Good evening,

In phylogenetic analyses, when we work with DNA we have to find the best DNA susbstitution model for our data, right? But in many cases our data is a mixture of different genes with different evolutionary histories; different mutation ratios, selection pressure, or even heritability (mitochondrial versus nuclear). This is also true even in a single gene, since the genetic code establish that a mutation in the third codon position is more probably to be synonymous compared with the first and second codon positions, and therefore more likely to be fixed in the genome.

Thus, a single model could not fit this variation. The way to solve this problem is to partitioning your data in "similar-evolutionary" units and look for the best model for each one of them.

how we check this? I suggest you to use PartitionFinder https://www.robertlanfear.com/partitionfinder/

This open source software will search for the best partitioning scheme for your data, and the best model for each of the different partitions.

Good luck!

David Barros-Garcia Thank you very much for your explanation. Can we consider the same situation for partitioning amino acid sequences?

You are welcome,

Well, I have never made phylogenetic analyses with aa, probably you should look into literature and see how others deal with that kind of data!