Installation of ResiDEM

Installation of ResiDEM requires few dependencies such as Computational Crystallography Toolbox (CCTBX). It also some requires some basic python packages for data processing such as numpy, pandas, scipy etc…

Installation

ResiDEM is a python tool written in python 3 version, and it is recommended that a new environment is created for better package management and to avoid any dependency conflicts.

  1. Creating a conda virtual environment and installation :

# you can change the name if you want
conda create -n residem python=3.10
conda activate residem

  1. Updating the environment

# After activation of residem environment, we can check the location with command `which python`
# do the following to install in particular residem environment

full_path=$(which python)
base_path=$(dirname $(dirname "$full_path"))
echo $base_path

conda env update --prefix $base_path --file environment.yml --prune

  1. Install ResiDEM package

pip install dist/residem-0.1.0.tar.gz

Additional Information

The ResiDEM package might also need CCP4 scaleit. It will be used for scaling between reference and triggered state if that particular user option in given. It also has inbuilt scaling method, which works well with isomorphous data sets.

To check if the CCP4 package or scaleit in the path, the following command can used.

if command -v ccp4 &> /dev/null
then
    echo "CCP4 in command line is Present"
else
    echo "CCP4 in command line is Not Present please install and give the path"
fi

Unit test

Testing the tool with Bacteriorhodopsin data. There are two python scripts which can be used to run, reproduce images as in the published article . The name of the scripts are residem_unit_test.py and SVD_unit_test.py which are available in test_data directory in tests as zip file. The main script residem_unit_test.py can be used to test the tool and reproduce certain figures as in published article . This script computes the isomorphous difference density for Bacteriorhodopsin(bR) for 13 time delays as published by Nango et al.

Single computation approximately takes around 3-5 minutes in personal laptop. The script residem_unit_test.py may take some time (~1 hour) to compute for 13 datasets.

The testing can be done using the following commands.

conda activate residem
cd tests
unzip test_data.zip
cd test_data
python residem_unit_test.py
# after completeion of the above the follwoing can be submitted.
python SVD_unit_test.py

After running the script, image corresponding to Figure 3, Figure 4 and Figure 6(a) as in published article can be reproduced. Ater successful completion, the image can be found in the following places.

  • Figure 3(a) can be seen in the file SVD/760ns/one_dimensional_linear_plot.pdf .

  • Figure 3(b) can be seen in the file SVD/negative/one_dimensional_linear_plot.pdf.

  • Figure 3(c) can be seen in the file SVD/negative/SVD_original.pdf.

  • Figure 4 can be seen in the file SVD/760ns/Difference_density_network_CA_nolabel.pdf. In the generated image, the edge and nodes will be same, but the layout might be different as random layout is generated every time.

  • Figure 6(a) can be seen in the file SVD/all/correlation_plot.pdf.