A comparison between dS estimates and phylogeny-based relative dating in several independent sources of phylogenetic trees. Boxplots represent the distribution of dS values of paralogs assigned to different phylogenetic categories. dS estimates correspond to the Maximum Likelihood method as implemented in codeml (see Section 2). Other methods (NG, YN) were evaluated, obtaining very similar results (data not shown). Horizontal dashed lines indicate the dS saturation threshold (dS = 1.5). X-axis indicates the granularity of the phylogenetic categories. Each panel evaluates a different seed species or tree source: (A) Human paralogs from ensembl compara trees; (B) Human paralogs from phylomeDB trees; (C) Mouse paralogs from ensembl compara trees; (D) Pufferfish from ensembl compara trees; and (E) Yeast from phylomeDB trees.