Rxivist logo

Rxivist combines preprints from bioRxiv with data from Twitter to help you find the papers being discussed in your field. Currently indexing 62,102 bioRxiv papers from 275,834 authors.

Practical Use of Methods for Imputation of HLA Alleles from SNP Genotype Data

By Allan Motyer, Damjan Vukcevic, Alexander Dilthey, Peter Donnelly, Gil McVean, Stephen Leslie

Posted 09 Dec 2016
bioRxiv DOI: 10.1101/091009

The human leukocyte antigen (HLA) genes play an essential role in immune function. Typing of HLA alleles is critical for transplantation and is informative for many disease associations. The high cost of accurate lab-based HLA typing has precluded its use in large-scale disease-association studies. The development of statistical methods to type alleles using linkage disequilibrium with nearby SNPs, called HLA imputation, has allowed large cohorts of individuals to be typed accurately, so that massive numbers of affected individuals and controls may be studied. This has resulted in many important findings. Several HLA imputation methods have been widely used, however their relative performance has not been adequately addressed. We have conducted a comprehensive study to evaluate the most widely used HLA imputation methods. We assembled a multi-ethnic panel of 10,561 individuals with SNP genotype data and lab-based typing of alleles at 11 HLA genes at two-field resolution, and used it to train and validate each method. Use of this panel leads to imputation accuracy far superior to what is currently publicly available. We present a highly-accurate new imputation method, HLA*IMP:03. We address the question of optimal use of HLA imputations in tests of genetic association, showing that it is usually not necessary to apply a probability threshold to achieve maximal power. We also investigated the effect on accuracy of SNP density and population stratification at the continental level and show that neither of these are a significant concern.

Download data

  • Downloaded 1,870 times
  • Download rankings, all-time:
    • Site-wide: 2,456 out of 62,102
    • In genetics: 231 out of 3,530
  • Year to date:
    • Site-wide: 7,486 out of 62,102
  • Since beginning of last month:
    • Site-wide: 48,780 out of 62,102

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide

Sign up for the Rxivist weekly newsletter! (Click here for more details.)