Table 1 Performance of SpecGlobX measured on the simulated dataset Dsim. At first, we provided the full list of ‘correct PSMs’ as input to SpecGlobX: each simulated spectrum is associated with the peptide sequence it derives from (row 1). Then, we executed SpecGlobX on the PSMs returned by three different OMS methods (rows 2 to 4). Several percentages of correct identifications are summarized with and without (W/o) SpecGlobX. As a first criterion (column 2), a PSM is counted as correct if the spectrum is associated with the peptide sequence it derives from; the second criterion measures the percentage of PSMs that exhibits the neutral loss among the 50,000 PSMs (columns 3 and 4); the third criterion counts the PSMs in which all detected modifications are correctly identified and placed related to all PSMs (columns 5 et 6); the last criterion refers to the percentage of modifications that are correctly identified relative to the number of modifications (about 103,000) incorporated in the simulated spectra (columns 7 and 8)

From: Fast alignment of mass spectra in large proteomics datasets, capturing dissimilarities arising from multiple complex modifications of peptides