Dinucleotide Composition in Animal RNA Viruses Is Shaped More by Virus Family than by Host Species

Abstract

Viruses use the cellular machinery of their hosts for replication. It has therefore been proposed that the nucleotide and dinucleotide compositions of viruses should match those of their host species. If this is upheld, it may then be possible to use dinucleotide composition to predict the true host species of viruses sampled in metagenomic surveys. However, it is also clear that different taxonomic groups of viruses tend to have distinctive patterns of dinucleotide composition that may be independent of host species. To determine the relative strength of the effect of host versus virus family in shaping dinucleotide composition, we performed a comparative analysis of 20 RNA virus families from 15 host groupings, spanning two animal phyla and more than 900 virus species. In particular, we determined the odds ratios for the 16 possible dinucleotides and performed a discriminant analysis to evaluate the capability of virus dinucleotide composition to predict the correct virus family or host taxon from which it was isolated. Notably, while 81% of the data analyzed here were predicted to the correct virus family, only 62% of these data were predicted to their correct subphylum/class host and a mere 32% to their correct mammalian order. Similarly, dinucleotide composition has a weak predictive power for different hosts within individual virus families. We therefore conclude that dinucleotide composition is generally uniform within a virus family but less well reflects that of its host species. This has obvious implications for attempts to accurately predict host species from virus genome sequences alone.IMPORTANCE Determining the processes that shape virus genomes is central to understanding virus evolution and emergence. One question of particular importance is why nucleotide and dinucleotide frequencies differ so markedly between viruses. In particular, it is currently unclear whether host species or virus family has the biggest impact on dinucleotide frequencies and whether dinucleotide composition can be used to accurately predict host species. Using a comparative analysis, we show that dinucleotide composition has a strong phylogenetic association across different RNA virus families, such that dinucleotide composition can predict the family from which a virus sequence has been isolated. Conversely, dinucleotide composition has a poorer predictive power for the different host species within a virus family and across different virus families, indicating that the host has a relatively small impact on the dinucleotide composition of a virus genome.KEYWORDS dinucleotide bias, evolution T he nucleotide composition of genomes is not uniform. This is clearly manifest in the variation in dinucleotide frequencies between organisms, which do not match the frequencies predicted from overall nucleotide composition. Most notably, the frequency of the TpA dinucleotide is suppressed across all animal genomes, with the

A decade of RNA virus metagenomics is (not) enough

“…Other than the retroviruses, RNA viruses do not readily provide host linkage information in metagenomics like DNA viruses and phages often do. Linking RNA virus and host has been imputed based on transcript levels or dinucleotide usage, although the latter has recently been shown to more correlated with viral family than host species (Kapoor et al, 2010;Giallonardo et al, 2017;Shi et al, 2016). Even when sequencing discrete organisms such as a singular honeybee, the co-existence of eukaryotic parasites means that imputation of viral host cannot be exact (Runckel et al, 2011).…”

ΦX174 Attenuation by Whole Genome Codon Deoptimization

“…However, there are many other compositional features within the DNA sequences of protein coding genes upon which selection acts. These features include: the genic GC content (4-6) , CpG or TpA dinucleotides (7)(8)(9)(10)(11) , codon pairs (12)(13)(14)(15) , endonuclease recognition sites (16)(17)(18)(19) , intron splicing motifs (20) , mRNA folding stability (21)(22)(23)(24)(25) , ribosomal pausing sites (26,27) , concentration of non-preferred codons at transcripts 5' end (28)(29)(30) , autocorrelation of codons on transcript (31) , and capacity of codon order to influence co-translational folding of proteins (32,33) . At any given site, natural selection may act on one or more of these features to favor the use of certain codons over other synonymous ones (34) .…”

Comparative genome analysis reveals a distinct influence of nucleotide composition on virus–host species‐specific interaction of prawn‐infecting nodavirus

“…The subsequent codon usage bias analyses of the nodavirus in this study were narrowed down to focus on beta‐ and gammanodavirus due to its significant impact on the cultured aquatic animal. The RSCU comparison of nodavirus and their hosts clearly showed the heterogeneous composition of mono‐ and dinucleotide among all strains, but reflected poorly on its host species especially gammanodavirus (Giallonardo, Schlub, Shi, & Holmes, ; Kumar et al, ). However, both viral genome and host mRNA are anticipated to share some common features due to constraint induced by host factors and their coexistence in the same cytoplasmic cellular environment (Kapoor et al, ).…”