OR4 High resolution haplotype analyses of classical HLA genes in families with multiple sclerosis

Abstract

Aim HLA alleles are observed in specific haplotypes because of linkage disequilibrium between particular alleles. Multiple sclerosis (MS) has been associated with HLA genes and haplotypes. Our goal was to identify candidate HLA gene alleles and haplotypes that are susceptible or protective to MS. Methods We applied a high-throughput, high-resolution NGS method to type 11 HLA loci in 481 trio families comprising of 962 unaffected parental controls and 481 children diagnosed with MS. We developed an in-house computer program named HaplObserve that builds haplotypes by comparing offspring and parents HLA genotypes from nuclear families. HLA haplotypes were built from the trio families using HaplObserve that also allows tracing HLA allele/haplotype transmissions from the parents to the offspring. We dissected the extended haplotypes generated into smaller haplotypes and alleles, then explored associations with MS using standard transmission disequilibrium test (TDT) and multiallelic TDT analyses. Results We built a pipeline to systematically analyze HLA alleles and various segments of haplotypes using standard TDT and multiallelic TDT tests. Standard TDT and multiple allelic TDT showed that the DRB5∗01:01:01 ∼ DRB1∗15:01:01:01 haplotype was significantly predisposing (standard TDT: p 2.2e−16). We identified that DRB1∗01:01:01 ∼ DQB1∗05:01:01:03 (standard TDT: p = 8.63e−06), C∗04:01:01:01 ∼ B∗35:01:01:02 (p = 0.001645) and A∗02:01:01:01 ∼ C∗03:04:01:01 ∼ B∗40:01:02 (p = 0.008318) haplotypes were protective. Our analysis indicated that DRB1∗11:01:01:01 (p = 0.001738), B∗27:05:02 (p = 0.0004531), B∗38:01:01 (p = 0.0031953) and C∗07:04:01:01 (p = 0.004126) were protective. Conclusions We have shown the suitability of the HaplObserve to efficiently build haplotypes from NGS data in a large number of families, as well as the efficacy and power of TDT analysis for MS association studies. By comparing statistical significance of various haplotype segments, we were able to fine-map MS candidate alleles/haplotypes while eliminating false signals resulting from ‘hitchhiking’ alleles.