Figure 3.

Multispecies protein alignment for IRX3. Coding sequence was extracted from whole-genome sequence alignments available from the UCSC Genome Browser and trimmed to include only coding regions based on matches with the human CDS. Alignment columns with gaps in the human sequence were excluded. Residues 1–180 of 501 are shown. Residues 36 and 44 were predicted to be positively selected in human with Bayes Empirical Bayes posterior probability of 0.886 and 0.997, respectively (Figure 4B).