Installation Guide¶

This guide provides detailed instructions for installing PRISM and its dependencies on various platforms.

Quick Start

git clone https://github.com/AIB001/PRISM.git && cd PRISM
pip install -e .
prism --help

Prerequisites¶

PRISM requires several components to be installed. Install them in the following order:

1. GROMACS (Required)¶

GROMACS is essential for PRISM to build and run molecular dynamics simulations. PRISM uses GROMACS for: - Processing protein structures (pdb2gmx) - Building simulation boxes and solvation (editconf, solvate) - Adding ions (genion) - Running MD simulations (mdrun)

Follow these steps to install GROMACS with GPU support:

Check CUDA Toolkit¶

First, verify that you have CUDA installed:

nvcc --version

If CUDA is not installed, please install it from NVIDIA's CUDA Toolkit page before proceeding.

Install Build Dependencies¶

Ubuntu/DebianCentOS/RHELmacOS

# Install essential build tools
sudo apt-get update
sudo apt-get install gcc g++ cmake
sudo apt-get install build-essential

# Install essential build tools
sudo yum groupinstall "Development Tools"
sudo yum install cmake gcc-c++

# Install using Homebrew
brew install cmake gcc

Download and Build GROMACS¶

# Download GROMACS 2024.3
wget https://ftp.gromacs.org/gromacs/gromacs-2024.3.tar.gz
tar xfz gromacs-2024.3.tar.gz

# Create build directory
cd gromacs-2024.3
mkdir build
cd build

# Configure with CMake (adjust paths as needed)
cmake .. -DGMX_MPI=ON \
         -DGMX_BUILD_OWN_FFTW=ON \
         -DGMX_GPU=CUDA \
         -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda \
         -DCUDA_INCLUDE_DIRS=/usr/local/cuda/include \
         -DCUDA_CUDART_LIBRARY=/usr/local/cuda/lib64 \
         -DCMAKE_INSTALL_PREFIX=/opt/gromacs-2024.3

# Build GROMACS (use all available cores)
make -j$(nproc)

# Run tests (optional but recommended)
make check

# Install GROMACS
sudo make install

# Source GROMACS environment
source /opt/gromacs-2024.3/bin/GMXRC

Add to your shell configuration

Add the following line to your ~/.bashrc or ~/.zshrc to automatically source GROMACS:

source /opt/gromacs-2024.3/bin/GMXRC

Install Bioconda (Optional)¶

For additional bioinformatics tools:

# Install bioconda channel
conda config --add channels bioconda
conda config --add channels conda-forge

2. Python Environment (Required)¶

PRISM requires Python 3.10 or higher. We recommend using Conda for environment management:

# Create a new environment for PRISM
conda create -n prism python=3.10
conda activate prism

Python Version

While Python 3.10 is recommended, PRISM supports Python 3.8 through 3.11. Python 3.10 offers the best balance of compatibility and performance.

3. Core Python Dependencies (Required)¶

Install essential Python packages:

# Activate PRISM environment
conda activate prism

# Install PDBFixer and core scientific packages
conda install -c conda-forge pdbfixer numpy scipy

# Install YAML support
pip install pyyaml

Force Field Specific Dependencies¶

PRISM supports 8+ ligand force fields. Install dependencies based on your needs:

For GAFF / GAFF2 Support (Most Common)¶

The General AMBER Force Field is the default and most widely used:

# AmberTools (provides antechamber, parmchk2)
conda install -c conda-forge ambertools

# ACPYPE for GROMACS conversion (required)
pip install acpype

# RDKit for molecular structure handling (recommended)
conda install -c conda-forge rdkit

For OpenFF Support¶

Modern, systematically improved force fields:

# OpenFF toolkit and interchange
conda install -c conda-forge openff-toolkit openff-interchange

# RDKit for molecular reading
conda install -c conda-forge rdkit

For CGenFF Support¶

CHARMM General Force Field (requires web download):

Upload your ligand to ParamChem
Download the generated parameter files
Provide the directory path to PRISM with --forcefield-path

# No additional software needed - uses downloaded files
prism protein.pdb ligand.mol2 -o output --ligand-forcefield cgenff --forcefield-path /path/to/cgenff_files

For OPLS-AA Support¶

Uses LigParGen web server (requires internet):

