I am trying to calculate dn/ds of about 3000 genes amongst 14 mammalian species via codeml.
Shall i use already established phylogeny tree of these 14 mammalian species or shall i generate phylogenetic tree of each genes separately before codeml analysis
My gut feeling would be that you may want to make new trees for all the genes. As its mammalian you don't have to worry so much about recombination (I would definitely do this for a bacterium for example). But, different genes will be evolving at different rates, and with different pressures, which the overall phylogeny might not capture accurately.
Thanks. I will generate phylogenetic tree for each gene then :)