Score Filters

Score filters are some of the most important filters in PIA, for example a protein inference usually should only use PSM(-set)s with a (Combined) FDR score below a certain threshold.

PSMs

On the PSM level, the score filters can be used to filter for all imported scores, probabilities and e-values in each input file (fileID > 0), but also for (Combined) FDR Scores on the overview (fileID = 0).

To apply a score filter on PSM level (also for protein inference), the filter's short name has to start with the prefix psm_score_filter_ followed by the scores internal short name. See the score names for a list of the mostly used score names. To filter e.g. for an X!Tandem expect value, the filter short name must be psm_score_filter_xtandem_expect.

Additionally the PSMs of each input file can be filtered for the "Top Identifications", i.e. the best identifications if a search engine applies more than one PSM to a spectrum. For these "Top Identifications" use the prefix psm_top_identification_filter_ and use the same rules as given for the normal scores.

Peptides

For score filtering on peptide level, the same rules as for PSMs apply, except the prefix is peptide_score_filter_. Usually this filter is not applied for protein inference.

Proteins

On the protein level, there is only the ProteinScore, which has the short name protein_score_filter