Dynamical Diffraction Analysis for 3D Electron Crystallography

Project Abstract This Phase II SBIR project aims to advance microcrystal electron diffraction (microED) methods, which combine the advantages of single-crystal X-ray crystallography with the flexibility of working with smaller crystals, but they do not routinely produce results of quality comparable to X-ray crystallography. Challenges include inefficiencies in data acquisition and analysis, the absence of dedicated methods for solving the phase problem in the presence of electron diffraction (ED)-specific systematic effects, and a lack of integration between software tools, resulting in information loss and the selection of suboptimal strategies by researchers to circumvent software issues. The overarching goal of the project is to enhance the usability of ED by developing innovative and integrated software solutions that address these challenges. We will improve data collection efficiency by innovating new data acquisition techniques, enabling faster selection of crystals and their quality evaluation with higher confidence. This will be achieved by calibrating hardware features affecting the experiment and by developing new indicators for detecting crystals on cryoEM grids. We will also advance ED data analysis by developing methods for indexing, integrating, and scaling electron diffraction patterns, including handling multiple scattering effects, and accelerating data throughput with GPU computing for real-time analysis. Furthermore, we will develop ED-specific phasing methods to address data insufficiency problems common in the field and to facilitate optimal structure refinement for better agreement between data and chemical models. This will involve dynamical refinement, determination of absolute configuration, dimensionality reduction methods for clustering isomorphous samples, and developing corrections for many systematic effects that are currently not modeled or are modeled suboptimally. Finally, we will equip users with tools for monitoring the progress and performance of complex experiments, instrument data analysis, and automating data preparation for deposition in scientific repositories. The outcome of the SBIR Phase II will be an easy-to-use, GPU-optimized, complete crystallographic package for microED, validated across a broad range of experimental conditions with submodules tailored to address specific scientific problems, e.g., analysis of small organic molecules or analysis of large macromolecules. This user-friendly, efficient, versatile software will be ready for commercial distribution and will benefit researchers who require atomic-level descriptions of chemical compounds, both in industry and academia.