Trinity: Bug in output collection when using genome-guided mode

[natefoo]

From Gitter:

@natefoo

It looks like Trinity with the genome-guided option selected produces outputs with -GG in the filename but the wrapper doesn't account for this? From the Trinity log:

Thursday, June 24, 2021: 06:41:20       CMD: find Dir_*  -name '*inity.fasta'  | /conda/envs/[email protected]/opt/trinity-2.9.1/util/support_scripts/GG_partitioned_trinity_aggregator.pl TRINITY_GG > /working/trinity_out_dir/Trinity-GG.fasta.tmp
Thursday, June 24, 2021: 06:41:24       CMD: /conda/envs/[email protected]/opt/trinity-2.9.1/util/support_scripts/get_Trinity_gene_to_trans_map.pl /working/trinity_out_dir/Trinity-GG.fasta > /working/trinity_out_dir/Trinity-GG.fasta.gene_trans_map


Finished. See /working/trinity_out_dir/Trinity-GG.fasta for reconstructed transcripts

And from Pulsar:

2021-06-24 06:46:43,659 WARNI [pulsar.managers.staging.post][[manager=bridges]-[action=postprocess]-[job=1234]] Failures collecting results [Exception('Attempt to post file /working/trinity_out_dir/Trinity.fasta to URL https://galaxy-web-03.galaxyproject.org/_job_files?job_id=deadbeef&job_key=deadbeef&path=/data/dataset_42.dat&file_type=output_workdir, but file does not exist.'), Exception('Attempt to post file /working/trinity_out_dir/Trinity.fasta.gene_trans_map to URL https://galaxy-web-03.galaxyproject.org/_job_files?job_id=deadbeef&job_key=deadbeef&path=/data/dataset_42.dat&file_type=output_workdir, but file does not exist.')]

@nsoranzo

We probably need to add a test for this to the wrapper.

[jennaj]

Hi -- we have a question about this functionality at Ghelp. I had forgotten it was an an open issue! I'm going to cross post to the tools working group here. Maybe someone can help. Or, we don't fix and state that clearer(remove functionality from tool form?).

Discussion

• https://help.galaxyproject.org/t/t8646/7
• note that the specific version of Trinity that had a native problem (same function) is not the version we have wrapped
• if I'm wrong about that, please correct me

Tests

• Imported the Ghelp test history (uses public test data) and tried doing some data cleanup to see if that was the issue. No, the wrapper is the problem from the tests that have completed so far. We are just not capturing that output. Tool doesn't fail -- green empty results.
• https://usegalaxy.eu/u/jenj/h/imported-trinity-genome-guided-test-httpshelpgalaxyprojectorgtt86467-9-27