Artificial Intelligence- and Machine Learning-Driven QSAR Modeling, Docking, and MD Simulation for the Discovery of 1,3,4-Thiadiazole Derivatives as VEGFR-2 Inhibitors

Introduction: Cancer remains a major global health burden, with breast cancer being one of the most prevalent malignancies worldwide. Current treatment options are hindered by toxicity, limited efficacy, and drug resistance, underscoring the urgent need for safer and more effective therapies. Targeting vascular endothelial growth factor receptor-2 (VEGFR-2), a key mediator of tumor angiogenesis, represents a promising therapeutic strategy. However, existing VEGFR-2 inhibitors often lack selectivity and are associated with adverse effects, limiting their clinical utility. Method: This study employs both qualitative and quantitative methods based on AI-ML approaches. It involves QSAR modeling, virtual screening, molecular docking, molecular dynamics (MD), and ADMET profiling to identify potent VEGFR-2 inhibitors among 1,3,4-thiadiazole scaffolds. Result: The dataset for modeling was retrieved from PubMed literature, while data for QSAR screening were collected from ChEMBL and analyzed using PaDEL descriptors. Machine learning models, including Support Vector Machines, k-Nearest Neighbors, Random Forest, Gradient Boosting, Linear Regression, and Decision Trees, were used to predict anticancer activity. Gradient Boosting was identified as the most robust model based on R², root mean square error (RMSE), and concordance correlation coefficient (CCCtest). Drug-likeness screening and molecular docking suggested drug-like characteristics and strong docking scores with favorable binding interactions with VEGFR-2 (4ASD). Among the screened compounds, compound 21604 was identified as the most promising candidate based on a 100 ns MD simulation performed using the OPLS_2004 force field. Pharmacokinetic and toxicity analyses revealed that the compounds were minimally toxic or non-toxic and exhibited good oral bioavailability. Discussion: The study successfully identified compound 21604 as a potential VEGFR-2 inhibitor with a higher docking score, stronger interactions, and more favorable pharmacokinetic properties than Axitinib and Sorafenib. Conclusion: The results demonstrate the efficacy of AI-ML approaches in accelerating anticancer drug discovery, providing a rational framework for identifying novel VEGFR-2 inhibitors and advancing targeted cancer therapeutics.