Dear Andrew,

I think you have to be very careful about defining what is meant by co-evolution, e.g. is it simply that substitution rates will concurrently increase or drop in corresponding lineages of two gene trees, or is it that some parts of the genes will evolve in concert (e.g. compensatory changes) - there are a multitude of possible definitions.

I concur with your concern about the confounding of founder effects and co-evolution when using methods devoid of phylogenetic correction. The distance which is based solely on substitution counts seems somewhat outdated as well.


Have you explored the relative ratio test to examine concurrent rate changes in the lineages of genes? We have done a similar analysis on genes involved in a tumor suppression pathway using relative ratio tests on dN and dS (Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login; Art Poon in our group here developed a number of methods (see datamonkey.org under the spidermonkey rubrik for details) that can map co-evolving sites in a gene - if your genes share the same phylogeny, you can conduct run the analysis on concatenated datasets.

Cheers,
Sergei