Mapping the Inter- and Intra-genic Codon Usage Landscape in Homo sapiens

The redundancy of the genetic code permits multiple codons to specify a single amino acid, but codon selection is far from random and reflects underlying functional significance. Most lifeforms exhibit the phenomenon of codon usage bias, where the usage of certain codons is favored over others and hence used more often throughout the genome. Codon bias is thought to provide a layer of post-transcriptional regulation that shapes mRNA stability, translation kinetics and co-translational protein folding. Despite extensive characterization of codon bias in bacteria, viruses and various model organisms, its prevalence and functional consequences remain understudied in humans. Accordingly, this study seeks to (1) characterize the genome-wide distribution of key codon-bias metrics in humans, (2) investigate how codon usage correlates with protein structural features, and (3) assess whether genes with exceptionally restricted codon repertoires are selectively constrained against synonymous mutations. Our study finds that codon usage bias is pronounced in structured protein domains relative to intrinsically disordered regions (IDRs) (Cliff’s Δ = 0.26; p < 2.2 × 10 ^-16 ); that the most frequently used codons are supported by higher tRNA gene copy numbers (r = 0.34; p < 0.05); and that genes exhibiting extreme codon bias (high ΔENC) harbor significantly fewer synonymous variants than expected (r = – 0.28, p < 2.2 × 10 ^-16 ). Collectively, our findings position codon usage bias as a central, evolutionarily conserved regulator of translational efficiency, protein folding, and genomic integrity in humans.

This offers a new framework for interpreting synonymous variation and guiding future efforts in functional genomics.