I am undergrad student, and I am studying the evolution of one gene in a subgroup of Drosophila species. For that, I am comparing the sequence of this gene of 12 Drosophila species, and I am looking for signs of selection. I have populational data from D. melanogaster and D. simulans to run HKA test, and I am also running Tajima relative rate tests in some points of the phylogeny.
I expect that this gene is under purifying selection (I'll compare the dn/ds with the average dn/ds of the whole genome)
What I did before was run dn/ds calculations on SNAP, and I got the dn/ds between Dmel and each of the other 11 species, but that didn't make sense because it didn't take phylogeny in consideration.
So what I need to do now is to find the dn/ds for this gene, and for that I am using CODEML, but I am confused about the model I should run. I have read the PAML manual, but I am still uncertain about what parameters I should use.
Should I run different models and then compare them?