I previously had issues on differences in positively selected sites between PAML and Datamonkey.  The conclusion was that the differences were attributed to PAML not allowing for synonymous rate variability while Datamonkey does.  It is unclear to me however, how, in PAML, a single dS rate is used.  Where does this single dS rate come from and how is it used in the calculations when estimating the omega category for individual sites?  Then again, I may be misinterpreting this concept of synonymous rate invariability in PAML.  Sorry if this is all is unclear but I am rather confused and not quite sure how to start asking questions on the subject.  Any bit of help/input/advice would be great.

Thanks!
- Denise

Dear Denise,


wrote on Jul 17^th, 2006 at 10:20am:

It is unclear to me however, how, in PAML, a single dS rate is used.  Where does this single dS rate come from and how is it used in the calculations when estimating the omega category for individual sites? Denise

In a codon model (like MG94 or GY94) one can generally estimate two rates: alpha (~dS) and beta (~dN). In PAML, the synonymous rate is effectively fixed at 1 (for technical reasons you can't estimate dS and tree length separately, only their product, so dS = 1 is a convenient relative scale). In datamonkey (and HyPhy), the synonymous rate can be set to one on average, but allowed to vary from site to site. This is the main difference. I strongly encourage you to read Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login where all these issues are touched upon.

The paper may be a bit technical, but it thorough. I'll be happy to assist you with further questions, if something in the manuscript is confusing.

Cheers,
Sergei