The ancestor of the Bacteria domain was a hyperthermophile

Brochier and Philippe (Nature, 417, 244) have recently re-analysed the phylogeny of ribosomal RNA using only multiple alignment positions with no phylogenetic noise. They conclude that the first branch of divergence in the Bacteria domain comprises Planctomycetales and not hyperthermophile bacteria as in classic phylogeny. In the present paper I examine the robustness of their conclusions. (1) A site-by-site reading of the RNA alignments of Brochier and Philippe seems to suggest that the number of nucleotide positions used in their analysis is not sufficiently high and their phylogenetic analysis is consequently not robust. Furthermore, (2) a different method for selecting positions with no phylogenetic noise from the rRNA alignment relocates the Aquificales and the Thermotogales as the first lines of divergence in the Bacteria domain, and sets Planctomycetales as the third branch of divergence in the phylogenetic tree built from these selected positions. These findings consolidate the hypothesis that the ancestor of the Bacteria domain was a hyperthermophile and, more generally, that the last universal common ancestor might also have been one. (C) 2003 Elsevier Ltd. All rights reserved.