Efficiencies of different genes and different tree-building methods in recovering a known vertebrate phylogeny

The relative efficiencies of different protein-coding genes of themitochondrial genome and different tree-building methods in recovering aknown vertebrate phylogeny (two whale species, cow, rat, mouse, opossum,chicken, frog, and three bony fish species) was evaluated. Thetree-building methods examined were the neighbor joining (NJ), minimumevolution (ME), maximum parsimony (MP), and maximum likelihood (ML), andboth nucleotide sequences and deduced amino acid sequences were analyzed.Generally speaking, amino acid sequences were better than nucleotidesequences in obtaining the true tree (topology) or trees close to the truetree. However, when only first and second codon positions data were used,nucleotide sequences produced reasonably good trees. Among the 13 genesexamined, Nd5 produced the true tree in all tree-building methods oralgorithms for both amino acid and nucleotide sequence data. Genes Cytb andNd4 also produced the correct tree in most tree-building algorithms whenamino acid sequence data were used. By contrast, Co2, Nd1, and Nd41 showeda poor performance. In general, large genes produced better results, andwhen the entire set of genes was used, all tree-building methods generatedthe true tree. In each tree-building method, several distance measures oralgorithms were used, but all these distance measures or algorithmsproduced essentially the same results. The ME method, in which manydifferent topologies are examined, was no better than the NJ method, whichgenerates a single final tree. Similarly, an ML method, in which manytopologies are examined, was no better than the ML star decompositionalgorithm that generates a single final tree. In ML the best substitutionmodel chosen by using the Akaike information criterion produced no betterresults than simpler substitution models. These results question theutility of the currently used optimization principles in phylogeneticconstruction. Relatively simple methods such as the NJ and ML stardecomposition algorithms seem to produce as good results as those obtainedby more sophisticated methods. The efficiencies of the NJ, ME, MP, and MLmethods in obtaining the correct tree were nearly the same when amino acidsequence data were used. The most important factor in constructing reliablephylogenetic trees seems to be the number of amino acids or nucleotidesused.