SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM

Thorsten Wagner,Sebastian Tacke,Daniel Prumbaum,Toshio Moriya,Tobias Raisch,Philine Hagel,Amir Apelbaum,Markus Stabrin,Oleg Sitsel,Felipe Merino,Claudia Antoni,Dennis Quentin,Pascal Lill,Stefan Raunser,Evelyn Schubert,Christos Gatsogiannis,Daniel Roderer,Tanvir R Shaikh,Birte Siebolds

doi:10.1038/s42003-019-0437-z

Thorsten Wagner, Sebastian Tacke + Show 17 more

Open Access

https://doi.org/10.1038/s42003-019-0437-z

Copy DOI

Abstract

Selecting particles from digital micrographs is an essential step in single-particle electron cryomicroscopy (cryo-EM). As manual selection of complete datasets—typically comprising thousands of particles—is a tedious and time-consuming process, numerous automatic particle pickers have been developed. However, non-ideal datasets pose a challenge to particle picking. Here we present the particle picking software crYOLO which is based on the deep-learning object detection system You Only Look Once (YOLO). After training the network with 200–2500 particles per dataset it automatically recognizes particles with high recall and precision while reaching a speed of up to five micrographs per second. Further, we present a general crYOLO network able to pick from previously unseen datasets, allowing for completely automated on-the-fly cryo-EM data preprocessing during data acquisition. crYOLO is available as a standalone program under http://sphire.mpg.de/ and is distributed as part of the image processing workflow in SPHIRE.

Highlights

Selecting particles from digital micrographs is an essential step in single-particle electron cryomicroscopy
The results demonstrate that crYOLO is able to accurately and precisely select single particles in micrographs of varying quality with a speed of up to five micrographs per second on a single graphics processing unit (GPU). crYOLO leads to a tremendous acceleration of particle selection, simplifying the training as no negative examples have to be labeled, and improving the quality of the extracted particle images and the final structure
We present a general model for crYOLO trained on more than 40 datasets and able to select particles of previously unseen macromolecular species, realizing automatic particle picking at expert level

Summary

Introduction

Selecting particles from digital micrographs is an essential step in single-particle electron cryomicroscopy (cryo-EM). The last resort is manual selection of particles, which is laborious and time intensive To solve this problem, two particle selection programs, DeepEM by Wang et al.[8] and DeepPicker by Zhu et al.[9], have been recently published, which employ deep convolutional neural networks (CNNs). Similar to common modern object detection systems, particle selection tools employ a specific classifier for an object and evaluate it at every position They are trained with positive examples of cropped out particles and negative examples of cropped out regions of background or contamination. The confidence of classification is transferred into a map and the object positions are estimated by finding the local maxima in this map Using this approach it is possible to select particles on more challenging datasets. As the classifier only sees the windowed region it is not able to learn the larger context of a particle (e.g., to not pick regions near ice contamination)

Methods

Results

Conclusion