Abstract
In the era of single-particle cryogenic electron microscopy (cryo-EM) and AI-driven protein structure prediction, obtaining high-resolution protein structures, either experimentally or computationally, has become increasingly routine. Yet studying and understanding protein dynamics remains challenging. In single-particle cryo-EM, protein dynamics are most obviously manifested as poor local resolution or disappearing densities in specific regions of a reconstruction. No method is yet available to computationally generate conformational ensembles that fully deconvolute these experimental observations. When dynamics are key to understanding protein function, it is clear to us that introducing new experimental approaches is necessary to close this gap and make sense of invisible densities in single-particle cryo-EM.