A comprehensive computational analysis investigating the relationships between phage codon usage, infection style,...

It has been known for decades that bacteriophages encode tRNA genes, but their function and the factors contributing to their acquisition and retention are unclear. Although tRNAs are found in a variety of phages infecting a variety of bacteria, many large-scale computational studies investigating tRNA acquisition and retention in phages are specific to Mycobacterium phages; however, these findings may not be representative of other phages or bacteria. This work uses a broader sampling of phages and hosts to investigate the relationships between codon usage bias, infection cycle, and tRNA gene numbers in phage genomes. We analyzed 154 phages infecting 7 host genera, including Gram-negative (Escherichia, Shigella, Salmonella) and Gram-positive (Bacillus, Lactobacillus, Staphylococcus, Mycobacterium) bacteria. Phages included temperate and virulent representatives, plus a range of tRNA numbers and morphologies. All phages and hosts were analyzed using four metrics: GC content, Effective Number of Codons, Relative Synonymous Codon Usage, and tRNA Adaptation Index. On a global scale, virulent phages with many tRNA genes show greater differences in codon usage and codon adaptation compared to their respective hosts. Gram-negative bacteria and their phages generally exhibit greater differences in codon usage compared to Gram-positive bacteria and their phages. Phages infecting Gram-negative hosts also tend to encode more tRNA genes. In nearly all genus-level comparisons, Mycobacterium phages were different from any other host and from global patterns. This suggests previous computational studies performed in Mycobacterium phages are likely not applicable on a global scale or to phages infecting other host genera.