Application Of Whole Genome Sequencing To Identify Mendelian Disease Candidates In Dogs Using a 1-Case + N-Control Design

The traditional paradigm for the identification of Mendelian disease genes involves first mapping the disease locus to a genomic region, followed by fine-mapping to refine location, the selection and sequencing of candidate genes to identify mutations, followed by the genotyping of candidate causal variants to identify those consistent with disease and carrier status.  Mapping has traditionally been performed by linkage and, more recently, by GWAS in a case-control design.  Current NGS technology permits the rapid and cost effective resequencing of individual genomes but is not yet sufficiently cost effective for GWAS particularly for single disease case-control designs in which large numbers of cases are matched to non-disease controls carefully selected to avoid stratification.  However, we hypothesized that a more efficient design was possible that capitalized on the simultaneous analysis of multiple diseases and that utilized whole genome sequence based variant detection, making the issue of stratification irrelevant.  Specifically, we propose a case-control design in which the genomes of N+1 individuals possessing N+1 mutually exclusive diseases are resequenced and that a case-control study is performed for each disease as 1 case versus N controls. For autosomal recessive diseases, only stretches of homozygosity within the case are considered and homozygous mutations are eliminated if they are found in any of the controls. We show that that this approach allows the generation of a small list of candidate causal variants that require validation.  We present the method using two diseases, Fanconi Syndrome and Multiple System Degeneration, to illustrate the design’s power and utility.