Tests for selection and recombination

In this study, only ARP ortholog sets that were detected in at least 50 strains per MABC subspecies were analyzed. Codon alignments of orthologous sequences were generated using PAL2NAL software version 14.0 [23], with protein alignments obtained in MUSCLE. To eliminate the impact of recombination on positive selection evaluations, PhiPack version 1.0 [24] was used to infer the presence of recombination in the ARPs. Three recombination-detection statistics: pairwise homoplasy index (PHI), maximum χ² (MaxChi²) and neighbor similarity score (NSS) were implemented to test for signatures of recombination. In all three tests, recombination was considered significant if the p-value was less than 0.1.

To reveal the selection pressures, Single-Likelihood Ancestor Counting (SLAC) [25], Fixed Effects Likelihood (FEL) [25], Fast Unconstrained Bayesian AppRoximation (FUBAR) [26] and Mixed Effects Model of Evolution (MEME) [27] models which were implemented in Hyphy package [28] were used. SLAC was carried out for the global estimation of the ratios of non-synonymous (dN) to synonymous (dS) substitution rates (ω = dN/dS) of the ARPs. The ARPs with ω < 1 and ω > 1 were inferred to have evolved under purifying selection and diversifying selection, respectively. FEL, FUBAR, MEME and SLAC were also used to detect the sites under diversifying or purifying selection. A significance level of posterior probability (PP) greater than 0.9, p-value less than 0.05 and ω < 1 or ω > 1 were considered as indicating sites under purifying and diversifying selection, respectively. A codon was considered to have undergone purifying or diversifying selection only if it was detected by at least two of the methods implemented.