Viral reference genomes to disentangle the recombinant phylogenetic history of the potyviruses

Potyviruses are a large genus of plant-infecting RNA viruses in the family
Potyviridae. Due to their rapid diversification and frequent
recombination, reconstructing the phylogenetic history of the potyviruses
has proven difficult. Phylogenies reconstructed from different
protein-coding regions of the viral genome often reveal conflicing or
discordant relationships. But the extent to which discordance is due to
interspecific recombination versus phylogenetic noise or errors in
reconstruction is unclear.     To explore the
recombinant history of the potyviruses, we assembled a dataset containing
referece genomes for 131 species of potyviruses. High-quality, full-length
reference genomes for all species were obtained form NCBI GenBank. Viral
genomes were carefully aligned at the codon-level and screened for
recombination. The full alignment was then partitioned into several
sub-alignments between each detected recombination event, such that each
sub-aligment corresponds to a non-recombinant block (NRB) free of detected
recombination events. Local phylogenetic trees for each NRB were then
reconstructed to explore how phylogenetic relationships varied across
different regions of the potyvirus genome.   We then
used our program Espalier to disentangle the phylogenetic history of the
potyviruses. Espalier reconciles and removes discordances between
phylogenetic trees that are likely attributable to phylogenetic error
while retaining recombination events that are strongly supported by the
sequence data. Applying Espalier to the potyviruses revealed that most
phylogenetic discordace between local trees is likely attributable to
phylogenetic noise. Removing the discordance attributable to phylogenetic
error allows us to much more clearly visualize the phylogenetic history of
the potyviruses.