Abstract/Summary

The study of human diseases has evolved from exploring non-causal pathophenotypes to discovering aetiological genetic factors through genome-wide association (GWA) studies. While GWA studies identify disease-associated genetic variants, they fail to reveal their functional implications. The emergence of network biology and molecular interaction mapping helped conceptualise disease as a breakdown of the protein-protein interaction (PPI) network or the ‘interactome’ due to these genetic variants. The interactome drives cellular processes and responds to genetic and environmental changes by leveraging its inherent interdependencies. Simultaneously, the focus has shifted from single-target drugs to the polypharmacological effects of drugs within the PPI network, inducing therapeutic and non-therapeutic effects. Despite these conceptual advances, several factors hinder the widespread application of interactome analysis in disease genetics and drug discovery. These include the lack of an integrated conceptual framework to derive biological and clinical insights from genetic data, the lack of context-sensitive interactomes, the failure to integrate computationally predicted PPIs to circumvent interactome sparsity, and the absence of methods to study correlations across multiple disease interactomes and drug target networks. This thesis proposes a two-pronged interactome-based framework to address these limitations. The first arm focuses on constructing the disease interactomes of complex and Mendelian disorders using both experimentally validated and computationally predicted PPIs, refining them using multi-omics datasets, and deriving insights into disease mechanisms using functional enrichment analyses, identifying repurposable drugs targeting the interactome using comparative analysis of drug-induced and disease-associated transcriptomes, and studying their activity in animal models. The second arm employs multivariate data analysis to explore relationships of multiple interactomes, revealing biological and clinical themes in cross-disorder relationships. This framework has demonstrated its potential by providing insights into eight disorders, identifying disease subgroups, and refining disease categorization based on genetic structures. The methodology yields clinically actionable results, including repurposable drugs and insights into drug activity that can inform safety and efficacy evaluations in clinical trials. This thesis proposes a comprehensive interactome-based framework to uncover hidden patterns in emerging multi-omics disease data and enhance our understanding of disease biology and therapeutics.