AlphaFold 3 predicts structures of proteins with DNA, RNA and small molecules

Google DeepMind/AlphaFold

Google DeepMind and Isomorphic Labs published AlphaFold 3, extending structure prediction from single proteins to complexes with DNA, RNA, small molecules and ions. For protein interactions with other molecule types it is at least 50% more accurate than existing methods; scientists can use most of it free through the new AlphaFold Server.

• Predicts structures for proteins, DNA, RNA, ligands and ions, handling chemical modifications controlling cell function.
• 50% more accurate than physics-based tools on drug design benchmarks without structural input.
• Computes entire molecular complexes holistically, revealing how multiple molecules interact simultaneously.
• Free access through AlphaFold Server; enables drug design and structural biology research at scale.