Method for differential detection and identification of components in protein mixtures analyzed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

We demonstrate that the semi-quantitative information in matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectra of tryptically digested protein mixtures can, via a systematic statistical approach, be utilized for the identification of a protein present in different concentrations in two samples. Multiple mass spectra were acquired from a series of tryptically digested test samples in which the concentration of one protein was varied and the concentrations of three other proteins were held constant. The mass spectra were subjected to soft independent modeling of class analogy (SIMCA) analysis assuming that spectra originating from two different samples belonged to different data classes. The SIMCA analysis yielded information on which individual m/z values discriminate between two classes. Protein identification by proteolytic peptide mass fingerprinting was performed with different numbers of mass values in the fingerprint according to the discriminatory information, beginning with the mass corresponding to the best discrimination, followed by the best together with the second best, etc. By using the Probity algorithm, which computes the statistical significance of each identification result, we demonstrate that the first protein identified at a desired significance level (0.001) is the protein that was present in a different concentration in the two samples. Differential analysis of expression is often performed by comparing 2D-gel-spot intensities followed by mass spectrometric identification of the respective protein in each spot that differs. The method presented here has the potential to allow identification of the protein component that differs in cases where a gel-spot is poorly resolved and contains several proteins. Copyright (C) 2004 John Wiley Sons, Ltd.