Generalized biomolecular modeling and design with RoseTTAFold All-Atom

• Published in: Science (American Association for the Advancement of Science) Vol. 384; no. 6693; p. eadl2528
• Main Authors: Krishna, Rohith, Wang, Jue, Ahern, Woody, Sturmfels, Pascal, Venkatesh, Preetham, Kalvet, Indrek, Lee, Gyu Rie, Morey-Burrows, Felix S., Anishchenko, Ivan, Humphreys, Ian R., McHugh, Ryan, Vafeados, Dionne, Li, Xinting, Sutherland, George A., Hitchcock, Andrew, Hunter, C. Neil, Kang, Alex, Brackenbrough, Evans, Bera, Asim K., Baek, Minkyung, DiMaio, Frank, Baker, David
• Format: Journal Article
• Language: English
• Published: United States The American Association for the Advancement of Science 19.04.2024
• Subjects: Accuracy; Additives; Amino acids; Amino Acids - chemistry; Assemblies; Binding; Bond graphs; Cell signaling; Chemical bonds; Chemical elements; Chemistry; Chirality; Coronary artery disease; Covalence; Crystal structure; Crystallization; Crystallography; Deep Learning; Design; Digoxigenin; DNA - chemistry; Gas pipelines; Graph representations; Graphical representations; Heart diseases; Heme; Homology; Ligands; Machine learning; Metal ions; Metals; Modelling; Models, Molecular; Networks; Neural networks; Nucleic acids; Optical activity; Organic Chemistry; Physical chemistry; Polypeptides; Predictions; Protein Engineering - methods; Protein structure; Proteins; Proteins - chemistry; Residues; Similarity; Structure-function relationships
• ISSN: 0036-8075; 1095-9203; 1095-9203
• DOI: 10.1126/science.adl2528

Deep-learning methods have revolutionized protein structure prediction and design but are presently limited to protein-only systems. We describe RoseTTAFold All-Atom (RFAA), which combines a residue-based representation of amino acids and DNA bases with an atomic representation of all other groups to model assemblies that contain proteins, nucleic acids, small molecules, metals, and covalent modifications, given their sequences and chemical structures. By fine-tuning on denoising tasks, we developed RFdiffusion All-Atom (RFdiffusionAA), which builds protein structures around small molecules. Starting from random distributions of amino acid residues surrounding target small molecules, we designed and experimentally validated, through crystallography and binding measurements, proteins that bind the cardiac disease therapeutic digoxigenin, the enzymatic cofactor heme, and the light-harvesting molecule bilin. Advances in machine learning have made protein structure prediction and design much more accurate and accessible in recent years, but these tools have generally been limited to polypeptide chains. However, ligands such as small molecules, metal ions, and nucleic acids are crucial components of most proteins, both in terms of structure and biological function. Krishna et al . present a next-generation protein structure prediction and design tool, RoseTTAFold All-Atom, that can accept a wide range of ligands and covalent amino acid modifications. The authors demonstrate superior performance on protein-ligand structure prediction relative to other tools, even in the absence of an input experimental structure. They also perform de novo design of proteins to bind cofactors and small molecules and experimentally validate these designs. —Michael A. Funk