Mechanism-based Stratification of Alzheimer's and Parkinson's Disease using Artificial Intelligence

The capacity to generate omics and clinical data in biomedical science is growing exponentially over the past decades. Additionally, recent advances in computational power and analyzing capabilities have resulted in overwhelmingly increased interest in the use of big data to solve most problems in biomedical science. Drug discovery and molecular disease taxonomies are two of the most pressing challenges in biomedical science that could be solved by the surge of big data. Hence, there is an urgent need for developing methods that incorporate biomedical data and prior knowledge for drug development and patient stratification in order to achieve the goal of stratified medicine. In this thesis, we address the aforementioned issues in the context of neurodegenerative diseases. First, we demonstrate a pure knowledge-driven approach for mechanism-based drug repositioning in Alzheimer's disease by curating and analyzing Alzheimer's disease knowledge assembly. Second, we present PS4DR, a drug repositioning workflow that is based on the combination of knowledge- and data-driven approaches. This work combines canonical pathway information and multi-omics data in order to predict drugs that can alter disease etiology. Finally, we showcase a hybrid artificial intelligence-based approach to jointly stratify Alzheimer's and Parkinson's disease patients based on the omics data and prior knowledge of shared molecular mechanisms of the two diseases. The established patient subgroups are reproducible and can be associated with different clinical and molecular disease features.
Finally, this thesis attempts to connect the knowledge- and data-driven strategy for solving two very interesting biomedical problems of drug discovery and patient stratification by using prior knowledge, multi-omics, imaging, and clinical data. Overall, this work is a step towards achieving a better targeted and thus more effective therapy in neurology to reach the ultimate goal of precision medicine concept.