Integrative computational and experimental approaches to visualize highly heterogeneous structural ensembles through cryo-EM and cryo-ET

The National Institute of General Medical Sciences awarded Massachusetts Institute of Technology $269,071 on July 1, 2026, under the Biomedical Research and Research Training program (CFDA 93.859) to develop and apply cryo-electron microscopy and cryo-electron tomography methods for visualizing protein machine assembly, function, and regulation at high resolution. The work emphasizes development of the CryoDRGN and TomoDRGN software ecosystem to enable structural analysis of complex samples in near-native cellular extracts and directly in cells, with applications focused on ribosome assembly and regulation, co-translational protein folding, and protein degradation. The project integrates computational and experimental approaches including direct-from-lysate imaging, time-resolved cryo-EM, and machine learning-based analyses to characterize structural heterogeneity and generate insights into cellular proteostasis mechanisms. Performance runs through June 30, 2031, at Cambridge, Massachusetts. The award is a Project Grant, the standard assistance type for investigator-initiated research under the NIGMS portfolio.

Abstract

Cells rely on precisely assembled and regulated protein machines to sustain life, and disruptions in these processes underlie many human diseases. This R35/mira grant supports our continued development and application of innovative cryo-electron microscopy (cryo-EM) and cryo-electron tomography (cryo-ET) methods designed to visualize these essential protein machines, emphasizing the resolution and interpretation of structural changes these machines undergo as they assemble, function, and are regulated. This work includes our continued development of the cryodrgn and tomodrgn ecosystem, which will broadly impact the structural biology community by enabling thorough structural analyses of complex samples both in near-native cellular extracts, and directly in cells. We will additionally apply these tools in our group's key areas of focus: ribosome assembly and regulation, CO-translational protein folding, and protein degradation. By integrating novel experimental strategies such as direct-from-lysate imaging, time-resolved cryo-EM, and advanced machine learning-based analyses, we expect this work to yield critical insights into fundamental cellular mechanisms supporting proteostasis. Moreover, the tools we develop will enhance community-wide capabilities for structural biology, which we expect to substantially improve the accuracy and interpretability of structural analyses critical to biological discovery.