Traditionally one can contrast HLA frequencies in cases vs controls. Another way to do it, and which mitigates the issue of population stratification, would be to contrast an affected child's HLA with that of his mother and father. One could do blunt frequencies in parents vs. affected children. Wouldn't it be a lot better to look at each trio and for each HLA type at each locus, count the number of transmissions and the number of untransmissions in one's sample of subjects. That is what the transmission disequilibrium test (TDT) is about. PLINK is software that can do that. However, that was designed for biallelic or at most quadraallelic variance. HLA types can be multi-allelic. Do you know an existing algorithm (preferentially implemented in R) that can do that? If I have to, I will write the code myself? Should I? What landmines await me?
