AutoSolveX: automated batch structure solution, initial refinement and guest screening from diffraction data
- 1Department of Chemistry, Stockholm University, SE-106 91, Sweden
Abstract
Modern three-dimensional electron diffraction (3D ED) workflows can collect and process many datasets in a single session, whereas downstream structure solution, refinement and model interpretation often remain manual. Here we present AutoSolveX, a cross-platform Python pipeline with a graphical user interface for automated batch structure solution and initial refinement from processed diffraction data. AutoSolveX integrates SHELXT, SHELXS and SHELXL, applies electron scattering factors to 3D ED data, and extends automation to guest identification by screening candidate molecules against experimental Q-peak clusters and generating guest-containing models for subsequent refinement. We demonstrate AutoSolveX using five deposited electron-diffraction structures from different chemical classes, seven 3D ED datasets of Co-MOF-74, and ten 3D ED datasets of guest-containing structures of Bi-MOF SU-100 and 18 single-crystal X-ray diffraction datasets of Al-MOF-520. For the five deposited structures, AutoSolveX recovered the reported space groups and produced initial models with coordinate root-mean-square deviations of 0.015–0.052 Å from the deposited models. Automated batch solution and refinement of the Co-MOF-74 datasets generated comparable structural models across crystals and identified a dataset with poorer refinement results, demonstrating the value of systematic multi-dataset analysis. Guest screening retained the expected guest among the top three candidates for 9 of 10 SU-100 datasets and among the top five for 12 of 18 MOF-520 datasets. Complete molecular geometries were generated for all but the largest guests. AutoSolveX therefore provides a unified workflow from diffraction data to structural models and guest interpretation, reducing repetitive crystallographic work and facilitating analysis of large numbers of datasets.
Related
- Automated multiphase identification and refinement in powder diffraction using mismatch-tolerant machine learning
- Rongzai agent: A Large Language Model-Based Autonomous Assistant for Rietveld Refinement of Neutron Diffraction Data
- High-Throughput Structure Determination Using Serial Electron Diffraction
- A Roadmap for Supercrystallography: From Single-Supercrystal Diffraction to the Atomic Structure of Colloidal Nanocrystals
- Continuous perception and adaptive metrics for scalable structure refinement
BibTeX
@misc{butonova2026autosolvex,
title = {{AutoSolveX: automated batch structure solution, initial refinement and guest screening from diffraction data}},
author = {Sofiia Butonova and Yinlin Chen and Xiaodong Zou},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.10002125/v2},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.10002125/v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - AutoSolveX: automated batch structure solution, initial refinement and guest screening from diffraction data
AU - Sofiia Butonova
AU - Yinlin Chen
AU - Xiaodong Zou
PY - 2026
DA - 2026/09/30
PB - ChemRxiv
DO - 10.26434/chemrxiv.10002125/v2
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.10002125/v2
AB - Modern three-dimensional electron diffraction (3D ED) workflows can collect and process many datasets in a single session, whereas downstream structure solution, refinement and model interpretation often remain manual. Here we present AutoSolveX, a cross-platform Python pipeline with a graphical user interface for automated batch structure solution and initial refinement from processed diffraction data. AutoSolveX integrates SHELXT, SHELXS and SHELXL, applies electron scattering factors to 3D ED data, and extends automation to guest identification by screening candidate molecules against experimental Q-peak clusters and generating guest-containing models for subsequent refinement. We demonstrate AutoSolveX using five deposited electron-diffraction structures from different chemical classes, seven 3D ED datasets of Co-MOF-74, and ten 3D ED datasets of guest-containing structures of Bi-MOF SU-100 and 18 single-crystal X-ray diffraction datasets of Al-MOF-520. For the five deposited structures, AutoSolveX recovered the reported space groups and produced initial models with coordinate root-mean-square deviations of 0.015–0.052 Å from the deposited models. Automated batch solution and refinement of the Co-MOF-74 datasets generated comparable structural models across crystals and identified a dataset with poorer refinement results, demonstrating the value of systematic multi-dataset analysis. Guest screening retained the expected guest among the top three candidates for 9 of 10 SU-100 datasets and among the top five for 12 of 18 MOF-520 datasets. Complete molecular geometries were generated for all but the largest guests. AutoSolveX therefore provides a unified workflow from diffraction data to structural models and guest interpretation, reducing repetitive crystallographic work and facilitating analysis of large numbers of datasets.
ER -