https://3demmethods.i2pc.es/index.php?action=history&feed=atom&title=2022Levy_CryoFire 2026-05-24T20:20:57Z Revision history for this page on the wiki MediaWiki 1.44.2 https://3demmethods.i2pc.es/index.php?title=2022Levy_CryoFire&diff=4337&oldid=prev 2023-06-16T09:21:55Z

<p>Created page with &quot;== Citation == Levy, Axel / Wetzstein, Gordon / Martel, Julien N. P. / Poitevin, Frederic / Zhong, Ellen Amortized Inference for Heterogeneous Reconstruction in Cryo-EM 202...&quot;</p> <p><b>New page</b></p><div>== Citation ==<br /> <br /> Levy, Axel / Wetzstein, Gordon / Martel, Julien N. P. / Poitevin, Frederic / Zhong, Ellen <br /> Amortized Inference for Heterogeneous Reconstruction in Cryo-EM <br /> 2022. Proc. Advances in Neural Information Processing Systems, Vol. 35, p. 13038-13049 <br /> <br /> == Abstract ==<br /> <br /> Cryo-electron microscopy (cryo-EM) is an imaging modality that provides unique<br /> insights into the dynamics of proteins and other building blocks of life. The algorithmic<br /> challenge of jointly estimating the poses, 3D structure, and conformational<br /> heterogeneity of a biomolecule from millions of noisy and randomly oriented 2D<br /> projections in a computationally efficient manner, however, remains unsolved. Our<br /> method, cryoFIRE, performs ab initio heterogeneous reconstruction with unknown<br /> poses in an amortized framework, thereby avoiding the computationally expensive<br /> step of pose search while enabling the analysis of conformational heterogeneity.<br /> Poses and conformation are jointly estimated by an encoder while a physics-based<br /> decoder aggregates the images into an implicit neural representation of the conformational<br /> space. We show that our method can provide one order of magnitude<br /> speedup on datasets containing millions of images without any loss of accuracy.<br /> We validate that the joint estimation of poses and conformations can be amortized<br /> over the size of the dataset. For the first time, we prove that an amortized method<br /> can extract interpretable dynamic information from experimental datasets.<br /> <br /> == Keywords ==<br /> <br /> == Links ==<br /> <br /> https://proceedings.neurips.cc/paper_files/paper/2022/file/54b8b4e0b4ba4aad112e84f32e3b5dbb-Paper-Conference.pdf<br /> <br /> == Related software ==<br /> <br /> == Related methods ==<br /> <br /> == Comments ==</div>

WikiSysop