Parameters for rescoring peptidomics data

[Dmorgen]

Hi,

Where can I change the enzyme parameters in the GUI application for peptidomics data? I've tried to input in the Mokapot FASTA parameters the Enzyme as '.' (as written in the Mokapot documentation) but I got an error that the Enzyme parameters are invalid. Is there anything I should do differently?

Thanks!
David.

[MatteoLacki]

Same question. Experimented with:

            "mokapot": {
                "write_flashlfq": true,
                "write_txt": true,
                "write_weights": true,
                "enzyme":"[KR](?!P)"
            }

and

"rescoring_engine": {
            "mokapot": {
                "write_weights": true,
                "write_txt": true,
                "write_flashlfq": false,
                "protein_kwargs": {
                    "enzyme":"[KR](?!P)"
                }
            }
        },

and it all fails. so it's back to the source and that's not simple to parse.

[MatteoLacki]

I manage to get it work: one needs to provide both fasta and protein_kwargs:

"rescoring_engine": {
            "mokapot": {
                "write_flashlfq": true,
                "write_txt": true,
                "write_weights": true,
                "fasta_file": "my_favorit_fasta.fasta",
                "protein_kwargs": {
                    "enzyme":"[KR](?!P)"
                }
            }
        },

[MatteoLacki]

Looking retrospectively: this is likely maybe not a bug, but an inconsistency: when I was running the command line tool, I did pass in the -f command. It was not passed on further to mokapot. I will make it into an enhancement report.

[RalfG]

Hi @Dmorgen,

Thanks for your interest in MS²Rescore!

Like Matteo mentioned, to enable protein inference and protein-level confidence estimation in Mokapot through MS²Rescore, you must provide both fasta_file and protein_kwargs options. These are passed to the mokapot.read_fasta function.

For a non-specific digest, such as in immunopeptidomics, the settings might look like this:

"rescoring_engine": {
    "mokapot": {
        "fasta": "uniprot-proteome-human-contaminants.fasta",
        "protein_kwargs": {
            "enzyme": ".",
            "missed_cleavages": 16,
            "max_length": 16,
            "min_length": 7,
            "semi": false
        }
    }
},

Note that in this case, the missed cleavages should be set to the same number as max_length, although 0 might also work (could not find that in the Mokapot docs).

One caveat here is that depending on the size of the search space (amount and length of proteins), Mokapot might take a very long time to calculate the full digest of proteins into peptides, and protein inference on immunopeptidomics experiments is perhaps also not what you are looking for.

Depending on the MS²Rescore PSM input file format (coming from the search engine), the mapping from peptides to proteins will already be there, and will be propagated to the PSMs TSV MS²Rescore output file.

Best,
Ralf

[Dmorgen]

Thank you Ralf and Mateo!