Researchers at ISTA and international collaborators published methods to guide AlphaFold beyond its tendency to collapse heterogeneous structures into a single dominant conformation. The work, reported in Nature Biotechnology, focuses on using experimental data to improve predicted protein ensembles consistent with measurement conditions. The approach addresses a known limitation in protein structure prediction where multiple conformational states may exist biologically, but predictions often default to one model. By incorporating experimental constraints, the team aims to better reflect real-world variability in structure under different measurement regimes. Separate Nature Biotechnology publications referenced in the same research set also highlight retargeted protein engineering tools, but this specific advance is the experimental-guided AlphaFold direction—relevant for drug discovery because it can alter how binding hypotheses and structure-based design decisions are made. For biotech drug developers, ensemble-aware structural prediction improves the quality of computational hypotheses when proteins behave dynamically, particularly in targets where conformational states matter for ligand binding and function.