Two - Phase Subsampling Designs for Genomic Resequencing Studies

Targeted resequencing of DNA at specific genes or other genomic loci is now feasible for hundreds or thousands of samples, and costs for larger-scale resequencing are decreasing rapidly. For at least the 1 2 T. LUMLEY ET AL. next few years, resequencing will need to be confined to small subsets of the large samples on which genome-wide association studies have been recently been performed. This paper describes some strategies for subsampling an existing cohort for resequencing, and flexibly analysing the resulting data. We illustrate these strategies by describing the actual design and planned analyses for the example that motivated our research, the CHARGES resequencing study carried out by the CHARGE (Cohorts in Heart and Aging Research in Ge-nomic Epidemiology) Consortium. 1. Introduction. The past few years have seen an explosion in the availability of genotype information. Genome-wide association studies have been performed with hundreds of thousands of genetic markers on sample sizes from tens to hundreds of thousands of people, and these have found many, mostly weak, associations between genetic variants and biological or clinical variables. The greatest likely benefit from these studies is in improving the understanding of biological processes in health and disease, rather than in direct prediction, but biological understanding is hampered by the fact that the association studies typically find a set of nearby genetic markers rather than the genetic variant or subset of variants that truly affects biological processes. One approach to finding functional genetic variants near a set of markers is to determine the complete DNA sequence of a region of the genome. Analysis of complete sequence data should sharpen the association estimates compared with analysis only of marker data. More importantly, as statistical association is unlikely to be sufficient to narrow the association down to a single variant, having sequence data facilitates biological investigation of candidate variants, whether in silico, in vitro, or in vivo. DNA resequencing is technologically feasible but very expensive at sample sizes sufficiently large for association studies. Although the costs are decreasing rapidly, over at least the next few years it will be necessary to resequence relatively small subsam-ples from the large samples that have been participated in genome-wide association studies. Two large resequencing projects in cardiovascular disease were funded by the US National Institutes of Health under the American Recovery and Reinvestment Act. In this paper we present the design and analysis principles for one of these studies, …