A machine learning approach to explore the spectra intensity pattern of peptides using tandem mass spectrometry data

Background: A better understanding of the mechanisms involved in gas-phase fragmentation ofpeptides is essential for the development of more reliable algorithms for high-throughput proteinidentification using mass spectrometry (MS). Current methodologies depend predominantly on theuse of derived m/z values of fragment ions, and, the knowledge provided by the intensityinformation present in MS/MS spectra has not been fully exploited. Indeed spectrum intensityinformation is very rarely utilized in the algorithms currently in use for high-throughput proteinidentification.Results: In this work, a Bayesian neural network approach is employed to analyze ion intensityinformation present in 13878 different MS/MS spectra. The influence of a library of 35 features onpeptide fragmentation is examined under different proton mobility conditions. Useful rulesinvolved in peptide fragmentation are found and subsets of features which have significant influenceon fragmentation pathway of peptides are characterised. An intensity model is built based on theselected features and the model can make an accurate prediction of the intensity patterns for givenMS/MS spectra. The predictions include not only the mean values of spectra intensity but also thevariances that can be used to tolerate noises and system biases within experimental MS/MS spectra.Conclusion: The intensity patterns of fragmentation spectra are informative and can be used toanalyze the influence of various characteristics of fragmented peptides on their fragmentationpathway. The features with significant influence can be used in turn to predict spectra intensities.Such information can help develop more reliable algorithms for peptide and protein identification.