants in a case-control analysis. Using samples drawn from 1000-Genomes, we illustrate the utility of PRs in identifying true variant and gene associations using a case-control study on a known Mendelian disease, cystic fibrosis (CF).