I would like to use MrBayes to test topological branch hypotheses. Specifically I want to create one reference tree with ingroups and outgroups and then test whether various sequences fall into the ingroup or the outgroup, using constraints (following instructions from section 4.4 in the MrBayes manual).
Since I plan on running this analysis many times with different candidate genes but the same reference tree, I was wondering if I could make my analysis more computationally efficient by using the topology of the reference tree and/or prior parameters for the reference tree as inputs for the Bayes factor comparison?
For the tree topology: Is it possible to use a user-defined tree that doesn't have all the species? (the alignment would have the reference genes + a single candidate gene)
For the prior parameters: Since the manual says that using more informative priors can reduce over-parameterization, does it make sense to use the prior parameter estimates from the reference tree as priors for the new reference+candidate tree?
Any insight or advice is greatly appreciated.