Identification and analysis of conserved rare codon clusters (CRCCs).

<p>(<b>A</b>) Overview of the analysis pipeline. From the ORFeomes of 76 species, homolog families were assigned using OrthoMCL [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1005531#pcbi.1005531.ref041" target="_blank">41</a>]. Protein sequences within each family were aligned using MUSCLE [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1005531#pcbi.1005531.ref042" target="_blank">42</a>]. ORFeome coding sequences were analyzed to locate rare codon clusters, and peak locations were spliced into alignments. These alignments of peaks were used as input for co-occurrence analysis. To avoid false positives due to closely related paralogs, homolog families were filtered to contain a maximum of one sequence from each species. Co-occurrence analysis determines if rare codon clusters in the resulting homolog families align more than expected by random chance. Terminal trimming removed 5’ and 3’ sequence termini, to identify co-occurrence of non-terminal rare codon clusters. Random reverse translations (RRTs) control for rare codon co-occurrence for reasons unrelated to codon rarity (see text for details). (<b>B</b>) Conservation of rare codons within homolog families is widespread. White bars indicate homolog families with conservation of rare codons at any position; black bars indicate homolog families with significant co-occurrence in regions other than the 5’ and 3’ termini.