Resequencing and fine-mapping of the chromosome 12q13-14 locus associated with multiple sclerosis refines the number of implicated genes

Multiple sclerosis (MS) is a common chronic inflammatory disease of the central nervous system. Susceptibility to the disease is affected by both environmental and genetic factors. Genetic factors include haplotypes in the MHC and over 50 non-MHC loci reported by genome-wide association studies. Amongst these, we previously reported polymorphisms in chromosome 12q13-14 with a protective effect in individuals of European descent. This locus spans 288kb and contains 17 genes; including several candidate genes which have potentially significant pathogenic and therapeutic implications. In this study we aimed to fine-map this locus. We have implemented a two phase study: a variant discovery phase where we have used next-generation sequencing and two target-enrichment strategies (long-range PCR and Nimblegen's solution phase hybridization capture) in pools of 25 samples; and a genotyping phase where we genotyped 712 variants in 3,577 healthy controls and 3,269 MS patients. This study confirmed the association (rs2069502, P=9.9 x 10-11, OR=0.787) and narrowed down the locus of association to a 86.5kb region. Although the study was unable to pinpoint the key-associated variant we have identified a 42 (genotyped and imputed) SNP haplotype block likely to harbour the causal variant. No evidence of association at previously reported low-frequency variants in CYP27B1 was observed. As part of the study we compared variant discovery performance using two target-enrichment strategies. We concluded that our pools enriched with Nimblegen's solution phase hybridization capture had better sensitivity to detect true variants than the pools enriched with long-range PCR, whilst specificity was better in the longe-range PCR enriched pools compared to solution phase hybridization capture enriched pools; this result has important implications for the design of future fine-mapping studies.