Work

Mass Spectrometry-based Proteomics: a 3d Approach to Data Handling and Quantification

Year: 2011
Type: dissertation
Abstract: This thesis describes the Ph.D. research project in Bioengineering for Computational Proteomics carried out during the last three years (January 2008 - January 2011). Activities focused on design and development of methods for the analysis of Quantitative Mass Spectrometry-based Proteomics data. The Introduction briefly elucidates the main themes developed in the thesis and how the work was schemed. It reviews the computational issues associated to both data handling and quantification, and introduces the solutions proposed in the following. The first two chapters are introductory to the Proteomics and Mass Spectrometry field. The objective is to provide the reader with the information needed to understand Quantitative Mass Spectrometry-based Proteomics. In particular, Chapter 1 explains how proteomics was born, as the –omics science of proteins. Then proteomics main applications and goals are illustrated, which are ranging from clinics and pharmaceutics to systems biology. Chapter 2 shows the main technologies and instrumentation exploited in Mass Spectrometry-based proteomics. The most common experimental setups are reported: among them, the Liquid Chromatography-Mass Spectrometry (LC-MS) technique is thoroughly explained since it is the principal technique for Quantitative Mass Spectrometry-based Proteomics. The third Chapter presents the main concepts necessary to introduce the reader to the main topic of the PhD research Project, that is the development of bioinformatics tools for the handling and quantification of Mass Spectrometry-based Quantitative Proteomics data, focusing on LC-MS quantitative data and their analysis. Indeed, LC-MS data are highly informative for quantification aims, but challenging to parse. Data features that were pivotal for the design of the proposed solutions (i.e., the 3D structure of LC-MS data and the high quality profile acquisition) are highlighted. In the fourth Chapter, the state of art both for data handling and quantification is described and available standard data formats and software are illustrated as well as related open challenges. In Chapter 5, the dataset used to carry out the analyses is technically described. It consists of LC-MS data from a labeled controlled mixture of proteins with known quantification ratios, acquired in profile acquisition mode and in triplicates. In particular, this thesis presents 2 software solutions to address the handling and quantification of Quantitative Mass Spectrometry-based Proteomics data: mzRTree and 3DSpectra, respectively. Chapter 6 presents the solution proposed for the data handling issue. The proposal is a scalable 2D indexing approach implemented through an R-tree-based data structure, called mzRTree, that relies on a sparse matrix representation of the dataset, which is appropriate for LC-MS data, and more in generally for MS-based proteomics data. mzRTree allows efficient data access, storage and enables a computationally sustainable analysis of profile MS data. Regarding the quantification, which is one of the most relevant problem in mass spectrometry-based proteomics, Chapter 7 illustrates the solution proposed for the quantification problem: 3DSpectra. It is an innovative quantification algorithm for LC-MS labeled profile data exploiting both the 3-dimensionality of data and the profile acquisition. 3DSpectra fits on peptide data the 3D isotopic distribution model shaped by a Gaussian Mixture Model including a noise component, using the Expectation-Maximization approach. This model enables the software to both recognize the borders of the 3D isotopic distribution and reject noise. 3DSpectra is a reliable and accurate quantification strategy for labeled LC-MS data, providing significantly wide and reproducible proteome coverage. In the conclusion section of this thesis future and ongoing research work, regarding further development of both the mzRTree data structure and 3DSpectra quantification software, are discussed.
Author: Sara Nasso
Language: English
Subfield: Spectroscopy
Field: Chemistry
Open Access status: green
OpenAlex ID: W2277691344
Cites: 46
Cited by:
Related to: 10