All proteins for comparison were multiply aligned using clustalw, and alignment columns with more than 30% gaps were removed to reduce the effects of poorly aligned regions.
The aligned core is comprised of alignment columns that contain all input genomes and no gaps.
Only those alignment columns that could be aligned to the structure sequence were used in our subsequent analyses.
All alignment columns containing ambiguities and gaps were excluded during the PAML analysis.
Since some functional regions contain unconstrained nucleotides, GERP allows for RS scores to be merged across unconstrained alignment columns.
In total, 861 domains are annotated, covering 6.34% of the amino acid residues and 6.69% of the protein sequence alignment columns in these single-copy protein sets.
Charge compensatory substitutions among the coevolved sites were investigated by calculating the frequency of charge compensatory substitution quads for each pair of coevolved alignment columns.
Occurrence of charge compensatory substitutions was investigated by calculating the frequency of charge compensatory substitution quads for each pair of coevolved alignment columns.
For each alignment, columns with more than 80% gaps were eliminated and the average nucleotide composition of the remaining columns from all the alignments was calculated.
The sum of deviation scores in consecutively constrained alignment columns provides a metric of evolutionary constraint, termed a rejected substitution (RS) score.
An IR is a maximal sequence of alignment columns having at least one gap character each and flanked by gapless columns.
