Generative adversarial networks as a tool to recover structural information from cryo-electron microscopy data

Cryo-electron microscopy (cryo-EM) is a powerful structural biology technique capable of determining atomic-resolution structures of biological macromolecules. However, the process of optimizing sample preparation and data collection requires expert-level users many months to years in order to find amenable conditions to collect and determine high-resolution structures due to the low signal-to-noise (SNR) ratio of the raw cryo-EM micrographs. To help address this problem, we have trained and tested generative adversarial networks, a form of artificial intelligence, to denoise and CTF-correct individual particles. This approach effectively recovers global structural information for both synthetic and real cryo-EM data, facilitating per-particle assessment from noisy raw images. Our results suggest that generative adversarial networks may be able to provide an approach to denoise raw cryo-EM images to facilitate particle selection and raw particle interpretation for single particle and tomography cryo-EM data.