I'm trying to run a GA locally. My input is:

5 branch categories
Universal code
Use codon file in nxs format (saved as complete data+partition+tree in HyPhy; running with ~26 sequences)
No site-to-site rate variation
012340 substitution model
Use tree included in codon file (this tree has branch lengths, which were read fine by HyPhy)

BranchSelector.bf was unchanged. The program took 7 CPU hours (28 processors). I ran ResultProcessor.bf and saw that all the branches had 0% above them and one apparent rate class. In the plain output file (not .cAIC or .samples) the branch assignments are zero also. I can't figure out what about my input would cause the problem. Could it be an error in the tree? I checked its structure in HyPhy; it looked fine. The postscript generated has an odd (vertical) branch, or maybe it's an artifact.

Also, in the messages.log file for the first (0th) processor, it says "Using model MG94_Model for definition of [node]." Where can I specify something else?

Any troubleshooting tips would be much appreciated.

Sarah

I think I'm under the cAIC threshold--there are 328 codons.

MG94 is great. I was just slightly worried when the program 'defaulted' to it with no mention of 012340.

Thanks for the help,
Sarah