Figure 5

Subset assignments for the sites from each codon position using the iterative k -means algorithm clustering using TIGER site rate estimates on the Kawahara 2013 dataset. Each row corresponds to a single gene and each column corresponds to a different codon position. The dotted red line represents the total number of sites in each codon position. In each chart, subsets are ordered by the mean site rate from slowest to fastest. First and second codon positions most closely align with “traditional partitioning”, while substantial variation exists among the 3^rd codon position sites.