pyCHARMM: Embedding CHARMM Functionality in a Python']Python Framework

CHARMM is rich in methodology and functionality as oneof the firstprograms addressing problems of molecular dynamics and modeling ofbiological macromolecules and their partners, e.g., small moleculeligands. When combined with the highly developed CHARMM parametersfor proteins, nucleic acids, small molecules, lipids, sugars, andother biologically relevant building blocks, and the versatile CHARMMscripting language, CHARMM has been a trendsetting platform for modelingstudies of biological macromolecules. To further enhance the utilityof accessing and using CHARMM functionality in increasingly complexworkflows associated with modeling biological systems, we introducepyCHARMM, Python bindings, functions, and modules to complement andextend the extensive set of modeling tools and methods already availablein CHARMM. These include access to CHARMM function-generated variablesassociated with the system (psf), coordinates, velocities and forces,atom selection variables, and force field related parameters. Theability to augment CHARMM forces and energies with energy terms ormethods derived from machine learning or other sources, written inPython, CUDA, or OpenCL and expressed as Python callable routinesis introduced together with analogous functions callable during dynamicscalculations. Integration of Python-based graphical engines for visualizationof simulation models and results is also accessible. Loosely coupledparallelism is available for workflows such as free energy calculations,using MBAR/TI approaches or high-throughput multisite lambda-dynamics(MS lambda D) free energy methods, string path optimization calculations,replica exchange, and molecular docking with a new Python-based CDOCKERmodule. CHARMM accelerated platform kernels through the CHARMM/OpenMMAPI, CHARMM/DOMDEC, and CHARMM/BLaDE API are also readily integratedinto this Python framework. We anticipate that pyCHARMM will be arobust platform for the development of comprehensive and complex workflowsutilizing Python and its extensive functionality as well as an optimalplatform for users to learn molecular modeling methods and practiceswithin a Python-friendly environment such as Jupyter Notebooks.