HyPhy message board - Print Page

HyPhy message board
http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl HYPHY Package >> HyPhy feedback >> Rooted Vs Unrooted Trees for Branch Site Analysis http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl?num=1186438083 Message started by Andrew_Roth on Aug 6^th, 2007 at 3:08pm

Title: Rooted Vs Unrooted Trees for Branch Site Analysis
Post by Andrew_Roth on Aug 6^th, 2007 at 3:08pm

Hi Sergei,
I am currently working with a data set containing four distinct clades corresponding to gene orthologues. I would like to conduct a branch site analysis and test each of the branches leading into the clades.

I am wondering what effect rooting the tree with one of the clades as an outgroup would have versus using an unrooted tree to conduct the analysis.

Cheers,
Andrew

Title: Re: Rooted Vs Unrooted Trees for Branch Site Analy
Post by Sergei on Aug 8^th, 2007 at 6:28am

Dear Andrew,

The model used by the branch-site test is time reversible - rooting should have no effect (i.e. the unrooted tree is used in any case). If you see a difference for two rootings of the same tree, this could be an indication of convergence issues.

Cheers,
Sergei

Title: Re: Rooted Vs Unrooted Trees for Branch Site Analy
Post by Andrew_Roth on Aug 8^th, 2007 at 10:54am

Thank you for the help Sergei.