Philosopher Protein ID and Gene in combined_proteins.tsv shows the full protein header

[marlenejarap]

- Upload your log file

log_2025-07-29_16-32-25.txt

- Describe the issue or question:
Hello, first thank you for developing Fragpipe and maintaining it.

I have an issue when trying to run a metaproteomics custom DB. The idea of this DB is to combine proteomes from Uniprot with the proteomes derived from a metagenomic assabled genomes (MAG DB). When running each of them separately there is not major issues (no issue in the case of the Uniprot DB metaproteome) and we get all the results that includes the file with the combined proteins. In this case there is only a minor issue which is that in the case of the MAG DB the Protein ID and Entry Name consist basically of the whole header and this is not a big issue as we can correct this in the ouput. However whant combining Uniprot DB and the MAG DB then we get an error that comes from Philosopher. As the custome DB does not have the header as Uniprot an attempt of mimicking the Uniprot header before joining the both DBs was not successful. Example is below:

MAG DB header example before modifying the headers

ds|KKJIJG_00005|donor1_SB_149 hypothetical protein OS=Bilophila wadsworthia

MAG DB header example after modifying the headers

ds|KKJIJG_00005|donor1_SB_149 hypothetical protein OS=Bilophila wadsworthia OX=unknown GN=unknown PE=unknown SV=unknown

I am not sure why this does not work. Maybe the tokens with unknown should be empty instead or filled with an space?

The error message is:

Process 'PhilosopherDbAnnotate' finished, exit code: 2
Process returned non-zero exit code, stopping

Cancelling 48 remaining tasks



Maybe I should also mention the the Decoys were added with Fragpipe tool. 

Thanks again

[AimeeD90]

@marlenejarap Thanks for reporting this, but Philosopher should work fine with both original and modified headers. Could you please share the combined modified fasta file so I can help resolve this issue?

Yamei

[marlenejarap]

Hi Yamei. Thank you very much for the quick response. After doing some sanity checks to the DB became apparent that this issue was caused by the names of some proteins, like in the case of this header ds|PPAJLN_06010|donor2_SB_2 3'->5' ssDNA/RNA exonuclease TatD ...... by the symbol ->. After renaming the proteins with this pattern philosopher completed the run.

The small issue of names allocations in the final combined proteins tsv still remains. Attached a sample of the headers. If tested then it should be done with fecal sample that will trigger the annotation with non unitprot proteins.

Custom_with_uniprot_test_sample.txt