Detection of Large Clonal Mosaic Events in Existing Genome-wide Association Study Data. M. Machiela1, W. Zhou1,2, M. Yeager1,2, K. Jacobs1,2, S. Berndt1, Q. Lan1, N. Rothman1,2, S. Savage1, P. Taylor1, N. Caporaso1, I. DeVivo3, R. Hoover1, M. Tucker1, S. Chanock1 1) Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD; 2) Cancer Genomics Research, SAIC-Frederick, Inc., NCI-Frederick, Frederick, MD; 3) Harvard School of Public Health, Boston, MA; Channing Division of Network Medicine, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA 02115.

   Mosaicism is defined as the presence of two or more genetically distinct cellular populations in an individual who developed from a single zygote. With the advent of genome-wide association studies, it is possible to evaluate the relative intensities of the genotype signals to detect large structural mosaicism in human populations. We performed an analysis of germline DNA derived from blood or buccal specimens from 24,849 individuals in 46 cancer-related studies using a modified mosaic alteration detection algorithm on renormalized B-allele frequencies and log2 relative probe intensity ratios from commercially available Illumina SNP arrays. Overall, we confirmed our previously reported findings (Jacobs et al Nat Gen 2012) including a similar distribution of types of events with respect to chromosomal position, an increase with age (overall and in cancer free-controls) and an increase in men versus women. Our segmentation algorithm detected 341 clonal mosaic events >2 Mb in size in 168 individuals, 91 in cases and 77 in controls. Among individuals with mosaicism, 27 had two or more events, including one individual with 34 events. The detected mosaic events include 163 (47.8%) copy-neutral losses of heterozygosity, 90 (26.4%) copy losses, 69 (20.2%) copy gains, and 19 (5.6%) complex events. The most common events detected were copy neutral losses of heterozygosity on chromosomes 4q and 9q and copy losses on 13q14 and 20q. The 13q14 deletion, commonly observed in chronic lymphocytic leukemia cases, was observed in our dataset among non-hematologic cancer cases suggesting a potential role of 13q14 deletions in these malignancies. A breakpoint analysis of the 13q14 deletion revealed that breakpoints clustered in genomic regions enriched with DNaseI peaks, transcription factors, and common repeat elements. An unadjusted analysis of age indicates the mosaicism increases with increasing age (p=5.3e-4), providing further support for age-related genomic senescence. Combined with independent data from our previous study of 57,835 individuals, analyses indicate mosaicism is associated with male sex with 353 of 37,287 males (0.95%) and 191 of 33,223 females (0.57%) having mosaic events (p=0.001). These results suggest that human mosaicism may be more common in the general human population than previously expected; underscore the importance of thorough characterization of germline DNA; and potentially represent the tip of the iceberg for smaller mosaic events.

You may contact the first author (during and after the meeting) at