Abstract 3983: Evaluation of data compression strategies for genetic variation calling using next generation sequencing data in normal and tumor samples

Abstract The per-base cost of whole genome sequencing has dropped dramatically due to recent advancement of next generation sequencing (NGS) technologies, and a number of large-scale re-sequencing projects, e.g. 1000 Genomes, ICGC, TCGA, GO-ESP, and CGCI etc., have been initiated to extend our knowledge of single nucleotide polymorphisms (SNPs), short insertions/deletions (INDELs) and structural variations (SVs), and relate these variants to human diseases. To date, various compression technologies, including NCBI's cSRA and EBI's CRAM etc., have been developed to significantly reduce the storage footprints and costs. However, whether or not such compression treatments have any impact on genetic variant callings has not been systematically assessed and validated with large scale NGS datasets as different calling algorithms for different variant classes (SNPs, INDELs, or SVs) may have different underlying data requirements. We have developed an integrated analysis framework (VarPipe) to profile genetic mutations using NGS data generated from various sequencing platforms in a uniform manner. This pipeline is also applied to the official data from 1000Genomes (normal samples) and Cancer Genome Characterization Initiative (tumor samples) projects to evaluate the compressions effects on variation callers. With defined quality metrics, the variant callsets from compression-treated and untreated data are carefully evaluated as well as compared to the official variation releases of the corresponding projects so that optimal compression strategy can be developed and recommended for variation detection using NGS data. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3983. doi:1538-7445.AM2012-3983