Genome-Wide Selection Signature Scan Using iHS

Using the phased BBJ GWAS haplotype data, we calculated genome-wide natural selection signatures based on iHS (Voight et al. 2006; Johnson and Voight 2018) separately for each short or long chromosome arm, using the multithread computing option of selscan (--threads, version 1.1.0b; Szpiech and Hernandez 2014). The genome-wide iHS z-scores were standardized through normalization within each derived allele frequency bin (bin widths = 0.01). The variants in the previously known loci with selection signatures in Japanese or Asians were also excluded in the process of normalization fitting (Hirayasu et al. 2008; Liu et al. 2017, 2018; Chiang et al. 2018; Okada et al. 2018). We estimated two-tailed P values of the SNP according to the normalized z-scores (=P_iHS). We set the significance threshold by considering the Boferroni correction based on the number of the assessed SNP (α = 0.05). Summary statistics of the SDS selection signals in Japanese were obtained from the previous study (Okada et al. 2018).