Codon based alignment for detecting positive selection

Hi everyone,

I’m trying to apply some site models to detect positive selection in proteins using ete3 evol. After a lot of troubleshooting, I realized it requires a ‘codon based alignment’. I currently just have a list of protein sequences in one file and their alignment in another file, how can I go back to the nucleotide level to generate this ‘codon based alignment’. Has anyone had this issue before?

Thanks.