Machine Learning Based ADMET Prediction in Drug Discovery

A drug is a substance or compound with the ability to treat or prevent disease in people or animals. Drug discovery is the process of identifying prospective medications for certain ailments, and it can take several years to accomplish. However, the introduction of machine learning has simplified the procedure. Virtual screening’s major goal is to minimize the virtual compound space of biological molecules. To do this, we developed the VIRTUAL SCREENING TOOL, a virtual molecular descriptor calculator that can analyze molecules and assess their drug-likeness based on specific rules. We used Django with PyCharm, a Python IDE, to create a molecular description calculator, fingerprint generator, and prediction interface. The tool begins by inputting SMILES, which is the chemical structure of the molecule, and outputs molecular descriptors, which are the properties of the molecules, including their physicochemical properties, absorption, distribution, metabolism, lipophilicity, excretion, toxicity, drug-likeness rules, and fingerprint generation. The outputs are available in CSV format for download.