HOTSPUR: A Real-time Interactive Preprocessing System for Cryo-EM Data

Single-particle cryo-electron microscopy enables structure determination of biological molecules and complexes at near atomic resolution with unprecedented speed. Yet in many instances, acquisition of a sufficient number of particles requires collection of thousands of micrographs, collected using direct electron detectors and comprised of dose-fractionated "movie" frames. To achieve the highest resolution reconstructions, the user must search the grid for optimal regions, typically ones with thin ice and dense yet non overlapping particle distribution. While software packages such as SerialEM [1] or EPU facilitate automated image acquisition, user intervention is required to choose the best acquisition areas. However, images require preprocessing-motion correction and contrast-transfer function (CTF) estimation -before they can be visually inspected. In addition to enabling visual inspection of the images, preprocessing yields quantitative information on image quality. To allow one to judge image quality in near real time we developed a software package, called HOTSPUR, that performs preprocessing without user intervention and that presents the results in an easy-to-use interface. HOTSPUR allows researchers to rapidly judge data quality and to focus on decision-making to obtain the highest quality data from limited microscope time.