DNA Hybridization as a Guide to Phylogenies: a Critical Analysis

Abstract— This article evaluates the use of DNA hybridization for estimating the extent of divergence among the single-copy fractions of vertebrate genomes. It focuses, in particular, on the nature and informational content of the melting profiles as a guide to phylogenetic relationships. While concluding that the DNA hybridization approach remains the best and most cost-effective guide to such relationships over its useful range, it demonstrates serious flaws in certain recent attempts to apply the method to specific cases among primates and birds. The major points are:

• 1 The T₅₀H statistic is flawed as a measure of mean sequence divergence, and also, therefore, as a measure of phylogenetic distance.
• 2 The T_mode statistic overcomes many of the problems inherent in interpreting thermal stabilities of DNA heteroduplexes for phylogenetic purposes.
• 3 The phylogenetic significance of ΔT_modes of > 15d? or so cannot be accurately assessed.
• 4 The putative slowdown in the rate of nuclear DNA sequence change among the lemurs is not justified by the data.
• 5 The claims of Sibley and Ahlquist to have resolved the human/chimpanzee/gorilla trichotomy are not supported by their data.
• 6 There are major problems in the published Sibley and Ahlquist avian phylogenies; in particular, with those containing evolutionary “staircases” of nodes separated by less than 1d? from one another.
• 7 There would appear to be a lineage misplacement involving a ΔT of at least 4d? in a recent publication on avian phylogeny.
• 8 Certain of the published ΔT₅₀ H values seem not to be representative of the actual data on which they are based.
• 9 Most important, it is recommended that no phytogenies based on DNA hybridization comparisons should be presented without being accompanied by the data relevant to each claim of a resolved lineage.