REPOSITIONING DRUG OF GLP-1 AGONIST IN PARKINSON DISEASE BASED ON MACHINE LEARNING AND MOLECULAR DOCKING

• Main Author: Hasnia Putri, Haliza
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81497
• Institution: Institut Teknologi Bandung

This study aims to identify potential inhibitors of GLP-1 agonist drugs against ?Synuclein protein, relevant for Parkinson's treatment. Using data from CHEMBL (CHEMBL6152) and the crystal structure of the target ?-Synuclein (PDB ID 3Q28), a machine learning model was built with scikit-learn to predict the pIC50 values of 224 GLP-1 agonist compounds. Molecular descriptors were calculated using RDKit and PaDEL-Descriptor to generate data in binary format. As a result, compounds CHEMBL3237911, CHEMBL250091, CHEMBL250310, and CHEMBL442281 emerged as promising candidates based on the predicted pIC50 with high values and showed binding affinity results from molecular docking ranging from 7.9740 to 8.5240 kcal/mol. The results of molecular docking provided strong interaction with the target based on residue analysis identifying several key residues in ligand binding. In conclusion, CHEMBL3237911, CHEMBL250091, CHEMBL250310, and CHEMBL442281 have significant potential as GLP-1 agonist inhibitors for the treatment of Parkinson's, however further experimental testing is required to validate that GLP-1 agonistt can be used as an alternative treatment of Parkinson's.