Comment on "Hexapod Origins: Monophyletic or Paraphyletic?"

Publisher Website
Google Scholar
Abstract
Nardi et al. drew their conclusions from maximum likelihood and Bayesian analyses at the amino acid level of four of the 13 mitochondrial proteins for both the original 35-taxon data set and a 15-taxon subset. However, phylogenetic analyses of amino acids carry several potential caveats. First, the currently available models of mitochondrial amino acid substitution are based on empirically deduced matrices from mammalian-dominated sequence databases. Second, the maximum likelihood analysis used in (1) does not model the variation of rate across sites, which is known to be one of the most important parameters of the likelihood model (3). Third, bias in nucleotide composition also affects the amino acid composition of the gene product, thereby causing potential problems for phylogenetic reconstruction (4).