TranSPHIRE
TranSPHIRE automates real-time processing of cryo-electron microscopy (cryo-EM) datasets during data acquisition to perform on-the-fly particle picking, 2D clustering, and 3D refinement for high-throughput structure determination.
Key Features:
- Full Automation: Automates the cryo-EM processing pipeline from data transfer through 3D refinement, integrating with ongoing data acquisition.
- On-the-Fly Processing: Processes images in real time as they are recorded to provide immediate analysis results.
- Machine Learning-Based Feedback Loop: Dynamically re-trains the particle picking model during live processing to adapt to specific datasets.
- Comprehensive Pre-Processing Steps: Automatically performs particle picking, 2D clustering, and 3D refinement as part of preprocessing.
- Data Management and Visualization: Collects and displays microscope settings and metrics generated by its tools for evaluation during acquisition.
- Versatility in Data Types: Supports automated processing of filamentous structures in addition to standard particle processing.
Scientific Applications:
- High-Throughput Structure Determination: Enables rapid, high-throughput determination of macromolecular structures by automating processing during acquisition.
- Drug Discovery: Supports structural workflows relevant to drug discovery by improving throughput and enabling on-the-fly optimization of data quality.
- Real-Time Data Quality Assessment: Provides immediate feedback during acquisition to optimize imaging conditions and data quality.
- Filamentous Structure Analysis: Enables automated analysis and processing of filamentous samples for structural studies.
Methodology:
Automated computational steps explicitly include data transfer integration, on-the-fly image processing as images are recorded, particle picking, 2D clustering, 3D refinement, collection and display of microscope settings and metrics, and a machine learning-based feedback loop that dynamically re-trains the particle picking model during live processing.
Topics
Details
- License:
- GPL-3.0
- Programming Languages:
- Python
- Added:
- 1/18/2021
- Last Updated:
- 3/3/2021
Operations
Publications
Stabrin M, Schoenfeld F, Wagner T, Pospich S, Gatsogiannis C, Raunser S. TranSPHIRE: Automated and feedback-optimized on-the-fly processing for cryo-EM. Unknown Journal. 2020. doi:10.1101/2020.06.16.155275.