Discovery of novel NLRP3 inhibitors based on machine learning and physical methods

Abstract

Abstract The NLRP3 inflammasome plays a crucial role in inflammatory responses, particularly in alcohol-related liver disease (ALD). Given that NLRP3 has emerged as a potential therapeutic target for ALD, the development of effective inhibitors is of great importance. In this study, we trained 11 regression models, and the results showed that LightGBM, Random Forest, and XGBoost performed the best, achieving R² values of 0.774, 0.755, and 0.719, respectively. Using machine learning models and physical methods, we screened more than 11.5 million compounds from Asinex, Princeton, UkrOrgSynthesis, Chemdiv, Chembridge, Alinda, Enamine, and Lifechemicals, which led to the identification of 26 potential NLRP3 inhibitors. Furthermore, molecular dynamics simulations and MMGBSA binding energy calculations confirmed the stability of the interactions between NLRP3 and three key molecules: 19,655,631 (source Chembridge), 38,214,692 (source Chembridge), and Z1180203703 (source Enamine). Additionally, ADMET analysis revealed their favorable pharmacokinetic properties. This study provides insights and candidate molecules for discovering NLRP3 inhibitors, potentially applicable in treating related diseases.

Keywords

• NLRP3 inflammasome
• NLRP3 inhibitors
• Machine learning
• Molecular docking
• Molecular dynamics simulation
• Drug discovery