Predictive Modeling of Detoxification Processes for Highly Toxic Organophosphorus Agents Using Interpretable Molecular Parameters

Highly toxic chemical warfare agents (CWSAs) based on organophosphorus compounds can be detoxified through hydrolysis. A new model has been introduced to predict the reaction barriers in the alkaline hydrolysis of G-series agents, focusing on their molecular structure. This model was developed using the largest dataset of reaction barriers ( ΔG _TS ) for 122 organophosphate compounds. It outperforms existing complex quantitative structure-activity relationship (QSAR) models, showing lower root mean squared errors (RMSE) across training, testing, and validation datasets. The new model's RMSE values are 4.74, 1.92, and 3.03, respectively, compared to the complex QSAR model's 8.00, 4.36, and 11.38. Additionally, it effectively covers 56 organophosphorus chemicals without measured ΔG _TS data, making it a robust tool for identifying improved simulants and examining structural factors affecting organophosphate reaction energetics.