Genome architecture, GC3 content, codon usage, and predicted translational adaptation of Salmonella temperate and virulent phages

Introduction Salmonella enterica phages are important drivers of bacterial evolution and potential tools for phage-based applications. However, differences in genome architecture, codon usage, and predicted translational adaptation between Salmonella temperate and virulent phages remain uncertain and need to be determined. Methods This study elucidates genome-wide and functional-module-specific molecular signatures across Salmonella phage lifestyles. We analyzed 41 temperate and 41 virulent Salmonella phage genomes comprising 7,759 protein-coding genes. Because lifestyle is a genome-level attribute, statistical inference employed equal-genome-weight comparisons, whole-genome permutations, genome-cluster bootstrapping, and sensitivity analyses. Results The virulent phages exhibited greater broad-scale genomic disconnectedness (81.7% versus 74.0% of within-group genome pairs lacking significant BLAST similarity) and larger mean genomes (80,784 versus 44,433 bp). The temperate phages showed higher genome-average overall GC and GC3 content, with “virulent”-minus-“temperate” differences of −4.484 and −7.706 percentage points, respectively (both q = 0.00013). Conversely, the virulent phages exhibited higher genome-average Codon Adaptation Index (CAI-2; difference = +0.0131) and Index of Translation Elongation (ITE; +0.0321), with associations persisting after adjustment for the nucleotide composition and coding-sequence length. The genome-equal effective number of codons (Nc) difference was not significant ( q = 0.383). Genome-equal PERMANOVA detected significant differentiation across the complete molecular dataset ( R 2 = 0.1518; q = 0.0001) and all five gene functional categories, with the greatest effects in the DNA repair/recombination, lysis-associated, and structural genes. Only regulatory genes showed significantly different genome-level multivariate dispersion. Genome-average PCA scores differed significantly along GC-associated PC1 and a second axis associated with CAI-2, ITE, and sequence size. GC-adjusted CAI-2 -ITE correlations remained positive across both phage lifestyles (ρ = 0.771 and 0.776), with genome-cluster bootstrap confidence intervals excluding zero. Genome-level relative synonymous codon usage analyses identified 23 of 59 codons with FDR-significant, large-effect differences. Principal findings persisted in functional-annotation and genome-relatedness sensitivity analyses. Discussion Collectively, these results identify complementary compositional and predicted codon-adaptation signatures in the selected phage lifestyle groups, while taxonomic imbalance and the absence of direct translational measurements limit further interpretation.