Selection pressure
We used a Bayesian approach, as incorporated in BayeScan (Foll & Gaggiotti 2008) to estimate candidate loci (island of differentiation) under natural selection between the Japanese and European populations. The approach estimates a posterior probability for each locus being under selection based on the allele frequencies of the two populations being significantly more or less divergent than the total shared allele frequencies of all loci in the dataset. The method incorporates uncertainties of small population sizes and varying effective population size, thus appropriate for our study. We analyzed the combined SNP set of Japanese and European isolates where the two populations were compared, using default settings. The resulting files were analyzed in R, after formatting with PGDSpider (Lischer & Excoffier 2011).
Signatures of selective sweeps can be investigated both within population using the integrated haplotype homozygosity (IHH) or between populations applying the population extended haplotype homozygosity (XP-EHH) (Voight et al. 2006). Considering that IHH is dependent on knowledge of the ancestral state, we focused on the XP-EHH approach to detect selective sweeps in which the selected allele has approached fixation within one population. The analysis was done using the R program rehh (Gautier & Vitalis 2012). Since calculation of XP-EHH is dependent on knowledge of haplotypes, the data was first phased using fastPHASE v.1.4 with default settings (Scheet & Stephens 2006).
All significant loci in predicted genes were further analyzed for function by InterProScan, using version InterPro 83.0 (www.ebi.ac.uk/interpro/). InterPro and PFAM domains were used for evaluation of gene function.