How do you test for significant dN/dS rates between paralogues?

[DNAngel]

If I have paralogous genes and I want to test if one gene is subject to more or less positive selection, do I have to run PAML on them individually various site models? How do I then compare the omega values obtained in the positive selection class are significant between the genes as I don't think it makes sense to concatenate the genes into one alignment file and then run PAML.

Your help is much appreciated! Thank you in advance.

[Ziheng]

not sure what you are trying to test, but there must be many papers which used codeml to analyze paralogous genes.  one possible analysis is use a big tree for all the orthologues and paralogues, and then identify branches right after gene duplication and test whether functional divergence accelerated amino acid changes after gene duplication.

perhaps you can glance over this old paper and see whether there is anything relevant:

Bielawski JP, Yang Z. 2003. Maximum likelihood methods for detecting adaptive evolution after gene duplication. J Struct Funct Genomics 3:201-212.

ziheng

[DNAngel]

Yes you are right, I created a tree with both paralogs and then simply set one as the foreground. I did have to manually copy the tree because I wanted to use the species tree, not a gene tree. I did try doubling my sequence alignment file to create a gene tree just to see hwo it looked and it did group all the paralogs together except one was embedded in between the second. I