Analysis of the HLA population data (AHPD) submitted to the 15th International Histocompatibility/Immunogenetics Workshop by using the Gene[rate] computer tools accommodating ambiguous data (ahpd project report)

Abstract

During the 15th International Histocompatibility and Immunogenetics Workshop (IHIWS), 14 human leukocyte antigen (HLA) laboratories participated in the Analysis of HLA Population Data (AHPD) project where 18 new population samples were analyzed statistically and compared with data available from previous workshops. To that aim, an original methodology was developed and used (i) to estimate frequencies by taking into account ambiguous genotypic data, (ii) to test for Hardy-Weinberg equilibrium (HWE) by using a nested likelihood ratio test involving a parameter accounting for HWE deviations, (iii) to test for selective neutrality by using a resampling algorithm, and (iv) to provide explicit graphical representations including allele frequencies and basic statistics for each series of data. A total of 66 data series (1-7 loci per population) were analyzed with this standard approach. Frequency estimates were compliant with HWE in all but one population of mixed stem cell donors. Neutrality testing confirmed the observation of heterozygote excess at all HLA loci, although a significant deviation was established in only a few cases. Population comparisons showed that HLA genetic patterns were mostly shaped by geographic and/or linguistic differentiations in Africa and Europe, but not in America where both genetic drift in isolated populations and gene flow in admixed populations led to a more complex genetic structure. Overall, a fruitful collaboration between HLA typing laboratories and population geneticists allowed finding useful solutions to the problem of estimating gene frequencies and testing basic population diversity statistics on highly complex HLA data (high numbers of alleles and ambiguities), with promising applications in either anthropological, epidemiological, or transplantation studies.

Retraction: Ana Strezova. Genotyping of HLA-A, HLA-C and HLA-B Loci with Reverse Line Strip Hybridization in Macedonian Population and Association of HLA Class I Alleles with Spondilitis [MSc thesis]

“…In the "Population comparisons" of the paper African, American, and European populations (including Macedonian population) were compared [5].…”

“…HLA DNA genotyped results from our laboratory were included in the analysis of the HLA population data (AHPD) submitted to the 15th International Histocompatibility/Immunogenetics Workshop by using the GENE[RATE] computer tools accommodating ambiguous data (AHPD project report) and are published [5].…”

“…Furthermore, there are a lot of significant differences between the results for Macedonian population analysed and published by the 15th International Histocompatibility and Immunogenetics Workshop [5] and the results in retracted MSc thesis [6]. These differences had tremendous impact on the frequencies of the HLA-A, HLA-B, and HLA-C alleles with erroneous frequencies in the retracted MSc thesis.…”

Allele Frequency Estimation from Ambiguous Data: Using Resampling Schema in Validating Frequency Estimates and in Selective Neutrality Testing

“…The method used here is an improvement of the procedure implemented in GENE[RATE] (Nunes 2004(Nunes , 2007Nunes et al 2010), which is an adaptation of the classical EW test to ambiguous data. It consists in estimating the p values of the usual EW test for nonambiguous random samples and in using the distribution of these p values to assess selective neutrality.…”

“…The two methods produce different genotyping ambiguities at each locus, and PCR-SSP reveals fewer ambiguities than SSO-Luminex. Allele frequencies were estimated for both methods and each of the three loci by using the GENE[RATE] implementation of the ExpectationMaximization (EM) algorithm (Dempster et al 1977;Nunes 2005), an extension capable of accommodating all kinds of ambiguous data (Nunes 2005(Nunes , 2007Nunes et al 2010). To test the equivalence of the two estimations, we developed an original resampling procedure to compare the values of the frequency estimates.…”