Analysis of redundant peaks in LC-MS/MS datasets

摘要

LC-MS/MS is an analytical technique used for protein identification and biomarker discovery. Typically, each peptide will be measured twice by the MS/MS, resulting in two or more mass-to-charge spectra for each peptide. The current study investigates various ways to combine these replicate measurements to improve the quality of the measured spectra, and in turn, the confidence and accuracy of the protein identification. Sample data was collected using a QSTAR XL hybrid quadrupole-time-of-flight mass spectrometer, given a input standard protein mixture of known composition. Spectrum alignment is used to identify replicate spectra. Various algorithms for combining these replicate measurements are investigated, and are compared to the current industry standard MASCOT algorithm. Results are judged based on protein identification rates following combination of replicate spectra. Algorithms which leverage the fact that mass-to-charge data collected above the parent ion mass are more informative for protein identification appear particularly effective, and warrants further investigation.