This course teaches learners how to use an ab initio algorithm for homogeneous reconstruction in cryo-electron microscopy (cryo-EM) to estimate 3D poses and electron scattering potential of biomolecules from noisy 2D images. The course covers the cryoAI algorithm, which combines a learned encoder for pose prediction with a physics-based decoder for volume aggregation. The teaching method involves direct gradient-based optimization and neural representation. This course is intended for individuals interested in computational biology, structural biology, cryo-EM, and algorithm development in the field of molecular imaging.
Inference Poses for Reconstruction of 3D Molecular Volumes from Cryo-EM Images
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Syllabus
Frederick Poitevin - Inference Poses for Reconstruction of 3D Molecular Volumes from Cryo-EM Images
Taught by
Institute for Pure & Applied Mathematics (IPAM)