The Molecular Biophysics and Biochemistry (MB&B) postdoctoral committee was excited to host Dr. Ryan Feathers for a workshop on the use of artificial intelligence (AI) in cryo-electron microscopy (cryo-EM). He is a presidential postdoctoral research fellow at Princeton University, advised by Ellen Zhong and Fred Hughson. Feathers holds a Ph.D. in biochemistry from Cornell University and a B.S. in microbiology and biochemistry from Oklahoma State University.
Cryo-EM provides researchers with the unprecedented ability to capture large, dynamic macromolecules in near-native states and visualize them at near-atomic resolution. As the use of cryo-EM can extend from purified macromolecules to the complex inner workings of a cell, researchers are faced with vast amounts of data that can be challenging to parse.
Computational approaches have been developed to address the different macromolecule dynamics a researcher might observe – a characteristic called conformational heterogeneity, as well as the different combinations of interacting macromolecules, termed compositional heterogeneity.
Feathers and his advisor Ellen Zhong are well known in the structural biology field for their pioneering machine learning software cryoDRGN. CryoDRGN gives researchers a unique tool to explore both the conformational and compositional heterogeneity of their biological systems. The software utilizes neural networks trained to learn the heterogeneity of a specimen. This “learned heterogeneity” can then be mapped back onto a continuous, low-dimensional space which can yield insights that are inaccessible by other methods.
Feathers’s workshop was a highly anticipated event, and the seminar room was packed with trainees and professors alike, all excited to learn about the latest advances in cryo-EM software. Feathers demonstrated the power of cryoDRGN in analyzing challenging datasets and gave a step-by-step tutorial on the usage of the software.
Dr. Shenhui Liu, a postdoc in MB&B department’s Tang Lab remarked, “Ryan gave a comprehensive introduction to cryoDRGN’s applications in both single particle cryo-EM analysis and cryo-electron tomography (cryo-ET) with practical tutorials. CryoDRGN provides a powerful solution to the challenges in my current project, and I am now implementing this tool to overcome existing bottlenecks.”
Another postdoc commented, “Ryan’s workshop was very inclusive for all levels of structural biologists, especially those of us who are brand new to the field of cryo-EM.” The faculty members in attendance also had great things to say. Jun Liu, Professor of Microbial Pathogenesis, of Molecular Biophysics and Biochemistry, and of Cell Biology, known for his innovative cryo-electron tomography (cryo-ET) research was very excited. Reflecting on the workshop, he said, “It was cutting edge!”
Ailong Ke, Eugene Higgins Professor of Molecular Biophysics and Biochemistry similarly commented, saying “Ryan brings a cutting-edge technique to solve complex problems in structural biology. This would benefit our students and postdocs in their structural biology research.”
AI tools in cryo-EM may prove invaluable in yielding new biological insights, transforming heterogeneity from a challenge to an opportunity.
One of Feathers’ postdoc advisors, Fred Hughson, was a Yale MB&B undergraduate, Class of 1984.
For more information on cryoDRGN, see the tutorial and slides provided at: http://tinyurl.com/cryodrgnyale (courtesy of Ryan Feathers and the Zhong Lab).
Written by Rithik Castelino.