Quantifying shifts in natural selection on codon usage between protein regions: a population genetics approach

Area of Science:

• Evolutionary biology
• Genomics
• Molecular biology

Background:

• Codon usage bias (CUB) is the non-uniform use of synonymous codons, influenced by selection for translation efficiency, accuracy, and protein folding.
• Previous studies on the relationship between CUB and protein structure yielded contradictory results, often due to methodological differences and unaddressed confounding factors like amino acid usage and gene expression.
• Understanding CUB's role in protein structure is crucial for deciphering protein function and evolution.

Purpose of the Study:

• To explicitly quantify codon-specific shifts in natural selection related to protein structure using a population genetics approach in *S. cerevisiae* and *E. coli*.
• To differentiate the effects of natural selection, mutation bias, and amino acid usage on codon usage in relation to protein structure.
• To re-evaluate previous findings on codon usage patterns at protein structural elements, such as helix termini.

Main Methods:

• Employed an explicit population genetics framework to analyze codon usage patterns.
• Separated the influences of natural selection (scaled by gene expression) and mutation bias.
• Utilized Bayesian model comparisons to assess selection variation across secondary structures (helix, sheet, coil) and between structured and intrinsically disordered regions.
• Used simulated data to test and correct for confounding factors in previous analyses.

Main Results:

• Found only slight variations in selection on codon usage across different secondary structures (helix, sheet, coil) and between structured and intrinsically disordered regions.
• Observed minimal variation in codon usage selection at helix termini in *E. coli*, contradicting some prior studies.
• Demonstrated that previous claims of 'non-optimal' codon enrichment at helix beginnings in *S. cerevisiae* were artifacts of uncontrolled confounding factors, not selection for cotranslational folding.

Conclusions:

• Established a weak relationship between codon usage and protein structure, with minor and idiosyncratic (codon-specific) selection shifts.
• Highlighted the power of an explicit evolutionary approach for studying selective shifts on genomic features.
• Discussed limitations and future research directions for understanding CUB and its structural implications.