Codon usage bias in human cytomegalovirus and its biological implication.

Abstract Human cytomegalovirus (HCMV) infection, a worldwide contagion, causes a serious disorder in infected individuals. Analysis of codon usage can reveal much molecular information about this virus. The effective number of codon (ENC) values, relative synonymous codon usage (RSCU) values, codon adaptation index (CAI), and nucleotide contents was investigated in approximately 160 coding sequences (CDS) among 17 human cytomegalovirus genomes using the software CodonW. Linear regression analysis and logistic regression were performed to explore the preliminary data. The results showed that, overall, HCMV genomes had low codon usage bias (mean ENC = 47.619). However, the ENC of individual CDS varied widely and was distributed unevenly between host-related genes and viral-self-function genes ( P  = 0.002, odds ratio (OR) = 3.194), as did the GC content ( P  = 0.016, OR = 2.178). The ENC values correlated with CAI, GC content, and the nucleotide composing at the 3rd codon position (GC3s) ( P