Peaks detection and alignment for mass spectrometry data

Abstract

The goal of this paper is to review existing methods for protein mass spectrometry data analysis, and to present a new methodology for automatic extraction of significant peaks (biomarkers). For the pre-processing step required for data from MALDI-TOF or SELDI- TOF spectra, we use a purely nonparametric approach that combines stationary invariant wavelet transform for noise removal and penalized spline quantile regression for baseline correction. We further present a multi-scale spectra alignment technique that is based on identification of statistically significant peaks from a set of spectra. This method allows one to find common peaks in a set of spectra that can subsequently be mapped to individual proteins. This may serve as useful biomarkers in medical applications, or as individual features for further multidimensional statistical analysis. MALDI-TOF spectra obtained from serum samples are used throughout the paper to illustrate the methodology.