Reader Comments
Post a new comment on this article
Post Your Discussion Comment
Please follow our guidelines for comments and review our competing interests policy. Comments that do not conform to our guidelines will be promptly removed and the user account disabled. The following must be avoided:
- Remarks that could be interpreted as allegations of misconduct
- Unsupported assertions or statements
- Inflammatory or insulting language
Thank You!
Thank you for taking the time to flag this posting; we review flagged postings on a regular basis.
closeaamodel estimation for homeodomain
Posted by Ferdi on 06 Feb 2007 at 13:22 GMT
a mixed rate model (aamodelpr = mix)
http://plosone.org/article/info:doi/10.1371/journal.pone.0000153#article1.body1.sec4.sec1.p4
Shouldn't we think that a blank prior for aamodel requires much more sites in the dataset to be properly estimated than a very short homeodomain. This could likely explain the lack of resolution of bayesian tree !
RE: aamodel estimation for homeodomain
Joseph_Ryan replied to Ferdi on 07 Feb 2007 at 14:13 GMT
It is true that more sites in the dataset would lead to a better estimation of the prior. However, the alternative to choosing the mixed model would have been to pick one of the twelve fixed models available in Mr. Bayes: poisson, jones, dayhoff, mtrev, mtmam, wag, rtrev, cprev, vt, blosum, equalin, or gtr. We felt that choosing one of these models would have been a more arbitrary decision than specifying the mixed model. Our hope was that by allowing Mr. Bayes to estimate the appropriate fixed-rate model we would at the very least avoid using an extremely inappropriate prior. In addition, we believe that the large number of mcmc generations we ran, were sufficient to overcome any deficiencies in the estimation of the prior.