# RDKit for structure handling
conda install -c conda-forge rdkit

# No additional software - uses LigParGen API

For MMFF / MATCH / Hybrid Support¶

Uses SwissParam web server (requires internet):

# RDKit for structure handling
conda install -c conda-forge rdkit

# No additional software - uses SwissParam API

Installing PRISM¶

Once all prerequisites are installed, you can install PRISM:

Recommended: Development Installation¶

This allows you to modify PRISM and immediately see changes:

# Clone the PRISM repository
git clone https://github.com/AIB001/PRISM.git
cd PRISM

# Install in development mode
pip install -e .

# Verify installation
prism --help
python -c "import prism; print(prism.__version__)"

Alternative: Direct Usage (Without Installation)¶

You can also use PRISM directly:

# Clone the repository
git clone https://github.com/AIB001/PRISM.git
cd PRISM

# Run PRISM module directly
python -m prism.builder protein.pdb ligand.mol2 -o output_dir

Verification¶

After installation, verify that everything is working correctly:

Test GROMACS Installation¶

# Check GROMACS version
gmx --version

# Should show version 2024.3 with CUDA support

Test Python Environment¶

# Activate PRISM environment
conda activate prism

# Test core functionality
python -c "import prism; print(f'PRISM version: {prism.__version__}')"
python -c "import pdbfixer; print('PDBFixer: OK')"
python -c "import yaml; print('YAML: OK')"
python -c "import numpy; print('NumPy: OK')"

Check Force Field Dependencies¶

# Check if AmberTools is available (for GAFF/GAFF2)
which antechamber && echo "AmberTools: OK"

# Check if ACPYPE is available
which acpype && echo "ACPYPE: OK"

# Check if OpenFF is available
python -c "import openff.toolkit; print('OpenFF: OK')" 2>/dev/null || echo "OpenFF: not installed (optional)"

# Check if RDKit is available
python -c "import rdkit; print('RDKit: OK')" 2>/dev/null || echo "RDKit: not installed (optional)"

List Available Force Fields¶

# Check installed protein force fields
prism --list-forcefields

Test System Building¶

# Quick test with example files (if available)
prism test_protein.pdb test_ligand.mol2 -o test_output

# Check output structure
ls -lh test_output/
ls -lh test_output/GMX_PROLIG_MD/

Environment Variables¶

Set these environment variables for optimal performance:

# Add to ~/.bashrc or ~/.zshrc

# GROMACS path
export GMXRC=/opt/gromacs-2024.3/bin/GMXRC
source $GMXRC

# CUDA paths (adjust as needed)
export CUDA_HOME=/usr/local/cuda
export PATH=$CUDA_HOME/bin:$PATH
export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH

# PRISM path (optional, for direct usage)
export PRISM_HOME=/path/to/PRISM
export PATH=$PRISM_HOME/bin:$PATH

Troubleshooting¶

Common Installation Issues¶

GROMACS build fails with CUDA errors

Solution: Ensure CUDA paths are correct:

# Find CUDA installation
which nvcc
ls /usr/local/cuda*

# Update CMake command with correct paths
cmake .. -DCUDA_TOOLKIT_ROOT_DIR=/correct/cuda/path ...

ImportError: No module named 'prism'

Solution: Ensure you're in the correct conda environment:

conda activate prism
which python  # Should point to conda environment
pip list | grep prism

PDBFixer installation fails

Solution: Install from conda-forge channel:

conda install -c conda-forge pdbfixer
# If still fails, try:
conda install -c conda-forge openmm pdbfixer

ACPYPE not found

Solution: Install with pip in the correct environment:

conda activate prism
pip install acpype --upgrade

Platform-Specific Issues¶

GPU Support on Linux¶

If GPU is not detected:

# Check NVIDIA driver
nvidia-smi

# Check CUDA installation
nvcc --version
ldconfig -p | grep cuda

# Reinstall CUDA toolkit if necessary

macOS Specific¶

For macOS users without CUDA:

# Build GROMACS without GPU support
cmake .. -DGMX_MPI=ON \
         -DGMX_BUILD_OWN_FFTW=ON \
         -DGMX_GPU=OFF \
         -DCMAKE_INSTALL_PREFIX=/opt/gromacs-2024.3

Additional Resources¶

Getting Help¶

If you encounter issues during installation: