ORC Priors

[Dylan O'Hearn]

Hello,

I'm performing a concatenated analysis of mtDNA and nuclear data and I'm partitioning the data with a relaxed clock model for each data type- one for mtDNA and one for nuclear.  Each data type is also further partitioned, with each codon position for mtDNA and each of three genes for the nuclear DNA having its own site model.  I have a couple questions.

1) By default when setting it up this way, one clock model has its rate un-estimated and the other is estimated relative to that first one.  In my case, it seems sensible to estimate the mtDNA rate, which I expect to be about ~25x the nuclear rate overall.  Which priors do I change to do that?  I'm a little confused about whether ORCRates, ORCucldMean, or both, need to be changed to be informative.

2) I expect rate variation between the nuclear genes and the codon positions, so I expect I need to check "estimate" for all the mutation rates in the site model tab, and let the mean be fixed to one so they're identifiable.  However, I don't know if the fact that two separate clock models are involved makes that more complicated.  Is estimating them all and fixing the mean at one still valid, or do I need to have one un-estimated for each clock model and estimate the others relative to that, or something?  A little lost in the options here.

Thanks!

Dylan