CpG Usage in RNA Viruses: Data and Hypotheses

CpG repression in RNA viruses has been known for decades, but a reasonable explanation has not yet been proposed to explain this phenomenon. In this study, we calculated the CpG odds ratio of all RNA viruses that have available genome sequences and analyzed the correlation with their genome polarity, base composition, synonymous codon usage, phylogenetic relationship, and host. The results indicated that the viral base composition, synonymous codon usage and host selection were the dominant factors that determined the CpG bias in RNA viruses. CpG usage variation between the different viral groups was caused by different combinations of these pressures, which also differed from each other in strength. The consistent under-representation of CpG usage in −ssRNA viruses is determined predominantly by base composition, which may be a consequence of the U/A preferred mutation bias of −ssRNA viruses, whereas the CpG usage of +ssRNA viruses is affected greatly by their hosts. As a result, most +ssRNA viruses mimic their hosts' CpG usage. Unbiased CpG usage in dsRNA viruses is most likely a result of their dsRNA genome, which allows the viruses to escape from the host-driven CpG elimination pressure. CpG was under-represented in all reverse-transcribing viruses (RT viruses), suggesting that DNA methylation is an important factor affecting the CpG usage of retroviruses. However, vertebrate-infecting RT viruses may also suffer host' CpG elimination pressure that also acts on +ssRNA viruses, which results in further under-representation of CpG in the vertebrate-infecting RT viruses.