Enhanced label-free discovery proteomics through improved data analysis and knowledge enrichment
Year: 2022
Type: dissertation
Abstract: Mass spectrometry (MS)-based proteomics has evolved into an important tool applied in fundamental biological research as well as biomedicine and medical research. The rapid developments of technology have required the establishment of data processing algorithms, protocols and workflows. The successful application of such software tools allows for the maturation of instrumental raw data into biological and medical knowledge. However, as the choice of algorithms is vast, the selection of suitable processing tools for various data types and research questions is not trivial. In this thesis, MS data processing related to the label-free technology is systematically considered. Essential questions, such as normalization, choice of preprocessing software, missing values and imputation, are reviewed in-depth. Considerations related to preprocessing of the raw data are complemented with exploration of methods for analyzing the processed data into practical knowledge. In particular, longitudinal differential expression is reviewed in detail, and a novel approach well-suited for noisy longitudinal high-througput data with missing values is suggested. Knowledge enrichment through integrated functional enrichment and network analysis is introduced for intuitive and information-rich delivery of the results. Effective visualization of such integrated networks enables the fast screening of results for the most promising candidates (e.g. clusters of co-expressing proteins with disease-related functions) for further validation and research. Finally, conclusions related to the prepreprocessing of the raw data are combined with considerations regarding longitudinal differential expression and integrated knowledge enrichment into guidelines for a potential label-free discovery proteomics workflow. Such proposed data processing workflow with practical suggestions for each distinct step, can act as a basis for transforming the label-free raw MS data into applicable knowledge.
Source: UTUPub (University of Turku)
Author: Välikangas, Tommi
Language: English
Subfield: Spectroscopy
Field: Chemistry
Domain: Physical Sciences
Open Access status: green
Funder: Turun yliopiston tutkijakoulu
OpenAlex ID: W7037775003
