Two possible cases of dS-based dating misestimations. Duplication nodes and paralogous sequences used in the analysis are marked in blue. The estimated topological relative age is shown at the base of duplication nodes. Branch support values, based on aLRTs, are indicated in red. dS estimation by different methods are shown below each tree (ML = Maximum Likelihood, YN = Yang and Nielsen, NG = Nei and Gojorobi). (A) Phylogenetic relationship between two primate-specific taste receptors, TAS2R9 and TAS2R8. As shown in the tree, a specific duplication affecting only the primate lineage is detected with high support by the relative topological dating approach. Similarly, the duplication leading to the Beta Defensin precursors DEFB127 and DEFB128 (B) is dated to mammals by the same method. Surprisingly, dS estimation for primate-specific TAS2R9/8 paralogs is higher than that of mammalian-specific DEFB127/128 paralogs. Similarly, the duplication of the homeobox proteins TGIF2LX and TGIF2LY (C) are dated by a topological approach in the human lineage. However, their dS rate is much greater than that of the pair MRGPRX3/MRGPRX4 (D), which is a highly supported primate-specific duplication. It must be noted that the phylogenetic distribution of the relevant paralogs in all these cases would support the given relative topological dating. All duplication events are supported by Maximum Likelihood analyses as implemented in the PhylomeDB phylogenetic pipeline and by Ensembl paralogy predictions. In the case of the human- specific duplication in tree C, further references supporting a human-specific duplication can be found in Blanco-Arias et al. (2002) and Skaletsky et al. (2003).