SandboxAQ launches AQPotency for virtual drug discovery

SandboxAQ has announced the general availability of AQPotency, a Large Quantitative Model (LQM) designed to accelerate drug discovery through virtual screening. The tool allows researchers to predict how effectively potential drug molecules will act on disease targets without requiring a previously solved protein structure.

Traditional computational methods often rely on detailed structural maps of disease targets, which can stall research when such data is unavailable. AQPotency addresses this by ranking enormous libraries of molecules in seconds using standard computing hardware. The model provides both a prediction of activity and a confidence assessment, informing scientists whether a target falls within the model’s reliable performance range.

The technology also supports reverse analysis, allowing researchers to input a single molecule to identify a ranked list of proteins it is likely to interact with. This can assist in identifying the mechanisms of action for molecules that show effects without a known cause. AQPotency is accessible via Claude through the Model Context Protocol (MCP) and the SandboxAQ website, with future plans for integration into Google Cloud’s Marketplace.