Source code and cofiguration files for "Determining Lennard-Jones parameters using multiscale target data through presampling-enhanced, surrogate-assisted global optimization"

• Python 3
• Gromacs >= 2016.3
• Amber

• ./coffe/ contains the code that was used to obtains the results that we present in our work. This is basically a version of COFFE(Comprehensive Optimization Force Field Environment), which is a python package that allows us to run Gromacs simulations on a slurm batch system from python code.
• ./coffe/coffe/grow/ contains the optimization code and code to steer the experiments that we did:
  • optimization_algorithms.py the CMA-ES(class CMAES) and MSAEOP (class BayesianOptimization) algorithms
  • sampling.py the Latin Hypercube Sampling and it's optimization
  • grow_sander_ff_opt.py the Energy Minimization steering
• ./inputs/ contains the required input files
• example_call.py / example.cfg script and config file to run an example optimization run
• pipfreeze.txt a dump of installed pip packages, that where required to run our software

Install required python packages, e.g.

pip -r pipfreeze.txt

Copy oplsaa-andi.ff to a directory that is in the Gromacs force field search path. This is a copy of the OPLSAA force field.

Install the optimization software:

cd coffe
python setup.py develop --user

Check that it works by spawning a new shell and typing:

which coffe
coffe --help

Make sure that ./out/ contains only the .keep file.

To run a "small" optimization task, type

python example_call.py

This might require additional configuration of your system, depending on your configuration of Gromacs and Amber.