search

Rappsilber-Laboratory / build-xiview

GNU General Public License v3.0
2 stars 0 forks source link

Unable to read mgf files #108

Closed bsakc closed 2 months ago

bsakc commented 2 months ago

log id:mmacgkfv5e00qtjsp7o96rt4pq_1719841584.log error type:Error message:duplicate column(s): matchscore id: undefined

I have created the mgf files on both proteome discoverer and MSconvert however, i always get this error message and cannot see the spectra matching to the crosslinks. the csv files were produced using xisearch. Please let me know if there is an issue with my workflow

colin-combe commented 2 months ago

Hi, the error message suggests there are two columns named "matchscore" in the CSV file you uploaded. It is case insensitive with regards column headers.

Please check the CSV files to see if it has two columns named 'matchscore'. If so, try deleting one of them and uploading again. Please let me know if this works or if there are still problems.

Best wishes, Colin

bsakc commented 2 months ago

Hi Colin, thank you for your quick response. I checked the csv file and there is only one column for matchscore. Is there anything else it could be?

Thanks Adam

colin-combe commented 2 months ago

Hmmm. Could you share the CSV file with me, please? If you don't want to post your data in public here, you can email it to me at colin.combe@ed.ac.uk.

colin-combe commented 2 months ago

it could be taking spaces out of column names also, so "matchscore" and "Match Score" would count as duplicate column names

bsakc commented 2 months ago

Hi Colin

i tried to send you an email yesterday with the csv file but the email could not send for some reason. i have attached it here. bsa dss_Xi1.7.6.7.csv

bsakc commented 2 months ago

i have seen your email about the file containing two different matchscores. which one would be best to delete?

colin-combe commented 2 months ago

yes... it has "MatchScore" and "match score" and the xiview parser thinks these are the same, which it maybe shouldn't do. But the easiest thing in short term is just to delete one. @lutzfischer might have some guidance on which is better to delete? Or you could make an arbitrary choice just to see if it works.

bsakc commented 2 months ago

thank you so much for the help I'll try those now

lutzfischer commented 2 months ago

The right one to delete would be MatchScore.

grandrea commented 2 months ago

I think those columns are part of the xiSEARCH csv output, not of the xiFDR CSM file that is the one that should be uploaded to xiView. Indeed the file you have uploaded here is not the CSM file, but it's the unfiltered csv file coming out of xiSEARCH as far as I can tell. See reply https://github.com/Rappsilber-Laboratory/XiSearch/issues/107

colin-combe commented 2 months ago

i think @bsakc has things working now.

You should upload the mzIdentML file produced by xiFDR.

https://github.com/Rappsilber-Laboratory/XiSearch/issues/107#issuecomment-2213176230