Skip to content

crYOLO

crYOLO is an application for fast and accurate cryo-EM particle picking.

crYOLO is available as a module on Apocrita.

Usage

To run the default installed version of crYOLO, simply load the cryolo module:

module load cryolo

Since crYOLO runs from a container, you need to run cryolo and then the command you wish to run, e.g.:

cryolo cryolo_train.py -h

crYOLO Graphical User Interface

The main crYOLO Graphical User Interface (cryolo_gui.py) is available when the cryolo module is loaded. However we strongly suggest users use a job script for training and analysis, then use an Open OnDemand Desktop Environment session to visualise results using napari_boxmanger.

Example job

crYOLO expects a GPU for training and analysis

crYOLO training and analysis should be run on a GPU node, as it relies on CUDA libraries for these workflows.

GPU job

Here is an example job running on 1 GPU, using the crYOLO reference example:

#!/bin/bash
#$ -cwd
#$ -j y
#$ -pe smp 8
#$ -l h_rt=240:0:0
#$ -l h_vmem=11G
#$ -l gpu=1

module load cryolo

# Run training
cryolo cryolo_train.py \
    -c config_phosnet.json \
    -w 5 \
    -e 5 \
    -g 0

# Run prediction
cryolo cryolo_predict.py \
    -c config_phosnet.json \
    -w model.h5 \
    -i unseen_examples/ \
    -o my_results

Visualising results

The example job produces the following output in the defined output directory "my_results":

$ ls -1 my_results
CBOX
CRYOSPARC
DISTR
EMAN
STAR

You can visualise these results in an Open OnDemand Desktop Environment session.

Use a CPU Desktop Environment session where possible

The crYOLO Napari boxmanager can run on CPU or GPU nodes. To avoid queueing, use a CPU Desktop Environment session for smaller visualisations and only use a GPU node for larger visualisations.

Once in the Desktop Environment session, load the cryolo module and the Napari boxmanager (napari_boxmanager) will be added to your path. To use it run a command such as:

napari_boxmanager 'unseen_examples/*.mrc' 'my_results/CBOX/*.cbox'

in the terminal window that appears. The defined results path will load in the Napari graphical user interface:

crYOLO Napari boxmanager visualisation

References