A Machine-Learning-Based Algorithm for Peptide Feature Detection from Protein Mass Spectrometry Data

Abstract

Liquid chromatography with tandem mass spectrometry (LC-MS/MS) has been widely used in proteomics. Two types of data, MS and MS/MS data, are produced in an LC- MS/MS experiment. The MS data contains signal peaks corresponding to the intact pep- tides in the samples being analyzed. However, research on protein mass spectrometry data has focused more on extracting information from MS/MS data than on MS data. To effectively utilize MS information, we propose a novel software tool, MSTracer, to detect peptide features from MS data. Two machine-learning-combined scoring functions were incorporated in the implementation: one for detecting the peptide features and another for assigning a quality score to each detected peptide feature. The software was compared with several other tools and demonstrated significantly better performance.