Enzymes are listed along the hydrogenotrophic pathway. Enzymes supporting the monophyly of methanogens are in italics (BV >75%). Enzymes supporting the monophyly of both Class I and Class II are in bold. Enzymes present in methanogens only are underlined. Genes present in Class I only are indicated by an asterisk. Phylogeny supporting the monophyly of Class II only are indicated by a +. Genes without lateral gene transfers between Class I/ Class II are in quotes. Genes with duplication in methanogenes, leading to incongruent position of a species in the tree are denoted by a superscript 2. Sixteen genes out of 20 coding for the seven core enzymes of hydrogenotrophic methanogenesis (cofD, fwdA, hmdI, mtrA, mtrB, mtrC, mtrD, mtrE, mtrF, mtrG, mtrH, mcrA, mcrB, mcrC, mcrD and mcrG) are consistent with the monophyly of methanogens. In other words, proteins of methanogenesis themselves seem in contradiction with the reference tree based on ribosomal proteins. In fact, 13 of the 15 proteins consistent with the monophyly of methanogens are exclusively present in methanogens (an additional protein is exclusively found in Class I, see Table 1). Such a restricted taxonomical sampling obviously increases the number of markers with monophyletic methanogens, but does not really help to test this monophyly. A crude interpretation could be that methanogenesis is a plesiomorphic trait, ancestral to most of the euryarchaea. It would have been lost several times independently in the Thermoplasmatales, Archaeoglobales and Halobacteriales lineages.