Matters Arising: Re-examining the correlations between codon usage and dihedral bond angles using a population genetics model

Abstract

Codon usage bias (CUB), or the non-uniform usage of synonymous codons, has been observed across all domains of life (1). CUB is driven by a combination of both non-adaptive (e.g. mutation biases) and adaptive (e.g. natural selection for translation efficiency/accuracy) evolutionary forces (2; 3; 4; 5; 6). Empirical work has shown that changes to synonymous codon usage can affect co-translational protein folding and various computational studies have sought to determine if there is a general connection between codon usage and protein structure (see (7) for a review). In a recent manuscript, (8) explored the relationship between synonymous codon usage and the dihedral bond angles that form a protein’s backbone. Using a method they developed to compare dihedral bond angle distributions across synonymous codons, they detected statistically significant differences between the dihedral bond angle distributions of synonymous codons within theE. coliproteome. Although they rightly note that correlation does not imply causation, they hypothesize that differences in dihedral bond angle distributions between synonymous codons could be due to differences in elongation speed (see Figure 5 in (8)). Here, we present results using simulated data suggesting the findings of (8) may be a statistical artifact due to failure to control for a significant factor shaping a gene’s synonymous CUB: gene expression.