Gaudelet, Thomas;
(2021)
Integration of multi-scale protein interactions for biomedical data analysis.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Main.pdf - Accepted Version Download (11MB) | Preview |
Abstract
With the advancement of modern technologies, we observe an increasing accumulation of biomedical data about diseases. There is a need for computational methods to sift through and extract knowledge from the diverse data available in order to improve our mechanistic understanding of diseases and improve patient care. Biomedical data come in various forms as exemplified by the various omics data. Existing studies have shown that each form of omics data gives only partial information on cells state and motivated jointly mining multi-omics, multi-modal data to extract integrated system knowledge. The interactome is of particular importance as it enables the modelling of dependencies arising from molecular interactions. This Thesis takes a special interest in the multi-scale protein interactome and its integration with computational models to extract relevant information from biomedical data. We define multi-scale interactions at different omics scale that involve proteins: pairwise protein-protein interactions, multi-protein complexes, and biological pathways. Using hypergraph representations, we motivate considering higher-order protein interactions, highlighting the complementary biological information contained in the multi-scale interactome. Based on those results, we further investigate how those multi-scale protein interactions can be used as either prior knowledge, or auxiliary data to develop machine learning algorithms. First, we design a neural network using the multi-scale organization of proteins in a cell into biological pathways as prior knowledge and train it to predict a patient's diagnosis based on transcriptomics data. From the trained models, we develop a strategy to extract biomedical knowledge pertaining to the diseases investigated. Second, we propose a general framework based on Non-negative Matrix Factorization to integrate the multi-scale protein interactome with multi-omics data. We show that our approach outperforms the existing methods, provide biomedical insights and relevant hypotheses for specific cancer types.
Type: | Thesis (Doctoral) |
---|---|
Qualification: | Ph.D |
Title: | Integration of multi-scale protein interactions for biomedical data analysis |
Event: | UCL (University College London) |
Open access status: | An open access version is available from UCL Discovery |
Language: | English |
Additional information: | Copyright © The Author 2021. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
UCL classification: | UCL UCL > Provost and Vice Provost Offices UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science |
URI: | https://discovery.ucl.ac.uk/id/eprint/10125012 |
Archive Staff Only
View Item |