{"entity":{"id":"bioemu","kind":"technology","name":"BioEmu (Microsoft)","aka":[],"tldr":"BioEmu is a Microsoft generative diffusion model that predicts the range of shapes a protein moves between, not one static structure, thousands of times faster than molecular dynamics simulation. For cancer drug discovery that can reveal transient pockets, as in KRAS, that static predictors miss, but its outputs are approximate and validated mainly on small proteins.","summary":"BioEmu from Microsoft Research is a generative diffusion model that emulates the equilibrium ensemble of shapes a protein moves between, trained on molecular dynamics simulations and experimental data. Rather than predicting one static structure, the Science 2025 paper shows it sampling conformational ensembles and estimating folding free energies thousands of times faster than simulation. For cancer drug discovery this matters for cryptic pockets, binding sites that only open transiently, as in KRAS, which static structure predictors cannot reveal. Its predictions are approximate, and the published work focuses on small proteins, so large complexes and membrane proteins remain beyond validated use. For a newcomer: BioEmu predicts how a protein wiggles, which can reveal hidden pockets a drug might fit into.","status":"emerging","asOf":"2026-09-08","links":[{"label":"Science 2025","url":"https://doi.org/10.1126/science.adv9817"}],"tags":["foundation-model","structure"],"related":[],"cancers":[],"sections":["drug-discovery","ai-computation"],"technologies":[],"targets":["kras"],"drugs":[],"companies":["microsoft-research"],"institutions":[],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-lewis-science"],"journals":[],"dependsOn":[],"notes":[],"principle":"Diffusion over conformational ensembles trained on MD and experiment.","strengths":["Dynamics at scale"],"limitations":["Approximate; small-protein focus"],"since":2025},"route":"/technologies/bioemu/","neighbours":{"section":[{"id":"ai-computation","kind":"section","name":"AI & Computation","route":"/fronts/ai-computation/"},{"id":"drug-discovery","kind":"section","name":"Drug Discovery Platforms","route":"/fronts/drug-discovery/"}],"target":[{"id":"kras","kind":"target","name":"KRAS","route":"/targets/kras/"}],"company":[{"id":"microsoft-research","kind":"company","name":"Microsoft (Research and Health AI)","route":"/companies/microsoft-research/"}],"paper":[{"id":"paper-lewis-science","kind":"paper","name":"Scalable emulation of protein equilibrium ensembles with generative deep learning","route":"/key-papers/paper-lewis-science/"}]}}