Use this URL to cite or link to this record in EThOS:
Title: Integration and visualisation of clinical-omics datasets for medical knowledge discovery
Author: Homola, Daniel
ISNI:       0000 0004 7657 991X
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
In recent decades, the rise of various omics fields has flooded life sciences with unprecedented amounts of high-throughput data, which have transformed the way biomedical research is conducted. This trend will only intensify in the coming decades, as the cost of data acquisition will continue to decrease. Therefore, there is a pressing need to find novel ways to turn this ocean of raw data into waves of information and finally distil those into drops of translational medical knowledge. This is particularly challenging because of the incredible richness of these datasets, the humbling complexity of biological systems and the growing abundance of clinical metadata, which makes the integration of disparate data sources even more difficult. Data integration has proven to be a promising avenue for knowledge discovery in biomedical research. Multi-omics studies allow us to examine a biological problem through different lenses using more than one analytical platform. These studies not only present tremendous opportunities for the deep and systematic understanding of health and disease, but they also pose new statistical and computational challenges. The work presented in this thesis aims to alleviate this problem with a novel pipeline for omics data integration. Modern omics datasets are extremely feature rich and in multi-omics studies this complexity is compounded by a second or even third dataset. However, many of these features might be completely irrelevant to the studied biological problem or redundant in the context of others. Therefore, in this thesis, clinical metadata driven feature selection is proposed as a viable option for narrowing down the focus of analyses in biomedical research. Our visual cortex has been fine-tuned through millions of years to become an outstanding pattern recognition machine. To leverage this incredible resource of the human brain, we need to develop advanced visualisation software that enables researchers to explore these vast biological datasets through illuminating charts and interactivity. Accordingly, a substantial portion of this PhD was dedicated to implementing truly novel visualisation methods for multi-omics studies.
Supervisor: Holmes, Elaine ; Nicholson, Jeremy ; Guo, Yike Sponsor: Wellcome Trust
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral