Cryo-electron microscopy (cryo-EM) of non-crystalline single particles is a biophysical technique that can be used to determine the structure of biological macromolecules and assemblies. Historically, its potential for application in drug discovery has been heavily limited by two issues: the minimum size of the structures it can be used to study and the resolution of the images. However, recent technological advances - including the development of direct electron detectors and more effective computational image analysis techniques - are revolutionizing the utility of cryo-EM, leading to a burst of high-resolution structures of large macromolecular assemblies. These advances have raised hopes that single-particle cryo-EM might soon become an important tool for drug discovery, particularly if they could enable structural determination for 'intractable' targets that are still not accessible to X-ray crystallographic analysis. This article describes the recent advances in the field and critically assesses their relevance for drug discovery as well as discussing at what stages of the drug discovery pipeline cryo-EM can be useful today and what to expect in the near future.

To automate electron microscopy tasks, the main challenge is to integrate image interpretation. This article presents a first achievement in the domain of electron microscopy. ANalysis of IMages for Automatic Targeting and Extraction of Data in Transmission Electron Microscopy (ANIMATED-TEM) is a software toolbox composed of a set of image analysis algorithms and an examination scenario. Combined with microscope software, it runs all the microscope tasks and manages the exploration strategy of the carbon-coated grids on which the sample of interest (viruses, proteins, bi-dimensional crystal of membrane proteins, etc.) has been deposited. ANIMATED-TEM realizes the automatic examination of biological samples, working at several magnifications without human intervention. Online image analysis of micrographs is an essential part both to extract characteristic data and to manage automation, as interesting regions need to be identified to trigger new acquisitions at higher magnifications. This toolbox has been developed to perform high-throughput screening of 2D-crystallization experiments. It is operational on a microscope equipped with an automatic grid loading system, allowing the continuous and automatic analysis of up to 96 samples. Intensive testing over a period of several months confirms that ANIMATED-TEM achieves full automation with an efficient target selection and in a suitable computational time.

No abstract