In order to elucidate the genetic mechanism of the association between IDDM and HLA, the ratio of epistatic variance to genotypic variance (VI/VG) was estimated on the assumption of linkage equilibrium between IDDM and HLA. Here, the allele responsible for IDDM was assumed to be dominant, and the penetrance of genotype for IDDM was considered. Using the same data as the present study, Furusho et al. (1982) undertook a similar analysis on the assumption that the allele responsible for IDDM was recessive, but some errors were made in the calculations. Thus, they were corrected in the present study.
It was found that the ratio (VI/VG) was 1/2.3 between IDDM and HLA-Bw54, 1/2.7 between IDDM and HLA-DYT, and 1/3.5 between IDDM and HLA-DRw 4, irrespective of the penetrance of the genotype for IDDM, whether the allele for IDDM was dominant or recessive. These estimates suggest the existence of epistatic interaction between loci but the difference of these estimates from zero could not be tested statistically.
It is very difficult to determine whether the association at the phenotype level is due to the linkage disequilibrium at the gamete level or to the epistatic interaction between loci.