Installation

1. Get sources from the GithubRepo 

The sources for Gromacs_py can be downloaded from the GithubRepo.

You can either clone the public repository:

$ git clone git://github.com/samuelmurail/gromacs_py

Or download the tarball:

$ curl -OJL https://github.com/samuelmurail/gromacs_py/tarball/master

Once you have a copy of the source, switch to the gromacs_py directory.

$ cd gromacs_py

2. Create Conda Environment

You need to create a conda environment to be able to use:

Gromacs
Rdkit Used for ligand parametrization, convert SMILE to pdb.
Antechamber Amber tools for ligand parametrization.
Acpype a python tool to use antechamber.
Apbs Pdb2pqr Protein protonation calculation.

Use conda en create to create it using the .conda.yml file. You can overide the environmnent name using the option --name YOUR_NAME.

$ conda env create -f .conda.yml

If you plan to use gromacs_py in jupyter notebook, you should try the jupyter version:

$ conda env create -f .conda_jupyter.yml

This will create an environmnet called gromacs_py (or the name you defined). You will then, need to activate the environmnent:

$ conda activate gromacs_py

Note

If you want to install yourself Gromacs to be able to use the GPU acceleration, you can use the .conda_no_gromacs.yml or .conda_jupyter_no_gromacs.yml:

$ conda env create -f .conda_no_gromacs.yml
$ conda activate gromacs_py

3. Install gromacs_py

Once you have a copy of the source and have create a conda encironment, you can install it with:

$ python setup.py install

4. Test Installation

To test the installation, simply use pytest:

$ pytest
=========================== test session starts ========================
platform linux -- Python 3.8.2, pytest-5.4.2, py-1.9.0, pluggy-0.13.1
rootdir: /home/murail/Documents/Code/gromacs_py, inifile: pytest.ini
plugins: cov-2.10.1
collected 30 items

gromacs_py/gmx.py .............                                   [ 43%]
gromacs_py/test/test_FreeEner.py ......                           [ 63%]
gromacs_py/test/test_GmxSys.py ..                                 [ 70%]
gromacs_py/tools/ambertools.py ....                               [ 83%]
gromacs_py/tools/monitor.py .....                                 [100%]

======================= 30 passed in 236.83s (0:03:56) =================

Conda installation

If you don’t need a GPU compiled version of Gromacs you can use directly the Gromacs_py conda package to install both Gromacs software and Gromacs_py library:

conda install -c bioconda gromacs_py

Pypi (Deprecated)

If gromacs (version >= 5.1) is already install, then install you need to install the gromacs_py library, and add the gromacs gmx command in the environmnent variable $PATH:

pip install gromacs_py

# Add gromacs 'gmx' path:
export PATH='*path_to_gromacs*/bin/':$PATH

Without Conda

Get the gromacs_py library from github.

git clone https://github.com/samuelmurail/gromacs_py.git
./setup.py install --user

# Add gromacs 'gmx' path:
export PATH='*path_to_gromacs*/bin/':$PATH

Prerequisites

python 3 libraries installed when you launch the pip command:
- numpy
- scipy
- pandas
- matplotlib
- Sphinx and sphinx-argparse (only for building documentation)
- Os_Command_py
- PDB_Manip_py
- PDB2PQR using the package pdb2pqr_htmd_propka30 a python 3 version developped by tonigi and adapted to use successfully propka3.0.

Gromacs

Get source code from gromacs website and follow the following command for a quick and dirty install (for more details see gromacs 2019 install guide)

In my case I add to change few options to cmake:

-DCMAKE_C_COMPILER=gcc-8, as gcc versions later than 6 are not supported.

-DGMX_GPU=on to use GPU acceleration

-DCMAKE_INSTALL_PREFIX=../../local-gromacs-2019.2/ to install gromacs in a non-standard location

# Specify the version:
version="2021.5"
# To modify:
dir_install="/home/murail/Documents/Software/local-gromacs-${version}"

wget https://ftp.gromacs.org/gromacs/gromacs-${version}.tar.gz
tar -xvf gromacs-${version}.tar.gz
cd gromacs-${version}
mkdir build
cd build
# In my case I needed to define ggc-8 because gromacs doesn't accept superior versions
cmake .. -DGMX_BUILD_OWN_FFTW=ON -DREGRESSIONTEST_DOWNLOAD=ON -DGMX_GPU=CUDA -DCMAKE_INSTALL_PREFIX=${dir_install} -DCMAKE_C_COMPILER=gcc-8
make
make check
make install
source ${dir_install}/bin/GMXRC
echo "export PATH=${dir_install}/bin/:\$PATH" >> ~/.bashrc

Ambertools

The easiest way is to use the conda package. However is can also be installed from source.