< Terug naar vorige pagina

Publicatie

Statistical Methods for the Analysis of High-resolution Mass Spectrometry Data

Boek - Dissertatie

Current developments in high-throughput techniques for life sciences have resulted in ever growing amounts of data. In turn, this has led to a situation where the interpretation of data and the formulation of hypotheses lag the pace, at which the information is produced. An example of such a high-throughput technique are technologies for mass spectrometry-based quantitative proteomics that measure the expression of thousands of proteins/peptides simultaneously (MS1). Another, less high-throughput, mass spectrometry-based application is the identification of the protein/peptide amino acid structure (MS2). Current analysis strategies involve data-driven methods based on the expression (MS1) to select proteins/peptides for a second interrogation on the mass spectrometer (MS2). This approach does not optimize the workload for the mass spectrometer, can have dramatic effects on the dynamic range, and can lead to undersampling of the proteins/peptides present in the biological sample. We argue that a targeted approach, which select proteins/peptides based on a statistical analysis from different biological samples, may improve the workflow, increase the dynamic range, and can avoid undersampling. However, this requires a conversion of the data flood into protein/peptide-related information, which constitutes a major challenge. The results presented in this dissertation address this problem in three ways by: (1) suggesting a method for the prediction of characteristic features related to protein/peptide peaks observed in a mass spectrum; (2) proposing an approach for the prior processing of mass spectra to extract and quantify biological relevant features, which may facilitate a down-stream statistical analysis. (3) developing an alternate strategy for the relative quantification of the protein/peptide abundance from enzymatic O-labeled mass spectra. ...
Jaar van publicatie:2008
Toegankelijkheid:Open