Positions of CRCCs relative to protein domains.

<p>(<b>A</b>) Comparison of the actual and expected frequency of CRCCs near (within +/- 20 codon positions) of domain boundaries. Expected frequency is based on random distribution, and p-values are calculated using a binomial test. (<b>B</b>) The majority of CRCCs in human and E. coli are found within mRNA regions encoding conserved structural domains. (<b>C</b>) Secondary structure classes of CATH domains in human and E. coli ORFeomes. CATH domains annotated by Gene3D were divided by secondary structure class. Graphs indicate the percent of domains that belong to each secondary structure class, for proteins in homolog families with (black) or without (white) conserved rare codon clusters (CRCC) (co-occurrence p-value ≤ 1E-4 after termini and RRT screening). Although human proteins from homolog families with CRCCs are significantly over-enriched in β-sheet-rich domains, this enrichment is most likely due to the longer average length of proteins with CRCCs, as similar results were seen for a control set of proteins with similar length but no CRCCs (gray).