Installing gmx_MMPBSA¶

Requirements¶

gmx_MMPBSA requires GROMACS (series 4.x.x or 5.x.x or 20xx.x) and AmberTools20 or 21 to be installed in your machine with Python3. gmx_MMPBSA has been tested with GROMACS 4.6.7, 5.1.2, 2018.3, 2020.4 and 2022.4, although it should run smoothly with any GROMACS present in the PATH and that is compatible with the files you are using.

Installing gmx_MMPBSA v1.5.x¶

Currently, gmx_MMPBSA can be installed using two ways:

conda environment

The conda environment provides a clean and efficient way of installing gmx_MMPBSA. It also allows to have different versions of gmx_MMPBSA in isolated environments, thus reducing the possibility of incompatibility with other packages. Installation time is also less since it does not require the compilation of AmberTools or GROMACS. (Recommended, especially if you want to keep older versions of gmx_MMPBSA)

AmberTools compilation

In this way, we assume that you have AmberTools compiled on your machine and that you want to do an installation without worrying about enabling or disabling conda environments. It also involves user compilation of GROMACS, which takes considerable installation time. This way also requires installed packages to be compatible and installation errors are more frequent.

// Create a new environment and use the *.yml file to install dependencies
$ conda env create -file env.yml

// To use gmx_MMPBSA, just activate the environment
$ conda activate gmxMMPBSA

conda env create -f env.yml                                                     # (1)
conda activate gmxMMPBSA                                                        # (2)

// Update conda
$ conda update conda

// Create a new environment and activate it
$ conda create -n gmxMMPBSA python=3.11.8 -y -q 
$ conda activate gmxMMPBSA

// Install mpi4py and AmberTools
$ conda install -c conda-forge "mpi4py=4.0.1" "ambertools<=23.3" -y -q

// Install dependencies for ploting
$ conda install -c conda-forge "numpy=1.26.4" "matplotlib=3.7.3" "scipy=1.14.1" "pandas=1.5.3" "seaborn=0.11.2" -y -q

// Install PyQt6 required to use the GUI analyzer tool (gmx_MMPBSA_ana). Not needed for HPC
$ python -m pip install "pyqt6==6.7.1"    

// (Optional) Install GROMACS
$ conda install -c conda-forge "gromacs<=2023.4" pocl -y -q

conda update conda
conda create -n gmxMMPBSA python=3.11.8 -y -q                                      # (1)
conda activate gmxMMPBSA                                                        # (2)
conda install -c conda-forge "mpi4py=4.0.1" "ambertools<=23.3" -y -q               # (3)
conda install -c conda-forge "numpy=1.26.4" "matplotlib=3.7.3" "scipy=1.14.1" "pandas=1.5.3" "seaborn=0.11.2" -y -q # (4)
python -m pip install "pyqt6==6.7.1"                                            # (5)
# Optional
conda install -c conda-forge "gromacs<=2023.4" pocl -y -q                         # (6)

// INSTALLATION
$ python -m pip install gmx_MMPBSA

// UPDATE
$ python -m pip install gmx_MMPBSA -U

// INSTALLATION
$ python -m pip install git+https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA

// UPDATE
$ python -m pip install git+https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA -U

// Install gmx_MMPBSA
$ amber.python -m pip install gmx_MMPBSA                                               

// Update gmx_MMPBSA
$ amber.python -m pip install gmx_MMPBSA -U

// Install gmx_MMPBSA
$ amber.python -m pip install git+https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA     

amber.python -m pip install git+https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA -U

amber.conda install -c conda-forge mpi4py=3.1.3

amber.conda install pip

$ curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
---> 100%

$ chmod +x Miniconda3-latest-Linux-x86_64.sh

$ ./Miniconda3-latest-Linux-x86_64.sh
---> 100%

Successful miniconda intallation

curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh   # (1)
chmod +x Miniconda3-latest-Linux-x86_64.sh                                      # (2)
./Miniconda3-latest-Linux-x86_64.sh                                             # (3) 

Extra Dependencies¶

gmx_MMPBSA uses some dependencies for other functions independent of calculations or in some cases they may be necessary due to the nature of your OS.

ParmEd

The current version of ParmEd implemented in AmberTools has some limitations that have been resolved in the GitHub repository by its author Jason Swails and others with our help.

Some of these limitations are:

• Error reading topology when it has insertion codes
• Error processing topologies generated with the Amber ff19SB force field
• New PBRadii sets for GAFF and CHARMM force fields

pip

In some cases, the miniconda environment created in the AmberTools compilation does not have the pip module, so any installation that depends on this package will fail. Required only if you did not install gmx_MMPBSA via conda

amber.conda install pip    

Git

Used by gmx_MMPBSA_test to download the GitHub repository to get the examples' folder or to install the development version.

conda install -c anaconda git

or

sudo apt install git 

mpi

In some cases it is necessary to install the MPI dependencies. Required only if you did not install gmx_MMPBSA via conda

sudo apt install openmpi-bin libopenmpi-dev openssh-client

libxcb

If you get an error related to Qt plugins

sudo apt install --reinstall libxcb-xinerama0

Troubleshooting after installation¶

Once the installation is completed, the following warning may appear:

WARNING: The scripts gmx_MMPBSA, gmx_MMPBSA_ana and gmx_MMPBSA_test 
are installed in '/home/user/path_to_amber_install/amber20/miniconda/bin'
which is not on PATH.

This warning is because pip installs the executables (gmx_MMPBSA, gmx_MMPBSA_ana and gmx_MMPBSA_test) in installation_path/amber20/miniconda/bin.

You have two options to solve this:

• Add this folder (*/amber20/miniconda/bin) to PATH:

  export PATH="/path_to_amber_install/amber20/miniconda/bin:$PATH"
  

  Tip

  • This option is more permanent and is recommended if you don't want to activate and deactivate the conda environment
  • Make sure to update path_to_amber_install in the PATH variable

• Initializing the environment of conda amber:

  amber.conda init bash
  

  You can deactivate like this:

  conda deactivate
  

If when running gmx_MMPBSA, you get an error like this:

ModuleNotFoundError: No module named 'parmed'

please see the following issue to see the solution

Autocompletion script¶

Since gmx_MMPBSA has many flags, we believe that this autocompletion can significantly improve productivity, be more user-friendly and reduce the number of unforced errors. That is why we created this script, which manages the autocompletion of the gmx_MMPBSA, gmx_MMPBSA_ana and gmx_MMPBSA_test.

Execution: Enter the following command in the terminal:

source /path/to/ambertools/lib/python3.x/site-packages/GMXMMPBSA/GMXMMPBSA.sh

Once you make the source of GMXMMPBSA.sh you can check its operation as follows:

All you have to do is enter the name of the program in the terminal and press the tab key twice:

gmx_MMPBSA <tab> <tab>

Testing the operation of gmx_MMPBSA¶

After preparing everything to run gmx_MMPBSA, it only remains to check its correct operation. To know how to do it, consult the documentation of gmx_MMPBSA_test