Hi everyone !

I am pretty new at Bayesian phylogeny and I just spent the last couple of days reading the manual of MrBayes software. I feel ready to use it but performing an analysis that make sense demands some understanding of the substitution models and other variable to be set up.

From my reading it seems that it exists nice alternatives to the model testing step by performing GTR cross-sampling or running a jumping model analysis. I understand that both of these procedures provide a result where several models have contributed. The jumping model procedure concerns only fixed rate models while GTR (General Time Reversible) is a variable rate model of substitution. However, in the 3.2 release paper of MrBayes, I read that "jumping model allows to sample across 203 possible time-reversible rate matrices".

I am not familiar with the terminology and the statistical background used in the manual and the paper I read. My question is:

Are the "Jumping model" and "GTR cross-sampling" two different procedures or not ?

The more I read, the less the difference are clear to me.

I would really appreciate any input on this naive question.

Thank you !

Anne-Lise D.