Comparison of the gene phylogenies present in the original dataset. Panel A gives the distribution of AU-test results. The AU-test was applied to the original dataset, and the distribution of AU-test results is depicted as a function of the significance level. Histogram bars give the number of all gene families that were different from the consensus tree (Figure 1) at the indicated significance level. Panel B gives data for the Symmetric Difference of Robinson and Foulds . Tree distances were calculated between the consensus tree and each individual gene tree. The mean distance between gene tree and consensus is 3.37 with the standard deviation of 2.32. Panel C gives the number of edits (SPR events corresponding to HGTs) calculated according to . To illustrate that the results do not depend on a particular reference tree, we used three different trees: Tree 1 is the consensus tree depicted in Figure 1, tree 2 and 3 are the trees that together with tree 1 were reported as showing the least amount of conflict with the individual gene phylogenies (trees 2 and 3 in Figure 2 of ).