Enabling lipidomics though bioinformatics and network biology
Publication Date
July 31, 2020
Creator
Abstract
Human disease and medical research have advanced greatly, thanks to developments in 'omics' techniques. Lipidomics focuses on the study of lipid profiles in biological samples. It is a growing area of research and development, especially with a view to finding clinical biomarkers and uncovering novel cellular mechanisms. However, the analysis and interpretation of raw data have been slow and disconnected from other information resources. Improvements to analysis time and novel capability can be enabled through the use of Bioinformatics. This thesis outlines three areas in which it has been applied.
In the first, a network, linking lipid molecules to the reactions in which they participate and the associated enzymes, can be queried using a set of web-enabled scripts. This software identifies and ranks the enzymes potentially causing the observed perturbations (between conditions) in lipid profiles and also generates a network specific to the perturbations. In turn, these enzymes can be used as queries to other databases to identify potential gene-expression and protein-interaction regulators. This tool has been queried with a polycystic ovary syndrome dataset, revealing an important role for Phospholipase D1. The results suggest the potential involvement of the transcription factors MAZ, NFAT and STAT5B in the disease.
The second area focuses on oxylipins, which are signalling molecules involved in inflammation and other physiological responses and are the subject of targeted lipidomic studies. An interface for converting lipid perturbations to a colour-coded lipidomic network has been developed and database searches with the enzymes concerned revealed that they form membrane-bound multi-protein complexes of both enzymes and specific oxylipin receptors. Furthermore, transcriptional factors for certain enzymes have been identified, which suggest potential regulatory mechanisms. This software was used to study 5 datasets from investigations into Osteo-Arthritis, and the results are consistent with both differential regulation of EPHX2 and CYP2B6, and thromboxane levels usually (but not exclusively) following those of prostaglandins.
The third area concerns a set of solutions to the most common challenges in lipidomics-data analysis. These scripts can also be used for metabolomics data in general. The specific software includes the design of a method to pair C13-marked molecular species, data 3 reformatting, improved lipid identification, and a study of lipid-adduct formation with recommendations on how to identify and compensate for their presence.
The outcome of this work has been much reduced data-analysis times and new tools to understand the biological significance of changes in lipid profiles.
Item Type
ethesis
Thesis Type
PhD
Supervisors
Subjects (LC)
Associated Schools / Departments
School of Biosciences (UK)
eprints ID
57166
UoN Repository URI
Except where otherwise noted, this item's license is described as
File(s)![Thumbnail Image]()
Name
Full-Thesis-Ferran-Casbas-8.0.pdf
Type
Full-text
Description
Examined
Size
5.24 MB
Format
Adobe PDF
Checksum (MD5)
ebb25aea57c51f083a411900778dcd83