Abstract 5320: Novel computational methods reveal subtle clonal mosaicism

Abstract

Recent genome-wide surveys have revealed age-related genetic mosaicism and established a link between clonal mosaicism and risk of hematological and solid-tumor cancers (Laurie et al, 2012, Nat Genet, 44:642; Jacobs et al, 2012, Nat Genet, 44:651; Forsberg et al, 2012, AJHG, 90:217). Existing methods to identify chromosomal aberrations leading to mosaic allelic imbalance (AI) rely on the detection of increased variance of the estimated within-sample allele frequencies at heterozygous loci, ignoring the dependence among allele frequencies in regions of an AI-inducing event, such as a deletion or copy-neutral loss of heterozygosity (LOH), expected from the loss or gain of a chunk of an inherited chromosome. As a result, they break down when the proportion of DNA from an aberrant source is below 10-15%. Here we present a suite of new methods (hapLOH, J-LOH, hapLOHseq) that harness the important information contained in the germline haplotypes to help distinguish signal from noise, accommodate multiple samples of different tumor purities (to deconvolute aberration type and tumor purity) and infer co-occurrence of aberrations in heterogeneous mixtures. For example, our initial method, hapLOH (Vattathil & Scheet, 2013, Gen Res 23:152), can identify 10 Mb regions with just 4% of the signal coming from the tumor genome in data simulated from a series of experiments from a 317K SNP microarray. Application of hapLOH to data analyzed by Laurie et al (2012) indicates that we have substantially greater sensitivity to detect clonal mosaicism, particularly events harbored in a small proportion of cells (< 5-10%). We identify more than twice as many mosaic events, demonstrating the potential role for our methods to aid in profiling cancer risk. Our suite of tools is designed for both microarray and next-generation sequencing data and is available at scheet.org. Citation Format: Paul Scheet, Rui Xia, F. Anthony San Lucas, Christina Hahn, Jerry Fowler, Selina Vattathil. Novel computational methods reveal subtle clonal mosaicism. [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 5320. doi:10.1158/1538-7445.AM2014-5320