Protein inference

Percolator provides two protein inference methods: Picked protein FDR (-f) and Fido (-A). We recommend to use the picked protein FDR, as Fido can report inaccurate FDR estimates in certain cases and is prohibitively slow on large data sets.

Picked protein FDR

Example:

percolator pin.tab -f protein_db.fasta

To use the picked protein FDR functionality, a fasta file needs to be specified to assist the protein grouping. If no fasta file is available, or if protein grouping is not desired, you can simply provide an empty file (e.g. touch empty.fasta on unix).

The database searches should be carried out on a target and decoy protein database. For each target protein, a corresponding decoy protein with the reversed sequence should exist with the naming convention <decoy_prefix>proteinA, e.g. decoy_proteinA.

For advanced options, see command line protein inference options. As a guide, the most relevant options are listed below:

percolator pin.tab -f protein_db.fasta [-P decoy_] [-c] [-g] [-l target_protein_result.tsv] [-L decoy_protein_result.tsv]