Detecting the convergent adaptative evolution by using branch-site model.

[金森駿介]

Dear, All

I want to detect convergent adaptive evolution by using branch-site model in the phylogeny attached.

Red branches indicate the branches on which I want to detect the adaptive evolution.

Now, I'm going to three separate runs of PAML, in which I specified the each red branches as foreground.

In each runs, Do I have to remove the other red branches?

Please make some suggestions to me.

Kanamori

[Ziheng]

I think it is better to keep them in the tree and treat them as background branches. Having positive selection on some of the background branches does not have much impact on your detection of positive selection on the foreground branch.
We looked at this issue in some simulation papers, but I am not sure whether it is
Fletcher & yang 2010
or an earlier paper. You can take a look.
Ziheng

At the end you can add another analysis in which you label all three red branches as the foreground branch.
Ziheng

[金森駿介]

Thank you for quick reply.

I've done the analysis which keep them as background branches.

Now I have another question about multiple testing.

In this case, should I do multiple testing?

If so, which is the reason why I sould do it, the number of genes or the number of foreground branches?

Please advise me.

[Ziheng]

If you suspect convergent evolution occurring along the three red branches, perhaps you should treat all three branches as foreground branches.
Ziheng

[Yangkang Chen]

Hi Kanamori, I'm facing the same problem. How did you finally solved that?