GroupFilter: a software tool for efficient filtering of Morpheus search engine results
T12N4
M.V. Ivanov, L.I. Levitsky, A.A. Lobas, M.V. Gorshkov
Morpheus is a search algorithm developed recently for high-resolution tandem mass spectra. According to the developers, its intrinsic property is discriminating short sequence length peptides. Therefore, elimination of direct comparisons between peptide spectrum matches (PSMs) for short and long peptides may potentially increase the search sensitivity for a given FDR level. In the proposed approach, all PSMs are grouped according to the number of matched fragment ions, followed by separate filtering of identifications in each group using target-decoy approach. The approach is applied to Morpheus output results and does not cause a significant increase in the overall data analysis time. The proposed approach was implemented as a Python command-line tool, called groupfilter. Several data sets from different types of mass spectrometers were used for testing of the software, including the data from the original Morpheus search engine paper. Separate FDR filtering for grouped identifications resulted in up to 18 % increase in the number of identified peptides filtered by GroupFilter compared with the default Morpheus’s post-processing procedure. The proposed approach can be considered as an addition to the Morpheus search engine.
