The p-value heatmap (Figure 3B) indicates that most gene trees showed divergent histories (depicted as groups of topologies with dissimilar p-value patterns) that are most likely due to horizontal gene transfer.
Support for the citrus 2 lineage was found in most gene trees when GCPSR criteria was applied.
Because species monophyly was supported by most gene trees, a quantitative assessment of the number of interspecies lateral gene transfer (LGT) per gene could be performed using the species tree as a control (supplementary fig. S2, Supplementary Material online).
In addition to the algal and bacterial affiliations, anomalous relationships among other taxa can be observed in most gene trees in our analyses, where multiple eukaryotic sequences sporadically branch with prokaryotic homologs.
We provide evidence that the evolutionary splits among primates, rodents and artiodactyls happened shortly after each other, with most gene trees favouring a topology with rodents as outgroup to primates and artiodactyls.
The two Lissotriton species are closely related; due to ongoing hybridization and incomplete lineage sorting, they share a large number of polymorphisms, and in most gene trees the species are not reciprocally monophyletic (Zieliński et al. 2013; Pabijan M, Zieliński P, Babik W, unpublished data).
Like most gene tree/species tree methods, BCA assumes topologically homogeneous loci, that is, that all sites within a given locus evolved under the same underlying tree topology.
Recently, however, it was shown that with 4 taxa the most probable gene tree does not necessarily reflect the species tree, if the species tree is unbalanced [ 16].
Furthermore, the most probable gene tree can have a different topology from that of the species tree if multiple branch lengths are small enough in coalescent units [ 16].
However, even in the presence of lineage sorting, for 3 species the most probable gene tree represents the species tree [ 16].
