dbli2000/filamenttools

Tools and scripts for processing cryoEM data of filaments

FilamentTools

Tools for processing cryoEM data of filaments.

• FiTSuite_typecluster allows for clustering and selection of particle groups/filament types
  • Can run on RELION Class2D or Select jobs post 2D classification
  • Example processing, data, and output provided in FiTSuite, ExampleData, and ExampleOutput folders respectively
  • Run time < 1 minute on a 2017 MacBook pro, even faster on a Linux Cluster
• FiTSuite_summarygenerator produces dataset summaries once types are separated (and optionally map slices after refinement/post-processing) [up soon, end of 8/23]
• FiTSuite_typeswitching, FiTSuite_deduplicate, FiTSuite_removeoverlaps also up soon, end of 8/23

Note that this is not yet compatible with the newly released Jupyter Notebook 7, so either stick with 6 or try NbClassic (not yet tested). Have tested on Mac, Macbook, and Unix systems.

Should be ready to go once you clone/download the repo. All scripts are self-contained in Jupyter Notebooks, but in order to avoid dependency errors it is probably a good idea to use a virtual environment set up either with venv or conda. An example of one way to set up such a virtual environment is below:

Virtual Environment Setup

General instructions

The following is a simple way to set up an virtual environment with venv if you don't know how to:

• Run the following commands in a new terminal
  • python3 -m venv FiTSuite
  • source FiTSuite/bin/activate.csh or source FiTSuite/bin/activate
  • python3 -m pip install -U -r requirements.txt (with the correct path to the requirements.txt in this repo, could also be pip3)
  • python3 -m ipykernel install --user --name=FiTSuite
• Run the command jupyter notebook or python3 -m notebook or whatever you normally do and open the relevant notebook (this is probably /FiTSuite/FiTSuite_typecluster_v0_072723-Template.ipynb)
  • If you choose to just install jupyterlab, FiTSuite is currrently incompatible with JupyterLab 4 due to plotly issues... use JupyterLab 3 e.g. pip install jupyterlab==3
• Upon opening, choose FiTSuite as Kernel by going to Kernel > Change kernel > FiTSuite

When done running the Jupyter notebook and shutting down the server, you can deactivate the virtual environment with deactivate in terminal

After you have done this once, next time just run jupyter notebook, and change kernel to FiTSuite again

LMB instructions

The easiest way is to source from my home directory -> source /lmb/home/dli/FiTSuite/bin/activate.csh. Then run jupyter notebook and open the file

The following is a simple way to set up an virtual environment with venv if you don't know how to:

• Run the following commands in a new terminal from your LMB home directory:
  • module avail python
    • Should see something like "python3/3.10.7" or higher
  • module load python3/3.10.7
    • or whatever the latest version of python is
  • python3 -m venv FiTSuite
  • source FiTSuite/bin/activate.csh or source FiTSuite/bin/activate
  • python3 -m pip install -U -r requirements.txt (with the correct path to the requirements.txt in this repo, could also be pip3)
  • python3 -m pip install jupyter notebook (could also be pip3)
    • If you choose to just install jupyterlab, FiTSuite is currrently incompatible with JupyterLab 4 due to plotly issues... use JupyterLab 3 e.g. pip install jupyterlab==3
  • python3 -m ipykernel install --user --name=FiTSuite
• Run the command jupyter notebook or python3 -m notebook or whatever you normally do and open the notebook
• Upon reopening, choose FiTSuite as Kernel by going to Kernel > Change kernel > FiTSuite

When done running the Jupyter notebook and shutting down the server, you can deactivate the virtual environment with deactivate

After you have done this once, next time just reactivate the virtual environment, run jupyter notebook to get back here, and make sure kernel is still set to FiTSuite