search

CCMS-UCSD / GNPS_Workflows

Public Workflows at GNPS
https://gnps.ucsd.edu/
Other
52 stars 43 forks source link

molecular-library-search-v2 bug at the get_dbannotations step #800

Closed adelabriere closed 2 years ago

adelabriere commented 2 years ago

I ran into a bug running the MOLECULAR-LIBRARYSEARCH-V2 workflow, I have an exception, possibly originating from an empty file somewhere during the processing. My input MGF is valid and can be opened programmatically using R MSnbase. The job is tagged as FAILED.

Link to the job here

The graph and exceptions are the following The graph The exception

Do you have any idea from where the problem can originates from and how to solve it ?

mwang87 commented 2 years ago

This means there are new hits, I need to fix this at some point so its more user friendly!

adelabriere commented 2 years ago

Well at the moment it is just crashing the job. Is there a solution so that I can get the list of matches for this set of MS2 spectra ?

mwang87 commented 2 years ago

Well there are no matches for this search so there is nothing to get.

adelabriere commented 2 years ago

Haaa thanks, since you said there "new" hits I was confused. Thanks it is clear now.

mwang87 commented 2 years ago

Oh man just looked at the actual data a bit, negative polarity is rough on public libraries. Just in case it's relevant, how did you format the mgf file?

adelabriere commented 2 years ago

It was a SLAW output, the spectra are consensus MS-MS spectra. It worked fine on the same dataset in positive mode with exactly the same formatting, so I don't think the format is a problem. The data seems really emptier in negative mode so it does not surprise me that there is just 0 matches.

mwang87 commented 2 years ago

Ok great. Yeah I'm always nervous when it comes to mgf since it's such a poorly defined format. One thing you could try in case you're interested is turning on analog search. We also have a few other libraries that are propogated that aren't selected by default.