This plugin provides a wrapper for TomoTwin software: Particle picking in Tomograms using triplet networks and metric learning
You will need to use 3.0+ version of Scipion to be able to run these protocols. To install the plugin, you have two options:
- Stable version
scipion installp -p scipion-em-tomotwin
Developer's version
- download repository
git clone -b devel https://github.com/scipion-em/scipion-em-tomotwin.git
- install
scipion installp -p /path/to/scipion-em-tomotwin --devel
TomoTwin software will be installed automatically with the plugin but you can also use an existing installation by providing TOMOTWIN_ENV_ACTIVATION (see below).
Important: you need to have conda (miniconda3 or anaconda3) pre-installed to use this program.
CONDA_ACTIVATION_CMD: If undefined, it will rely on conda command being in the PATH (not recommended), which can lead to execution problems mixing scipion python with conda ones. One example of this could can be seen below but depending on your conda version and shell you will need something different: CONDA_ACTIVATION_CMD = eval "$(/extra/miniconda3/bin/conda shell.bash hook)"
TOMOTWIN_ENV_ACTIVATION (default = conda activate tomotwin-0.9.1): Command to activate the TomoTwin environment. Tomotwin uses cuda-11.8, so you might want to activate specific CUDA libs via e.g. TOMOTWIN_ENV_ACTIVATION = . /etc/profile.d/lmod.sh && module load cuda/11.8 && conda activate tomotwin-0.9.1
TOMOTWIN_MODEL (default = software/em/tomotwin_model-092023/tomotwin_model_p120_092023_loss.pth): Path to the pre-trained model.
NAPARI_ENV_ACTIVATION (default = conda activate napari-0.4.19): Command to activate the Napari viewer environment.
To check the installation, simply run the following Scipion tests:
scipion tests tomotwin.tests.test_protocols_tomotwin.TestTomoTwinRefBased
scipion tests tomotwin.tests.test_protocols_tomotwin.TestTomoTwinClusterBased
0.8.0, 0.9.0b1, 0.9.1
- clustering-based picking (step 1)
- clustering-based picking (step 2)
- create tomo masks
- reference-based picking
- TomoTwin: Generalized 3D Localization of Macromolecules in Cryo-electron Tomograms with Structural Data Mining. Gavin Rice, Thorsten Wagner, Markus Stabrin, Stefan Raunser. https://www.biorxiv.org/content/10.1101/2022.06.24.497279v1