Ancestor reconstructions

[bredelings]

I'm very impressed at what historian can do! It seems like it can/should replace PRANK, at the very least. I will make sure to include it in any future benchmarks.

For the ancestral sequence reconstructions, two thoughts:

• maybe you could use N/X instead of * for an unknown letter? I think that * can mean "stop codon" in some cases, and I don't think many people in bioinformatics use *
• how hard would it be to sample actually letters instead of N/X? In mcmc mode where you write a number of different alignments, that could be useful.

[ihh]

Thanks Ben, sorry for the slow reply.

You're probably right that "N/X" would be better than "*", I've added this as a separate issue (#7).

It should sample ancestral sequences already if you specify the "-ancseq" option, was that not working for you or was it just unclear from the usage text? (The help message is a bit cryptic on this point, it's kind of a buried feature.)

I really need to update it to use differential automata, but this is an unfunded project for me at this point...

[ihh]

I've now created a separate test for the -ancseq option (#11) so I am closing this issue.

[ihh]

@bredelings FYI, the use of N/X for ancestral wildcard characters is now allowed as of 3a292a7. Historian still uses asterisks for wildcards internally, which means that you still can't use "*" to denote a stop codon (changing that would have required a bit more refactoring than I had time for), but at least it issues a warning if you try to use "*" as an alphabet character. I think that modeling stop codons as allowable substitution states is a sufficiently niche application that it's probably fine to ask those users to represent their stop codons with something other than an asterisk.