Tools for building, analyzing and evaluating HLA haplotypes from families

• Published in: Human immunology Vol. 80; no. 9; pp. 633 - 643
• Main Authors: Osoegawa, Kazutoyo, Mack, Steven J., Prestegaard, Matthew, Fernández-Viña, Marcelo A.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.09.2019
• Subjects: 17th International HLA and Immunogenetics Workshop; Alleles; Child; Gene Frequency - genetics; Genetic Loci; Haplotype; Haplotypes - genetics; Heterozygote; HLA; HLA Antigens - genetics; Humans; Linkage disequilibrium; Linkage Disequilibrium - genetics; Next generation sequencing; Nuclear Family; Pedigree; Software; GL; HML; Linkage disequilibrium; CSV; Next generation sequencing; IHIW; NMDP; 17th International HLA and Immunogenetics Workshop; STR; CWD; IMGT; SNP; NGS; HLA; IPD; MIRING; Haplotype
• ISSN: 0198-8859; 1879-1166
• DOI: 10.1016/j.humimm.2019.01.010

The highly polymorphic classical human leukocyte antigen (HLA) genes display strong linkage disequilibrium (LD) that results in conserved multi-locus haplotypes. For unrelated individuals in defined populations, HLA haplotype frequencies can be estimated using the expectation-maximization (EM) method. Haplotypes can also be constructed using HLA allele segregation from nuclear families. It is straightforward to identify many HLA genotyping inconsistencies by visually reviewing HLA allele segregation in family members. It is also possible to identify potential crossover events when two or more children are available in a nuclear family. This process of visual inspection can be unwieldy, and we developed the “HaplObserve” program to standardize the process and automatically build haplotypes using family-based HLA allele segregation. HaplObserve facilitates systematically building haplotypes, and reporting potential crossover events. HLA Haplotype Validator (HLAHapV) is a program originally developed to impute chromosomal phase from genotype data using reference haplotype data. We updated and adapted HLAHapV to systematically compare observed and estimated haplotypes. We also used HLAHapV to identify haplotypes when uninformative HLA genotypes are present in families. Finally, we developed “pould”, an R package that calculates haplotype frequencies, and estimates standard measures of global (locus-level) LD from both observed and estimated haplotypes.