A strategy for the prior processing of high‐resolution mass spectral data obtained from high‐dimensional combined fractional diagonal chromatography

Abstract

Combined fractional diagonal chromatography (COFRADIC) is a novel suite of gel-free technologies for the identification of biomarkers in complex peptide mixtures. For this purpose, reversed-phase high performance liquid chromatography (HPLC) technology and, in this case, matrix assisted laser desorption /ionization-time of flight (MALDI-TOF) mass spectrometers are extensively used. The particular characteristic of COFRADIC mass spectrometry data is the high number of chromatographic fractions, over which a peptide can be scattered. This can obstruct the quantification of the peptide abundance in the biological sample, which is required for statistical analysis. On the other hand, because of the superior peptide sorting properties of the methodology, the mass spectra become less crowded. Consequently, each peptide appears in a mass spectrum as a series of peaks with peak heights proportional to the probability of occurrence of the isotopic variants of the peptide. In this manuscript, we propose an analysis strategy concerned with the preprocessing of COFRADIC mass spectra prior to a downstream statistical analysis. The preprocessing algorithm produces for each mass spectrum a peptide list by exploiting the characteristic features that should be associated with peaks corresponding to an isotopically resolved cluster of peptide peaks. This reduction step is necessary to facilitate the clustering used in a next step to assemble the validated monoisotopic peptide peaks found over several fractions into a single peptide abundance. To assess the performance of the algorithm, two technical experiments were conducted. The proposed strategy is memory and computationally efficient.

A Bayesian Model Averaging Approach to the Quantification of Overlapping Peptides in an MALDI-TOF Mass Spectrum

“…It should be noted that the validity of this approach is based on a proper-preprocessing procedure (for details of preprocessing, refer to Vaikenborg et al [11]). More specifically, it assumes that a cluster of peptide peaks is correctly found after noise filtering.…”

A Bayesian Markov-Chain-Based Heteroscedastic Regression Model for the Analysis of 18O-Labeled Mass Spectra

“…To this aim, we use the strategy proposed by Valkenborg et al [5]. The pre-processing strategy extracts the information about the m/z location and the height (intensity) of peaks, which are most likely due to a peptide.…”

A Markov-chain-based regression model with random effects for the analysis of18O-labelled mass spectra

“…To this aim, we use the strategy proposed by Valkenborg et al [10]. The pre-processing strategy extracts the information about the mass location (m/z) and the height (intensity) of peaks, which are most likely due to a peptide.…”