P265 HLA haplotype structure in a Chinese population

Abstract

Aim In order to study the gene structure of HLA haplotypes, these haplotypes must be determined in a large sample from a given population. Diploid combinations of alleles from different loci must be placed in phase with each other. In this study, this accomplished by family segregation analysis. Methods Samples come from a Chinese population in Taiwan collected for disease association. 490 mother-father-child trios were used in this study. HLA typing is based on Mia Fora next-generation sequencing by Immucor. The following loci are included: A, B, C, DRB1, DRB345, DQA1, DQB1, DPA1 and DPB1. A simple propositional-logic algorithm based on basic Mendelian genetics principles was used to do family segregation analysis and determine the haplotypes in each family. Results Full 2 × 2 contingency tables, for each pair of variables, with their corresponding chi-square, p value and other measures of correlation are presented for the following pairs of HLA loci or groups of loci: 1) B vs C, 2) DRB1 vs DRB345, 3) DQA1 vs DQB1, 4) DPA1 vs DPB1, 5) A vs B-C, 6) DRB1-DRB345 vs DQA1-DQB1, 7) DPA1-DPB1 vs DRB1-DRB345-DQA1-DQB1, and 8) A vs DRB1-DRB345-DQA1-DQB1. Conclusions Results may be biased by the disease association in this project. (The disease in question cannot be disclosed at this time.) Nevertheless, the bias effect will hardly diminish the value of the linkage disequilibrium data presented here. It is possible that haplotype frequencies are over-represented or under-represented, but the identity of the haplotype itself will not be affected by this bias. One of the main finding in this study is that, contrary to what is generally believed, NGS does not disclose a hidden abundance of HLA alleles, it rather helps us characterise in full detail alleles we were already familiar with. This study shows the distribution of HLA alleles in a given population, but their inter-loci structure, their association with alleles in other loci. The second main finding in this study is that in those cases where more than one genetic variant is represented by a single previously known protein allele, the linkage disequilibrium measures increase significantly when this genetic variants are taken into account individually.