AutoRevDock: An open-source toolkit for scalable reverse docking.

Reverse docking is a pivotal computational strategy for drug repurposing and polypharmacology studies, yet existing tools often suffer from limitations in throughput, accuracy, and reliance on centralized servers. To overcome these challenges, we present AutoRevDock, an open-source Python toolkit designed to streamline and enhance the reverse docking workflow. Key features include: (1) support for two established docking engines (AutoDock Vina and idock) with a hybrid scoring scheme (Vina_SFCT, combining the Vina score with a scoring function correction term (SFCT)); (2) pre-processed target libraries covering the human proteome and DrugBank pharmacologically active targets; (3) support for custom target libraries and fully automated local execution. Benchmark evaluations demonstrate that idock operates over 40 times faster than AutoDock Vina. For multiple-target drugs, Vina_SFCT outperforms the default scoring function in identifying biologically relevant targets. Furthermore, incorporating protein family information leads to increased hit rates, suggesting enhanced predictive power for real-world applications. By combining robust methodology with user-centric design, AutoRevDock offers a scalable solution for high-throughput target fishing in drug discovery. The toolkit is freely available at https://github.com/AI4Bio-GuoLAB/AutoRevDock.git.