Inter- and Intra-Island Genetic Diversity in Adriatic Populations of Croatia: Implications for Studying Complex Diseases in Isolated Populations

TLDR: The high mutation rate at these STR loci diluted the effect of small effective size in producing correlation of alleles within individuals, and the non-significant gene differentiation between villages within the islands is an advantage for disease-gene association studies using microsatellite loci.

Abstract: Adriatic populations residing in over 7 islands of Croatia have been extensively studied for over the past 30 years. Linguistic, social, physical and biological anthropological data indicate that the island populations have a considerable degree of genetic isolation from each other. Within the islands as well, some degree of isolation exists between villages of several islands. To examine the effects of such genetic isolation on patterns of gene diversity at microsatellite loci, this investigation studied the coefficient of gene differentiation (GST) between villages within islands as well as between islands using allele frequency data at 9 short tandem repeat loci (D3S1358, vWA, FGA, THO1, TPOX, CSF1PO, D5S818, D13S317, and D7S820) in 31 villages of 4 islands (Hvar, Krk, Brac, and Korcula). Overall, none of the villages exhibits any significant departure from Hardy-Weinberg expectations of genotype frequencies at these loci. Decomposition of heterozygosity (H) as well as allele size variance (V) indicated significant inter-island gene differentiation (GST based on H = 0.0032, P < 10-4 ; GST based on V = 0.0036, P = 0.007), while the between villages within island component of genetic variation remain non-significant (GST(H) = 0.0029, P = 0.073 ; GST(V) = 0.0018, P = 0.315). The coefficient of coancestry (q) for the sampled unrelated individuals vary from village to village, ranging from 0.0 to 0.022, with an average of 0.0098. Comparison of estimates of inbreeding coefficient (F) based on genealogical data from these villages indicates that while q and F values are moderately correlated across villages, q is comparatively smaller (nearly one-half) than F. In other words, the high mutation rate at these STR loci diluted the effect of small effective size in producing correlation of alleles within individuals. The non-significant gene differentiation between villages within the islands is an advantage for disease-gene association studies using microsatellite loci, since data from several villages within each island may be pooled to enlarge the sample size without introducing complications of population substructure effects on the association results. (Researh supported by NIH grants GM 41399, GM 52601, and GM 58545).