Explore the fundamentals of cryo-electron microscopy and single-particle analysis (SPA) in this comprehensive webinar presented by Prof. Michael Unser from EPFL. Delve into the high-resolution imaging of biosamples in their native state, understanding the X-ray transform model for electron ray transmission, and the challenges of tomographic reconstruction in SPA. Learn about current reconstruction paradigms, regularized reconstruction schemes, and the innovative CryoGAN algorithm. Examine synthetic experiments, discriminator architecture, and results in structural reconstruction. Address the key challenge of structural heterogeneity, exploring methods for reconstructing both continuous and discrete sets of conformations. Gain valuable insights into advanced signal processing techniques applied to cutting-edge biological imaging.
Cryo-Electron Microscopy and Single Particle Analysis - Webinar 3
IEEE Signal Processing Society via YouTube
Overview
Syllabus
Intro
EPFL Outline
EPFL Cryo-electron microscopy: imaging biosamples at high-resolution in their native state
EPFL The fundamental principles of SPA
EPFL The transmission of straight electron rays is modeled with the X-ray transform
EPFL The 3D object is represented in a shift-invariant ba
EPFL Tomographic reconstruction in SPA is a major technical challenge
EPFL Current reconstruction paradigms achieve high-res through iterative-refinement procedures
EPFL Regularized reconstruction scheme for SPA
EPFL The joint scheme appropriately refines the structure and the orientations
EPFL The scheme compares favorably with
EPFL Classical GANs estimate data distributions through an adversarial learning scheme
EPFL CryoGAN relies on a mathematical framework that provides strong recovery guarantees
EPFL A closer look on the min-max problem
EPFL The CryoGAN algorithm
EPFL Synthetic Experiment (beta-gal)
EPFL Discriminator architecture
EPFL Results (Structure)
SPRL Reconstruction (Structure)
EPFL Structural heterogeneity: the key challenge ahead
EPFL Reconstruction of a continuum of conformations
EPFL Reconstruction for a discrete set of conformations
EPFL Conclusion
Taught by
IEEE Signal Processing Society
