What kind of analysis are you trying to perform with the MrBayes package??

Hello Ijad, i am allready working with MrBayes and it works quit good.

Iam doing phylogenetic tree estimations from bacteria and archea 16S rDNA sequences.

By now I use the GTR model (set nst=6 rates=invgamma) and I am not sure if there is any better model.

Do you have experience with MrBayes and what multiple sequence alignments algorithm are you using.

reg Frank

Frank, it is really depending on how similar the sequences are which you are trying to visualize in a phylogenetic tree. Very similar (around 80%) UPGMA, between 50-80% NJ and lower than that Maximum Likelihood. Have you tried to use MEGA Software package?

Hello Ijad,

how do you know about these values for the orientation? Do you have a artikel about this.?

Reg Frank

Yes, I have this from a famous bioinformatics book. Problem: Have to find this book. :). I think I cited it in the penton protein publication. Give you heads up, when I found it!

I must say that I have to disagree with Ijad - UPGMA is an unrealistically simplistic approach and is unlikely to make it past peer-review in any respectable journal, these days. The problem is that if there is variation in evolutionary rate among different lineages, then UPGMA is known to be problematic - pushing rapidly-evolving sequences towards the root. I'm also not a fan of NJ, because it doesn't explore treespace, doesn't permit the substitution model to change during the analysis and is not built on the robust foundations of maximum likelihood and bayesian inference, so testing alternative topologies or models is not possible. Model selection is possible using MrBayes and I suggest reading the paper by Ronquist in Systematic Biology in February for some more info.