Fig. 1: The 3DFlex model represents the flexible 3D structure of a protein as deformations of a single canonical 3D density map V.
From: 3DFlex: determining structure and motion of flexible proteins from cryo-EM
Under the model, a single-particle image is associated with low-dimensional latent coordinates that encode the conformation for the particle in the image. A neural flow generator network fθ converts the latent coordinates into the flow field u and a convection operator then deforms the canonical density to generate a convected map W. This map can then be projected along the particle viewing direction determined by the pose ϕ, CTF corrupted and compared against the experimental image.
