Drug repositioning for neurodegenerative diseases based on bioinformatics and text mining analysis

Neurodegenerative diseases (NDDs) are a group of complex diseases with limited treatment options. This thesis presents a comprehensive bioinformatics approach to identify repurposed drug candidates for NDDs, focusing on Alzheimer's Disease (AD), Parkinson's Disease (PD), Huntington's Disease (HD), and Amyotrophic Lateral Sclerosis (ALS). Transcriptomic data including disease and healthy samples were selected and grouped. Differentially expressed genes were identified and co-expression network analysis was performed. Machine learning (ML) algorithms were applied to identify prominent gene clusters. These significant gene clusters were used for drug repositioning analysis. Five ML algorithms (Linear Regression, SVR, Random Forest, Gradient Boosting and Neural Network) were performed to predict the fold change of potential drug candidates. The novelty of these drugs was verified by text mining analysis. For AD, we identified 5 candidate drugs: Dmnq (2,3-dimethoxy-1,4-naphthoquinone), Interferon beta-1b, Cyfluthrin, Torcetrapib, and Vx. For PD, 9 candidate drugs: Interferon beta-1b, Aplidin, Androstanolone, Ribavirin, Dmnq, Natural alpha interferon, Interferon beta-1a, Clinafloxacin, and Bicalutamide. Lastly, for ALS, 5 candidate drugs: Nilotinib, Trovafloxacin, Apratoxin A, Carboplatin, and Clinafloxacin. Among these novel drugs, Dmnq and Interferon beta-1b drugs were common for AD and PD, while Clinafloxacin is common for ALS and PD. Validation of these findings with wet lab studies may provide new treatment options for NDDs. Keywords: drug repurposing, machine learning, neurodegenerative diseases