Electron Microscopy Interest Group Meeting

Speakers:  Joshua Carmichael, Garrett Wright, Christopher Langfield, Princeton University

Topic: ASPIRE – A Python Package for Single Particle Reconstruction

We demonstrate ASPIRE’s functionality by running through a typical reconstruction pipeline. After showcasing the basics of loading and unloading real datasets, and the command-line options, we generate a synthetic dataset of projection images using ASPIRE’s Simulation module, into which we feed a molecular volume map and apply custom noise and CTF filters. This results in a set of noisy, CTF corrupted, projection images taken from random viewing angles. Using these images as our experimental base, we apply CTF correction, and class averaging to prepare the images for orientation estimation. We then apply a common-lines algorithm to estimate the relative orientations of our images. Finally, using ASPIRE’s unique 3-D Fourier basis representation we perform a mean volume reconstruction of the molecule., 2-D classification