conda create -n pl python=3.9
conda activate pl
conda install mpich
conda install mpi4py
conda install openmm=8.0 -c conda-forge
conda install plumed=2.8.2=mpi_mpich_h7ded119_0 -c conda-forge
conda install py-plumed -c conda-forge
conda install cmake swig
conda install pandas
conda install mdtraj -c conda-forge
git clone https://github.com/husseinmur/OpenMM-Plumed-MPI
cd OpenMM-Plumed-MPI
mkdir build install openmm -p plumed/include -p plumed/lib
unzip openmm.zip -d openmm
unzip plumed_include.zip -d plumed/include
unzip plumed_lib.zip -d plumed/lib
cd build
ccmake ..
click c to configure and set (replace paths as required):
CMAKE_INSTALL_PREFIX {git clone path}/OpenMM-Plumed-MPI/install
MPI4PY_DIR {user path}/miniconda3/envs/pl/lib/python3.9/site-packages/mpi4py/include
OPENMM_DIR {git clone path}/OpenMM-Plumed-MPI/openmm
PLUMED_INCLUDE_DIR {git clone path}/OpenMM-Plumed-MPI/plumed/include
PLUMED_LIBRARY_DIR {git clone path}/OpenMM-Plumed-MPI/plumed/lib
Keep everything else as is and click c then g then:
make
make install
make PythonInstall
cd ../install/lib
cp -r * ~/miniconda3/envs/pl/lib
In the script folder, run simulate.py
as follows:
mpirun -np {number of replicas} python simulate.py
Note: you might need to replace the checkpoint, the forcefield xml, and other files if you need to run the simulation on a different system. The provided files are prepared for TDP-43.
To generate the forcefield xml and the exclusions pickle for a different system, you can use the gen_xml_and_constraints.py
script in the scripts folder. It takes a fasta file as a parameter, and the file should only include a fasta sequence.