OR53: HLA-EpiDB: A DATABASE FOR THE ANALYSIS OF HLA EPITOPE FREQUENCIES IN WORLDWIDE POPULATIONS

Abstract

Aim The presence of anti-HLA donor-specific antibodies in transplant patients is a crucial factor related to tissue and graft rejection. These antibodies target specific regions of HLA proteins that are different from the transplant patient’s HLA proteins – termed HLA epitopes. Current efforts in matching for kidney transplantation often attempt to minimise the number of allele-level mismatches, and yet there is growing evidence that an HLA epitope-focussed approach for minimising mismatches may be more appropriate. In this study, we present the development of a new section within the Allele Frequency Net Database (AFND, www.allelefrequencies.net ), called HLA-EpiDB, for the storage of HLA epitope frequencies across worldwide populations. Methods To generate epitope frequencies we used seventeen genotype datasets and twenty-four haplotype datasets from 41 worldwide populations comprising more than 36,000 individuals, which were also stored in AFND. All populations datasets used in this study were typed at high resolution level for HLA-A, -B and -DRB1 loci, and nineteen also included HLA-C. To map the >2000 HLA Class I alleles that have been reported in IMGT/HLA database as of release 3.16.0 (April, 2014), we used the definition of HLA epitopes and mapping to HLA alleles established by the HLA Epitope Registry ( http://www.epregistry.ufpi.br/ ). Epitope frequencies (percentage of individuals carrying the epitope) were estimated using Hardy–Weinberg proportions for haplotype-derived data and by direct counting from genotype data. Results We created a new section in AFND called HLA-EpiDB, which allows users to filter for epitopes present in an individual HLA genotype, and analyse the likelihood of a population containing individuals with a closely matching epitope set. Another feature of HLA-EpiDB is that it permits users to display a cross-population comparison of epitope frequencies in a heatmap, or as a world map distribution for single epitopes. Conclusions We expect HLA-EpiDB to help transplant clinicians and the immunogenetics research community to better understand and determine HLA mismatch acceptability across worldwide populations. At present, we have estimated HLA epitope frequencies for HLA Class I alleles but Class II alleles will be added in the near future.