SAIR: Accelerating Pharma R&D with AI-Powered Structural Intelligence

AI Impact Summary

SandboxAQ has released SAIR, a massive dataset of 5.24 million co-folded protein-ligand structures paired with experimentally measured IC50 values. This dataset, generated using a cluster of NVIDIA H100 GPUs, addresses a critical data gap in AI-powered drug design by providing high-accuracy structural information for targets previously lacking experimental validation. The dataset’s creation involved significant computational optimization, achieving a 4x speedup compared to the initial estimate, and incorporates rigorous validation using PoseBusters to ensure structural quality.