Empirical Evaluation of Methods for Identifying Recurrent Copy Number Variations Across Multiple Samples

• Activity Number: 133
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Jeanette E. Eckel-Passow*+ and Elizabeth J. Atkinson and Vernon S. Pankratz and Christopher G. Scott and Janet E. Olson and Julie M. Cunningham and Fergus J. Couch and Thomas Sellers and Celine M. Vachon
• Companies: Mayo Clinic and Mayo Clinic and Mayo Clinic and Mayo Clinic and Mayo Clinic and Mayo Clinic and Mayo Clinic and H. Lee Moffitt Cancer Center & Research Institute and Mayo Clinic
• Keywords: copy number ; plink ; composite ; cover ; cnvpack

Abstract

Copy Number Variations (CNVs) are structural changes to regions of chromosomal DNA, resulting in a change in the normal diploid copy number. CNVs are detected for each sample independently and as a result of the inherent noise in the technology, most CNVs will contain sample-specific breakpoints. In order to perform downstream association analyses it is necessary to first define recurrent CNV regions across samples, specifically, to define a common chromosomal start and stop location across samples for each defined recurrent CNV. We empirically evaluated methods for identifying recurrent regions using the GENetic Epidemiology of MAMmographic Density (GENEMAM) study that evaluated 472 subjects from 90 families. The data were generated using the Illumina 660 SNP array. Sample-specific CNV regions were identified using PennCNV. Recurrent regions were identified using three methods: Plink and the COMPOSITE and COVER methods available in the R package cnvpack. We tested for an association between each recurrent insertion and deletion with breast density adjusting for familial correlation. The influence of each of the three methods on the association results will be presented.