pyCHARMM

pyCHARMM provides Python bindings to the CHARMM molecular dynamics and modeling engine to enable programmatic construction, simulation, and augmentation of biomolecular systems and energy functions.

Key Features:

• Python bindings: Exposes CHARMM functions and modules as Python-callable interfaces for scripted workflows.
• System variables access: Provides programmatic access to psf, coordinates, velocities, forces, atom selection variables, and force field parameters.
• CHARMM force fields: Supports CHARMM parameter sets for proteins, nucleic acids, lipids, sugars, and other biologically relevant entities.
• Energy/force augmentation: Enables inclusion of additional energy terms or force contributions implemented in Python, CUDA, or OpenCL as callable routines.
• Loosely coupled parallelism: Supports loosely coupled parallel execution for complex workflows and high-throughput calculations.
• Free energy methods: Facilitates free energy calculations including MBAR and thermodynamic integration (TI) and high-throughput multisite λ-dynamics (MSλD).
• Enhanced sampling: Supports replica exchange and string path optimization calculations.
• Molecular docking: Implements a Python-based CDOCKER module for molecular docking workflows.
• Accelerated kernels integration: Integrates CHARMM-accelerated kernels via the CHARMM/OpenMM API, CHARMM/DOMDEC, and CHARMM/BLaDE API.
• Visualization integration: Integrates with Python-based graphical engines for visualization of simulation models and results.

Scientific Applications:

• Biomolecular simulation: Molecular dynamics and structural modeling of proteins, nucleic acids, lipids, sugars, and complexes with small-molecule ligands.
• Force-field development: Development and testing of CHARMM-compatible force field parameters and ML-augmented energy terms.
• Alchemical free energy calculations: End-point and pathway free energy estimation using MBAR, TI, and multisite λ-dynamics (MSλD).
• Enhanced-sampling studies: Conformational sampling and pathway optimization using replica exchange and string path methods.
• Molecular docking workflows: Docking and scoring of ligands using the Python CDOCKER implementation for pose generation and evaluation.
• High-performance simulations: Large-scale or accelerated simulations leveraging OpenMM, DOMDEC, and BLaDE integrations.

Methodology:

Uses Python bindings to CHARMM to access psf, coordinates, velocities, forces, atom selections and force field parameters; allows callable energy terms implemented in Python, CUDA, or OpenCL; and supports CHARMM/OpenMM API, CHARMM/DOMDEC, CHARMM/BLaDE API, loosely coupled parallelism, MBAR/TI, MSλD, string path optimization, replica exchange, and CDOCKER.