Inference Poses for Reconstruction of 3D Molecular Volumes from Cryo-EM Images

Explore a 46-minute lecture on advanced techniques in cryo-electron microscopy (cryo-EM) reconstruction presented by Frederick Poitevin from SLAC National Accelerator Laboratory. Delve into the challenges of processing large cryo-EM datasets and learn about cryoAI, an innovative ab initio algorithm for homogeneous reconstruction. Discover how this approach combines a learned encoder for particle image pose prediction with a physics-based decoder to create neural representations of scattering potential volumes. Gain insights into the potential of this inference method for handling arbitrarily large datasets and its implications for structural biology research.