MIT Develops A Machine-learning Model That Can Help Determine Protein Structures logo

MIT Develops A Machine-learning Model That Can Help Determine Protein Structures

Feb 04, 2021
In a new Nature Methods study, the Davis and Berger labs combine cryo-electron microscopy and machine learning to visualize molecules in 3D — easily identifying the possible conformations that a protein may take. The department is home to approximately 180 undergraduates, 200 graduate students, 200 postdoctoral researchers, and more than 60 world-renowned faculty, including: - 3 Nobel laureates - 30 members of the National Academy of Sciences - 15 Howard Hughes Medical Institute (HHMI) investigators - 4 recipients of the National Medal of Science We promote a highly collaborative environment that allows for a free exchange of ideas across research areas and academic disciplines. The result is a rigorous, creative, dynamic culture in which scientists and students tackle the important problems and questions in biology and related fields. Our location — at the heart of one of the world’s most important biotech/pharma hubs — creates exciting opportunities for research, employment, and the commercialization of new discoveries. Abstract: Cryo-electron microscopy (cryo-EM) single-particle analysis has proven powerful in determining the structures of rigid macromolecules. However, many imaged protein complexes exhibit conformational and compositional heterogeneity that poses a major challenge to existing three-dimensional reconstruction methods. Here, we present cryoDRGN, an algorithm that leverages the representation power of deep neural networks to directly reconstruct continuous distributions of 3D density maps and map per-particle heterogeneity of single-particle cryo-EM datasets. Using cryoDRGN, we uncovered residual heterogeneity in high-resolution datasets of the 80S ribosome and the RAG complex, revealed a new structural state of the assembling 50S ribosome, and visualized large-scale continuous motions of a spliceosome complex. CryoDRGN contains interactive tools to visualize a dataset’s distribution of per-particle variability, generate density maps for exploratory analysis, extract particle subsets for use with other tools and generate trajectories to visualize molecular motions. CryoDRGN is open-source software freely available at