Gene-to-gene correlation of phylogenetic profiles of population structure in CSF and SWAB HSV-1 genomes

The phylogenetic relationships between consensus genomes were dissected for each genelocus using the bayesbipartprofile tool (https://github.com/flass/genomescans) (Lassalle et al., 2017).<br>Bayesian gene tree samples were computed for each genes with Mrbayes v3.2.8 (2,000,000 generations, sampling every 500 trees, discarding the first 25% as burn-in).<br>For each gene locus, phylogenetic compatibility of tree splits with the alignment was computed for eachsplits observed in at least one Bayesian gene tree sample at a minimum frequency of 0.1 andseparating clades of at least 3 sequences. <br><br>The gene-to-split compatibility scores were thenused to compute correlation (r²) between gene pairs, displayed as a heatmap in file "test_longLD_bipart.pdf"; genes areordered following their position in the genome. <br>Genes physically clustered together showsignificant correlation of phylogenetic structure (i.e. related history of descent), but notthose located far apart in the genome. Elevated compatibility of UL45 with most other genesindicates limited phylogenetic information available in this gene’s alignment.<br><br>We also report the phylogenetic support for grouping of pairs of strains as closest relatives across the genome in the file "top_long_tracks.pdf". This file presents “iceberg” plots for the most clonally related strain genome pairs. In these plots, the top panel shows the posterior probability support in individual gene trees across the genome (following reference strain 17 gene order), and the bottom panel (with an inverted y-axis) shows the phylogenetic compatibility index, i.e. the proportion of topologies in each gene tree sample that do not reject the grouping; the compatibility metric is able to reveal hidden signals of strain pairing, hence the iceberg analogy.