Molecular criteria for genus and species discrimination within the family Potyviridae

A phylogenetic analysis of the optimised nucleotide (nt) alignment of the entire ORFs of a representative of each fully-sequenced species in the family Potyviridae provided strong support for several subgroups within the genus Potyvirus. A complete set of two-way comparisons was done between the sequences for the entire ORF and for each gene amongst all the 187 complete sequences from the family. Most species had 50-55% nt identity to other members of their genus in their ORFs but there were significant groups of more closely related species and species demarcation criteria were <76% nt identity and <82% amino acid identity. The corresponding thresholds for species demaracation using nt identity values for the individual genes ranged from 58% (P1 gene) to 74-78% (other genes) although a few comparisons between different species exceeded these values. For the entire ORF, genus demarcation criteria were <46% nt identity but this did not separate rymoviruses from potyviruses. Comparisons in the CI gene most accurately reflected those for the complete ORF and this region would therefore be the best for diagnostic and taxonomic studies if only a sub-portion of the genome is to be sequenced. Further comparisons were then made using all the 1220 complete capsid protein (CP) genes. These studies suggest that 76-77% nt identity is the optimal species demarcation criterion for the CP. The study has also helped to allocate the correct virus name to some sequences from the international databases that currently have incorrect or redundant names. The taxonomic status of the current genus Rymovirus and of three unassigned species in the family is discussed. Significant discontinuities in the distributions within and between the currently defined species suggest that the continuum of variation that is theoretically available is constrained or disrupted by molecular barriers that must have some biological significance.