TranSPHIRE: automated and feedback-optimized on-the-fly processing for cryo-EM

Markus Stabrin,Christos Gatsogiannis,Thorsten Wagner,Fabian Schoenfeld,Sabrina Pospich,Stefan Raunser

doi:10.1038/s41467-020-19513-2

Markus Stabrin, Christos Gatsogiannis + Show 4 more

Open Access

https://doi.org/10.1038/s41467-020-19513-2

Copy DOI

Abstract

Single particle cryo-EM requires full automation to allow high-throughput structure determination. Although software packages exist where parts of the cryo-EM pipeline are automated, a complete solution that offers reliable on-the-fly processing, resulting in high-resolution structures, does not exist. Here we present TranSPHIRE: A software package for fully-automated processing of cryo-EM datasets during data acquisition. TranSPHIRE transfers data from the microscope, automatically applies the common pre-processing steps, picks particles, performs 2D clustering, and 3D refinement parallel to image recording. Importantly, TranSPHIRE introduces a machine learning-based feedback loop to re-train its picking model to adapt to any given data set live during processing. This elegant approach enables TranSPHIRE to process data more effectively, producing high-quality particle stacks. TranSPHIRE collects and displays all metrics and microscope settings to allow users to quickly evaluate data during acquisition. TranSPHIRE can run on a single work station and also includes the automated processing of filaments.

Highlights

Single particle cryo-EM requires full automation to allow high-throughput structure determination
TranSPHIRE implements a fully-automated pipeline to process cryo-EM data onthe-fly during data acquisition
TranSPHIRE is designed to allow users to make the best use of their available resources by prioritizing data analysis, presenting early results, and using machine learning tools to identify and process only those parts of the data that contribute to high-quality results

Summary

Introduction

Single particle cryo-EM requires full automation to allow high-throughput structure determination. With the number of modern high-throughput microscopes growing rapidly, there is an urgent demand for a robust, automated processing pipeline that requires little to no user intervention This need is felt especially in the field of drug discovery[3]. CryoFLARE8, Focus[9], Scipion[10,11], Appion[11], WARP12, RELION IT13, and CryoSPARC Live perform live analysis and processing in parallel to data acquisition This is done by chaining together the different tools of the single particle analysis pipeline, but it lacks any automatic data optimization. It combines deep learning tools for particle picking and 2D class selection with a novel, feedback-driven approach to re-train the integrated crYOLO particle picker[17] during ongoing pre-processing This allows TranSPHIRE to perform GPU accelerated 2D classification to provide high-quality 2D class averages and, subsequently, 3D reconstructions from clean data. It allows users to perform automated high-throughput on-the-fly screenings for different buffer conditions or ligands of interest as well as to fine-tune the workflow for the respective target-protein and perform digital purification during image acquisition

Methods

Results

Conclusion