
In the standard genetic code, most amino acids are indeed encoded by two or more codons, with only methionine and tryptophan having a single codon. The degeneracy of the standard genetic code was postulated by Crick et al. Findings from this communication can be utilized for the optimization of gene expression.

We propose that readthrough efficiency and context effects could explain the prevalence of UAA over UAG, particularly in highly expressed genes.

In contrast, increased GC content results in a decrease in UAA abundance with a concomitant increase in UGA abundance. The strength of the selective bias against UAG in highly expressed genes varies among bacterial genomes, but it is not affected by the GC content of these genomes. Of the 253 highly expressed genes, only two end with an UAG codon. In contrast, UAG (amber) is the least frequent termination codon, e.g., only 321 occurrences (7.4%) in E. The occurrence of the three translation termination codons is highly biased, with UAA (ochre) being the most prevalent in almost all bacteria. Here, we analyze the frequencies of stop codons in a group of highly expressed genes from 196 prokaryotes under strong translational selection.

Thus, codon frequencies differ across organisms, including the three termination codons in the standard genetic code. The genome hypothesis postulates that genes in a genome tend to conform to their species’ usage of the codon catalog and the GC content of the DNA.
