Protein Ids missing for some identifications about percolator

[sdgroeve]

Hello,

I have the same problem. I'm using 3.6 as:
percolator --subset-max-train 5000000 --maxiter 5 --verbose 2 --results-psms result.out --decoy-results-psms result.decoy --trainFDR 0.01 --protein-decoy-pattern decoy_ --picked-protein concat.fasta --results-proteins result.proteins.tsv --weights result.weights --protein-enzyme trypsin --spectral-counting-fdr 0.01 --unitnorm --only-psms --post-processing-tdc --protein-report-duplicates --protein-report-fragments --decoy-results-proteins proteins_decoy.tsv --protein-report-duplicates --protein-report-fragments input.pin

There is also an issue with the protein result file. In the pin I have proteins matched with many peptides that are not in the protein result file. E.g. UBE2C. This protein is present twice (with one of them a subset of the other), so I can see this is a special case, but at least one of them should be reported no?

The pin file and result files are attached.
files.zip

Thank you for your reply!

[bmx8177]

I have also come across this issue on the most recent version of Percolator. To me it appears that the column does not have a value when the peptide is only found in a single protein.

[wsnoble]

@bmx8177 can you see if the problem is still there when you get rid of all the extraneous command line options? Can you check to see if it's still there if you remove the "filename" column from the input file?

[bmx8177]

@wsnoble Yes, the problem is there under both conditions you asked for.

/home/lina627/bin/percolator/Release/src/percolator --results-psms percolator.target.psms.txt make-pin.pin

/home/lina627/bin/percolator/Release/src/percolator --results-psms percolator.target.psms.remove-filename.txt make-pin-remove-filename.pin

percolator.target.psms.remove-filename.txt
percolator.target.psms.txt
make-pin-remove-filename.pin.txt
make-pin.pin.txt