AQPotency ranks drug candidates in seconds, SandboxAQ says

SandboxAQ has launched AQPotency, a new Large Quantitative Model that rapidly ranks potential drug candidates, even when detailed protein structures are unavailable. Unlike existing computational screening methods limited by the need for structural data, AQPotency delivers results in seconds using standard computing resources.

This capability expands access to targets previously inaccessible to virtual screening, addressing a costly and time-consuming bottleneck in early-stage drug discovery. “AQPotency has given us and our customers a faster, scalable and reliable way to prioritize compounds,” says Andrea Bortolato, SandboxAQ vice-president of drug discovery, “without needing a 3D crystal structure of the target.”

AQPotency Ranks Drug Candidates Without Structural Data

Existing computational methods are often hampered by their reliance on detailed structural data of the target molecule, restricting access to a significant number of potential targets. AQPotency, a Large Quantitative Model, operates effectively even when a detailed protein structure is unavailable, unlike these established techniques. This capability expands the scope of virtual screening, allowing researchers to evaluate compounds for targets previously inaccessible to computational analysis.

The model not only predicts the activity of a candidate molecule but also provides a confidence assessment, indicating the reliability of its performance range. This functionality proves valuable when a molecule exhibits effects with unclear mechanisms, potentially revealing previously unknown biological pathways.

SandboxAQ reports that the model has already been successfully implemented in eight customer programs, with experimental validation confirming its impact, and Bortolato added that “the confidence intervals make the output actionable for biopharma companies.” AQPotency is accessible both through Claude via the Model Context Protocol and directly on the SandboxAQ website, with future availability planned for Google Cloud’s Marketplace.