OpenMM can be installed with conda or pip. The official guide recommends Miniconda for an isolated, reproducible environment.
conda create -n md-openmm python=3.10
conda activate md-openmm
conda install -c conda-forge openmm
If you have an NVIDIA GPU and want CUDA:
conda install -c conda-forge openmm cuda-version=12
pip install openmm
If you use an NVIDIA GPU with CUDA 12:
pip install "openmm[cuda12]"
If you use an AMD GPU (HIP):
pip install "openmm[hip6]"
python - <<'PY'
import openmm
print('openmm', openmm.__version__)
PY
The openmm-setup app generates ready-to-run scripts and fixes common structural issues.
conda install -c conda-forge openmm-setup
To launch it:
openmm-setup
Inside the md-openmm environment, install the packages we use in the episodes:
conda install -c conda-forge jupyterlab mdanalysis mdtraj deeptime openff-toolkit openmmforcefields pdbfixer
PyEMMA is optional (the project is frozen). If you need it, install from PyPI:
pip install pyemma
To follow the episode 06 workflow you can also build PyEMMA from source inside md-openmm (after pinning to Python 3.9) or inside a dedicated pyemma environment. With the environment activated, run:
git clone https://github.com/markovmodel/PyEMMA.git
cd PyEMMA
conda install python=3.9
conda install pybind11 cython setuptools numpy scipy matplotlib
python setup.py develop
The python setup.py develop step installs PyEMMA in editable mode, so the episodes can import it without reinstalling. If you keep working in md-openmm, rerun conda install python=3.9 before the remaining commands so the interpreter matches PyEMMA’s expectations.
We use an external directory controlled by COURSE_DIR:
export COURSE_DIR=~/Concepcion26
if [ ! -d "$COURSE_DIR" ]; then
mkdir -p "$COURSE_DIR"
fi
Download data:
cd "$COURSE_DIR"
curl -L -o Course-MD-Data.zip https://github.com/Biocomputing-Teaching/Course-MD-Data/archive/refs/heads/main.zip
unzip -q Course-MD-Data.zip
rm -rf "$COURSE_DIR/data"
mkdir -p "$COURSE_DIR/data"
cp -R Course-MD-Data-main/* "$COURSE_DIR/data"
rm -rf Course-MD-Data-main Course-MD-Data.zip
Download scripts and notebooks:
cd "$COURSE_DIR"
curl -L -o Curso-MD-Analisis.zip https://github.com/Biocomputing-Teaching/Curso-MD-Analisis/archive/refs/heads/main.zip
unzip -q Curso-MD-Analisis.zip
mkdir -p "$COURSE_DIR/scripts" "$COURSE_DIR/notebooks"
find Curso-MD-Analisis-main/docs/episodes/scripts -type f -name "*.py" -exec cp -i {} "$COURSE_DIR/scripts" \;
find Curso-MD-Analisis-main/docs/episodes/notebooks -type f -name "*.ipynb" -exec cp -i {} "$COURSE_DIR/notebooks" \;
rm -rf Curso-MD-Analisis-main Curso-MD-Analisis.zip
Download the Amber dataset archive referenced by the course and keep it alongside the other inputs:
cd "$COURSE_DIR/data"
curl -L -o 1.AOM_amber.tar.gz https://www.dropbox.com/scl/fi/vltio5d6l3ghg3n5gb7tu/1.AOM_amber.tar.gz?rlkey=nv7fi1lp6k27u0iccdnlv22ms&st=kt8d35xd&dl=0
tar -xzf 1.AOM_amber.tar.gz
rm 1.AOM_amber.tar.gz
The extracted files remain in $COURSE_DIR/data, ready for the Amber routines that come later in the syllabus.
Expected structure:
data/ input filesresults/ outputs and analysisscripts/ course scriptsnotebooks/ course notebooksYou can also download the official OpenMM repository to access the example scripts:
cd "$COURSE_DIR"
git clone git@github.com:openmm/openmm.git
cd openmm/examples/python-examples