veg / hyphy

HyPhy: Hypothesis testing using Phylogenies
http://www.hyphy.org
Other
201 stars 68 forks source link

HyPhy Vision fails to open aBSREL output .json file #1582

Closed madeleineaaseremedios closed 1 year ago

madeleineaaseremedios commented 1 year ago

I ran aBSREL locally using hyphy for several gene families, and only one of them is having problems with HyPhy Vision "Something went wrong when rendering the results page. Please file a bug.". I know that it has tried to open the .json file because the input data line on the webpage updates to the name of my file. I also reran aBSREL on the dataset just in case the output file had been messed with, but the problem was the same.

I would run aBSREL on the Datamonkey website, but I would like to specify the phylogeny and that doesn't seem to be possible on the web version.

I've attached the codon alignment and phylogeny I used to run aBSREL, as well as the .json output (as a txt file so it could be uploaded).

Any help would be appreciated! (edit: I have removed my files for privacy, now that I fixed the problem)

madeleineaaseremedios commented 1 year ago

I would like to add that I already checked for the known 'inf' error and that's not the problem.

madeleineaaseremedios commented 1 year ago

Maybe it is because in the 'Testing selected branches for selection' section of the screen log output, there is a branch that says "Not selected" and "for testing" under the columns for 'Test LRT' and 'Uncorrected p-value'?

I am not sure how this would mess up the .json file though, or what I could change to fix it.

madeleineaaseremedios commented 1 year ago

I think I fixed the issue - it was a typo... There was a problem with the input files where in one sequence name in the alignment, a hyphen should have been an underscore, so even though hyphy renames sequences to replace characters it doesn't like with underscores, it then got confused between the original name and the new edited name. Weirdly this didn't make an error about different sequence names between the files though... It did fix the "Not selected" and "for testing" issue, as that branch now exists (because the sequence label now exists) and everything works fine.