Evaluation of Data Analysis Strategies for Improved Mass Spectrometry-Based Phosphoproteomics

摘要

Here we describe a set of enhanced data processing andnfiltering methods to improve significance and coverage ofnphosphopeptide identifications by mass spectrometry. Wendemonstrate that for samples of limited complexity,nspectra-based estimation of false discovery rates will leadnto overprediction of confidently identified phosphorylatednpeptides due to a bias caused by multiple fragmentationnof highly abundant peptide species. We further providenevidence that fragmentation of abundant peptides at thentails of their chromatographic peaks is a major source fornfalse positive peptide matches and that overall confidencenin phosphopeptide identifications can be improved by anchromatographic peak-based aggregation scheme, intensitynrank-based neutral loss and optimized mass errornfilters. When replicate runs of a standard sample werenperformed using different fragmentation techniques on annOrbitrap mass spectrometer we observed improvementsnof 7-31% in phosphopeptide coverage depending on thenfragmentation method and the desired false discoverynrate.