For this analysis, we used as input a set of OGs generated by

the DomClust program [24] (see selleck inhibitor “”Phylogenetic profile analysis”" section below for details about identification of OGs by DomClust). Absence of a gene in some genomes (at least half of the genomes) in each OGs among the core is allowed. In addition, as identified OGs are at the domain level, if a counterpart of a gene in one genome is split in another genome, different number of genes can participate in the OGs in different genomes. Thus, the number of core genes in each genome can vary. Still, the numbers of core genes varied less (1364-1424; SD = 13.5) than the total number of genes among the strains (1465-1593; SD = 33.9) (Table 1). Among those core OGs, 1079 OGs were universally conserved (conserved in the all genomes),

non-domain-separated, with one-to-one correspondence, and designated “”well-defined core OGs”". Those 1079 OGs were used for phylogenetic analysis (Figure 1). Nucleotide sequences of genes in well-defined core OGs were aligned by the Mafft program [128], click here from which conserved blocks were extracted by the Gblocks program [129]. Phylogenetic profile analysis Phylogenetic profiling was carried out using the set of OGs generated by DomClust [24]. We identified OGs with East Asian-specific features as those whose phylogenetic profiles were highly correlated to the template pattern (taking 1 for find more hspEAsia and 0 for hpEurope). The DomClust clustering program can identify OGs at the domain level, and was used to identify genes truncated in particular strains. Clustering was performed based on PAM (point accepted mutation) distance rather than score to ensure proper evaluation of evolutionary distances, even if one gene was truncated; in the latter case, scores may underestimate evolutionary relatedness. To clarify differences 3-oxoacyl-(acyl-carrier-protein) reductase in gene-splitting patterns among strains, we did not use DomClust options to suppress domain splitting. To identify genes with characteristic

patterns of hspEAsia strains, we constructed a phylogenetic profile for each OG as a vector of examined property values (e.g., number of domains or number of duplications). For surveying patterns of gene splitting and deletion, a phylogenetic profile was constructed for each OG using the number of domains for each gene that resulted from the clustering. For surveying patterns of gene duplication, a phylogenetic profile was constructed using the number of duplicated genes (in-paralogs). To find OGs with a characteristic hspEAsia pattern, equality of the medians among different populations was tested by Kruskal-Wallis test. Tests between East Asian and European strains used the six hspEAsia strains and the seven hpEurope strains.