Answers are grounded in this paper and cite their sources.
After sparking a revolution in protein structure prediction, machine learning (ML), and artificial intelligence (AI) are now reshaping the modeling of protein conformational ensembles and dynamics. A fast, scalable, and accurate representation of protein conformational ensembles could enable genome-scale studies of protein function, with broad applications in drug discovery and protein engineering. The challenges associated with this goal, however, are far greater than those involved in single structure prediction because of the broad timescales spanned by protein motion, the complex relationship between sequence and dynamics, and the scarcity of datasets with experimental measurements. Despite these difficulties, many promising solutions have emerged in recent years, ranging from accelerated molecular dynamics to direct sampling of the conformational distribution, as well as some hybrid methods. Here, we review recent AI-powered approaches to tackle protein dynamics and conformational sampling, and outline their current limitations.
No takes yet. Share an insight, caveat, or question.
Pasos-Trejo et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: