I attempted to upload an alignment of a few hundred sequences, each about 7,000 bases long.  I got a "data file too large error" but it did not tell me what the maximum file size is.

Dear Brian,

There are two limitations on file size in datamonkey:

1). Raw file size (4MB) - basically to prevent hackers from trying to crash the server - these files never get processed.
2). For valid alignments, current size limitations are 150 sequences and/or 10000 codons. This is mostly to keep a single job from running over 3-4 minutes. You can always run a larger analysis locally in HyPhy (it is am option in Standard Analyses); data sets larger than 1000 sequences can cause underflow problems (we are working on fixing that, but it is rather involved).

Internal stop codons are indeed disallowed. The analyses are mostly targeted towards a single-gene setting, in which there should be no premature stop codons.

HyPhy has a tool ('Standard Analyses: Data file tools: CleanStopCodons.bf) which reads in an alignment and replaces all stop codons with gaps. This might be a reasonable way to preprocess alignments with premature stop codons (due to sequencing errors, e.g).

If there is enough interest, I could implement a user-selectable option for the datamonkey upload module to

(a). Replace all detected stop codons with gaps
(b). Replace stop codons with site consensus (good for sequencing error correction).

HTH,
Sergei