/pafnucy-convert

Primary LanguagePython

Pafnucy Data Conversion Scripts

Setting up your environment

The following instructions recount how I've gone about setting up my environment on the rice machines.

Installing Python requirements

I created a Python3 virtual environment, though you could probably skip this and move right to instaling requirements.

virtualenv -p python3 venv

Activate the virtual environment by running:

source venv/bin/activate

Install requirements by running:

pip install -r requirements.txt

You can deactivate the virtual environment with:

deactivate

Installing AutoDock Vina

Navigate to http://vina.scripps.edu/download.html. Copy the link to the Linux download. Run the following commands.

curl -O [link]
tar xzvf autodock_vina_1_1_2_linux_x86.tgz

I recommend adding this line to your .bashrc file, since the docking script relies on this variable being set.

export VINA_ROOT="[wherever you installed AutoDock Vina]"

Example:

export VINA_ROOT="$HOME/autodock_vina_1_1_2_linux_x86"

Installing MGL Tools

Navigate to http://mgltools.scripps.edu/downloads. Copy the link to the proper tarball installer. Run the following commands.

curl -O [link]
tar -xzf [filename]
cd [folder name]
./install.sh

I recommend adding this line to your .bashrc file, since some of these scripts rely on this variable being set.

export MGL_ROOT="[wherever you installed MGL Tools]"

Example:

export MGL_ROOT="$HOME/mgltools_x86_64Linux2_1.5.6"

Installing PyMol

The version of PyMol pre-installed on the rice machines is old, so you must install a new version. Navigate to https://pymol.org/2/#download. Copy the link for the proper install file. Run the following commands.

curl -O [link]
tar -jxf [filename]

Do NOT add pymol/bin to your path. It includes an anaconda installation that may mess with things. Instead, create a symlink to the PyMol binary somewhere in your path. For example, I ran

cd ~
mkdir bin
cd bin
ln -s ~/pymol/bin/pymol pymol

and added this line to my .bashrc file:

export PATH="$HOME/bin:$PATH"

Cloning Pafnucy

Clone the repository:

git clone git@gitlab.com:jspayd/pafnucy.git

Add this line to your .bashrc:

export PAFNUCY_ROOT="[wherever you cloned pafnucy to]"

For example:

export PAFNUCY_ROOT="$HOME/pafnucy"

Converting ligand SMILES strings into PDB files

Use the ligand_smiles_to_pdb.py script for this. It reads from smiles.csv and outputs to the ligand_pdb directory.

If you haven't already, activate your virtual environment:

source venv/bin/activate

Then run the script:

python ligand_smiles_to_pdb.py

Getting relative side chain distances from amino acids

Use the side_chain_dists.py script for this. It reads from the aminoacid_pdbs directory and outputs to the aminoacid_dists directory.

If you haven't already, activate your virtual environment:

source venv/bin/activate

Then run the script:

python side_chain_dists.py

Getting the PDB files of translations from hNaV1.7

Use the create_protein_translation_pdbs.py script for this. It reads from the aminoacid_dists directory and the base_protein.pdb and protein_translations.csv files and outputs to the protein_pdb directory.

If you haven't already, get the relative side chain distances:

python side_chain_dists.py

Then run the script:

python create_protein_translation_pdbs.py

Converting ligand PDB files into PDBQT files

Make sure you have your MGL_ROOT variable set. Then run:

./ligand_pdb_to_pdbqt.sh

This takes files from the ligand_pdb directory and puts the converted files in the ligand_pdbqt directory.

Converting protein PDB files into PDBQT files

Make sure you have your MGL_ROOT variable set. Then run:

./protein_pdb_to_pdbqt.sh

This takes files from the protein_pdb directory and puts the converted files in the protein_pdbqt directory.

Docking the ligands to the proteins

Make sure you have your VINA_ROOT variable set. Then run:

python3 dock.py

This takes files from the protein_pdbqt and ligand_pdbqt directories and outputs files in the docked directory. It uses the configuration file conf.txt. Logs are placed in the vina_logs directory.

By default, this command will not dock pairs that have already been docked. To do so, add the -o command line option:

python3 dock.py -o

Blocking the docked pairs

Make sure you've installed PyMol and are running version 2.3.4.

./prep_pockets.sh

This takes files from the docked directory, creates a directory for each file in the blocked directory, and creates protein and ligand files in each created directory.

Preparing HDF files for running through pafnucy

Make sure you have your PAFNUCY_ROOT environment variable set. If using the GPU environment, use one of the oat machines by running the following command from one of the rice machines:

srun --pty --partition=gpu --gres=gpu:1 --qos=interactive $SHELL -l

Then run:

./pafnucy_prepare.sh [options]

This reads from the blocked directory and the affinities.csv file and outputs HDF files in the datasets directory for the training, validation, and test datasets. By default, these files are data_train.hdf, data_val.hdf, and data_test.hdf respectively, but these can be configured with the options below.

Options (all are optional)

  • --shuffle, -s: shuffle the dataset before partitioning. The default is to not shuffle if this option is not specified.
  • --train, -tr: a number between 0 and 1 representing the fraction of the dataset that should be used for training. This is 1 by default.
  • --val, -v: a number between 0 and 1 representing the fraction of the dataset that should be used for validation. This is 0 by default.
  • --prefix, -p: the prefix to use for the data files. This is "data" by default.

The portion of the dataset that is not partitioned to the training or validation set will be partitioned to the test set.

Example:

./pafnucy_prepare.sh --shuffle --train 0.8 --val 0.1 --prefix data2