What is the best codeml model to determine if a gene is undergoing positive, negative or neutral selection? At this point I am not interested in site specific omega values, but rather want to estimate a value for the whole gene. ete3 toolkit offers models from M0 all the way to M13. For the most part they seem to follow the same trend. I was wondering if M0 is okay to use for getting an average omega for all sites in a gene.

More info on purpose: I have some virus-host protein-protein interaction factors and want to determine from a longer list the pairs of virus-host genes that have higher omega values indicating rapid evolution.