A Machine Learning and Networks Approach to Infer Disease Mechanisms
Abstract
All biological problems, including cancer, mental disorders, rare diseases, and immunological disorders, are complex multisystem pathophysiological shifts facilitated by functional changes. Multiple points of origin and mechanisms can contribute, including genetic mutations, epigenetic factors, protein misfolding, and differential gene expression, which promote such functional abnormalities. This dissertation describes mega-analysis and machine learning methods to identify and select essential entities that are prime candidates for drivers of such disorganization. Furthermore, heterogeneous networks have been employed to achieve a more thorough pathway analysis that takes protein-protein interactions into account. This dissertation contains four projects, investigating PPARD’s role in major depressive disorder, exploring the link between schizophrenia and myocardial infarctions, identifying the connections between atopic dermatitis and major depressive disorder, and finally, biomarker identification for survival status prediction of patients with brain cancer.
Description
Keywords
Biomarker Discovery, Graph Databases, Machine learning, Mage-Analysis, Pathway Analysis, Protein-Protein Interactions