Guo, A. C. | Search bioRxiv

bioRxiv2025

SMART: An Integrated Platform Revolutionizing Cryo-Electron Microscopy Workflow from Data Acquisition to Atomic Modeling

Recent advancements in cryo-electron microscopy (cryo-EM) have revolutionized structural biology, yet persistent workflow fragmentation between data acquisition, reconstruction, and modeling limits its full potential. This paper presents SMART, the first fully integrated cryo-EM platform unifying three critical AI-driven modules: DataSmart (automated data collection), CryoSmart (3D reconstruction), and ModelSmart (end-to-end atomic modeling). The platform embeds more than 10 distinct AI algorithms across its workflow, achieving unprecedented automation levels. These advancements not only lower the expertise barrier for cryo-EM adoption but also establish a new paradigm for AI-empowered structural biology, with broad implications for high-throughput drug discovery. This protocol provides a comprehensive delineation of the operational procedures for the SMART platform. As an example, this protocol enabled the determination of the 2.1 [A] resolution structure of TRPML1, a Ca2+-permeable, nonselective, six-transmembrane tetrameric cation channel found in late endosomes and lysosomes (LELs) of mammalian cells. This integrated workflow enables users, including those with minimal cryo-EM experience, to efficiently complete the entire pipeline from data collection to atomic model building.

bioinformatics↗

bioRxiv2023

An end-to-end approach for protein folding by integrating Cryo-EM maps and sequence evolution

Protein structure modeling is an important but challenging task. Recent breakthroughs in Cryo-EM technology have led to rapid accumulation of Cryo-EM density maps, which facilitate scientists to determine protein structures but it remains time-consuming. Fortunately, artificial intelligence has great potential in automating this process. In this study, we present SMARTFold, a deep learning protein structure prediction model combining sequence alignment features and Cryo-EM density map features. First, using density map, we sample representative points along the predicted high confidence areas of protein backbone. Then we extract geometric features of these points and integrate these features with sequence alignment features in our proposed protein folding model. Extensive experiments confirm that our model performs best on both single-chain and multi-chain benchmark dataset compared with state-of-the-art methods, which makes it a reliable tool for protein atomic structure determination from Cryo-EM maps.

bioinformatics↗