Abstract:In recent years, cryo-electron microscopy (cryo-EM) has become a practical and effective method of determining structures at previously unattainable resolutions due to advances in detection, automation, and data processing. However, sample preparation remains a major bottleneck in the cryo-EM workflow. Even after the arduous process of biochemical sample optimization, it often takes several iterations of grid vitrification and screening to determine the optimal grid freezing parameters that yield suitable ice thickness and particle distribution for data collection. Since a high-quality sample is imperative for high-resolution structure determination, grid optimization is a vital step. For researchers who rely on cryo-EM facilities for grid screening, each iteration of this optimization process may delay research progress by a matter of months. Therefore, a more strategic and efficient approach should be taken to ensure that the grid optimization process can be completed in as few iterations as possible. Here, we present an implementation of Design of Experiments (DOE) to expedite and strategize the grid optimization process. A Fractional Factorial Design (FFD) guides the determination of a limited set of experimental conditions which can model the full parameter space of interest. Grids are frozen with these conditions and screened for particle distribution and ice thickness. Quantitative scores are assigned to each of these grid characteristics based on a qualitative rubric. Input conditions and response scores are used to generate a least-squares regression model of the parameter space in JMP, which is used to determine the conditions which should, in theory, yield optimal grids. Upon testing this approach on apoferritin and L-glutamate dehydrogenase on both the Vitrobot Mark IV and the Leica GP2 plunge freezers, the resulting grid conditions reliably yielded grids with high-quality ice and particle distribution that were suitable for collecting large overnight datasets on a Krios. We conclude that a DOE-based approach is a cost-effective and time-saving tool for cryo-EM grid preparation.

CryoRL: Reinforcement Learning Enables Efficient Cryo-EM Data Collection

Robust ab initio solution of the cryo-EM reconstruction problem at low resolution with small data sets

A comprehensive foundation model for cryo-EM image processing

Deep-learning with synthetic data enables automated picking of cryo-EM particle images of biological macromolecules.

CryoFormer: Continuous Heterogeneous Cryo-EM Reconstruction using Transformer-based Neural Representations

Attention-guided Quality Assessment for Automated Cryo-EM Grid Screening

Cryo-ZSSR: multiple-image super-resolution based on deep internal learning

Accounting Conformational Dynamics into Structural Modeling Reflected by Cryo-EM with Deep Learning

Machine Learning for Structure Determination in Single-Particle Cryo-Electron Microscopy: A Systematic Review

Protein complex structure modeling by cross-modal alignment between cryo-EM maps and protein sequences

A strategic approach for efficient cryo-EM grid optimization using design of experiments

CryoGEM: Physics-Informed Generative Cryo-Electron Microscopy

A deep convolutional neural network approach to single-particle recognition in cryo-electron microscopy

Multiple-image super-resolution of cryo-electron micrographs based on deep internal learning

CryoTRANS: predicting high-resolution maps of rare conformations from self-supervised trajectories in cryo-EM

Cryo2RT: a high-throughput method for room-temperature macromolecular crystallography from cryo-cooled crystals

CryoMAE: Few-Shot Cryo-EM Particle Picking with Masked Autoencoders

CryoTransformer: a transformer model for picking protein particles from Cryo-EM micrographs

Efficient high-resolution refinement in cryo-EM with stochastic gradient descent

Improvement of cryo-EM maps by simultaneous local and non-local deep learning

Learning to automate cryo-electron microscopy data collection with Ptolemy