Spatially Aware Diffraction Mapping Enables Fully Autonomous MicroED

Eremin, Dmitry B.; Jha, Kunal K.; Delgadillo, David A.; Zhang, Hongyu; Foxman, Samuel H.; Johnson, Samuel N.; Vlahakis, Niko W.; Cascio, Duilio; Lavallo, Vincent; Rodríguez, Jose A.; Nelson, Hosea M.

doi:10.1021/jacs.5c10751

Published November 19, 2025 | Version Published

Journal Article Open

Spatially Aware Diffraction Mapping Enables Fully Autonomous MicroED

1. California Institute of Technology
2. University of California, Los Angeles
3. University of California, Riverside

In the hands of experts, microcrystal electron diffraction (microED, a 3D ED method) is a powerful tool for structural chemistry and chemical discovery. To expand the accessibility and utility of microED, we introduce Reciprocal Eyes (REyes), an autonomous and intelligent platform (available for academic use) that combines diffraction-based particle selection with real-time data processing to deliver crystal structures without human intervention. REyes spatially maps diffraction signal and autonomously selects crystallites of interest, relying on lattice-quality metrics to acquire and index high-resolution data sets from diverse compounds. Tested on four different transmission electron microscopes (TEMs), it consistently yields preliminary ab initio structural solutions from single crystallites of materials, peptides, metal complexes, natural products (NPs), and proteins.

Copyright and License

Acknowledgement

Dedicated to the memory of Prof. George M. Sheldrick, in honor of his lifelong contributions to science. The authors thank Alex Lisker (UCLA) for computational infrastructure management; Cameron Flowers (UCLA) for the samples of lysozyme crystals; Dr. William Palfey and Prof. George Rossman (Caltech) for the samples of grossular; Dr. Adam Stieg, Dr. Matthew Mecklenburg, and Dr. Jungyoun Cho (CNSI/UCLA) for providing access to Spectra 300C microscope; Dr. Scott Virgil (Caltech) and the Center for Catalysis and Chemical Synthesis for the LC-HRMS infrastructure and support; Katya Kadyshevskaya (USC) for the design of REyes icon; and Dr. Jessica Burch and Isabel Hernandez Rodriguez for their pioneering work on computer vision-based microED workflows. The authors thank the Resnick High Performance Computing Center for providing computational resources. The code development was supported by Claude (Anthropic) and ChatGPT (OpenAI) resources. ChatGPT (OpenAI) was used to assist with language editing and proofreading of this manuscript. Figure 1 and TOC were created with BioRender.com. Development of the platform for diffraction data labeling has been supported by the NSF CCI Center for Computer Assisted Synthesis (CHE-1925607). Structure elucidation of natural products has been supported by NIH NCCIH (R01AT011990) and Howard Hughes Medical Institute Emerging Pathogens Initiative (HHMI-EPI). Use of Spectra 300C was supported by the BioPACIFIC Materials Innovation Platform of the NSF (DMR-1933487). V.L. is grateful for partial NSF support of this project (CHE-2003418).

Data Availability

MicroED structure of the lysozyme-ligand complex is available via PDB entry: 9P6R. All other data are provided in the Supporting Information. REyes Python package, associated SerialEM script suite, installation manual, user guides, support tools, and hyperlink to the raw data are available at https://github.com/theNelsonLab/pyREyes. The example raw and processed data set for (S,S)-salen ligand 1 are available at 10.5281/zenodo.16971798.

Supplemental Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c10751.

Experimental details, REyes data, and microED structural data (PDF)

Additional Information

Accession Codes:

Deposition Numbers 2449546–2449550, 2449554–2449556, and 2449594–2449604 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via the joint Cambridge Crystallographic Data Centre (CCDC) and Fachinformationszentrum Karlsruhe Access Structures service.

Files