CHARMM

CHARMM
Name	CHARMM
Developer	Martin Karplus laboratory; Harvard University group; community contributors
Released	1980s
Latest release	ongoing
Programming language	Fortran, C, C++
Operating system	Unix, Linux, macOS, Microsoft Windows
Genre	Molecular simulation
License	Academic site-license; commercial options

CHARMM

CHARMM is a molecular simulation program widely used for modeling the structure, dynamics, and thermodynamics of biomolecular systems. Developed by teams led by Martin Karplus and maintained by an international community, CHARMM provides algorithms for molecular mechanics, molecular dynamics, Monte Carlo, and free-energy calculations applied to proteins, nucleic acids, lipids, carbohydrates, and small molecules. It interfaces with experimental and computational resources created at institutions such as Harvard University, Rutgers University, and national laboratories, and it has influenced software like NAMD, GROMACS, AMBER, and LAMMPS.

History

CHARMM originated in the 1970s–1980s from theoretical chemistry research groups including Martin Karplus and collaborators at Harvard University and Columbia University. Early releases implemented energy functions and dynamics algorithms developed in parallel with force-field work by groups linked to Richard A. Friesner and Arieh Warshel. Over decades CHARMM evolved through contributions from laboratories at Rutgers University, University of California, San Diego, and national centers such as Brookhaven National Laboratory, expanding support for biomolecules studied by researchers at Max Planck Society and CNRS. Key historical milestones include incorporation of extended force fields in the 1980s, development of generalized Born implicit solvent methods in the 1990s, and integration of polarizable force fields and QM/MM coupling in the 2000s. CHARMM’s development paralleled experimental advances at facilities like Brookhaven National Laboratory and Argonne National Laboratory and computation platforms established by National Science Foundation initiatives.

Theory and Force Field Development

CHARMM implements classical mechanics frameworks grounded in potential-energy functions parameterized as force fields developed by groups associated with Charles L. Brooks III, Martin Karplus, Alexander D. MacKerell Jr., and collaborators. The functional form includes bonded terms (bonds, angles, dihedrals) and nonbonded terms (Lennard-Jones, Coulomb) with extensions such as CMAP corrections, Drude oscillator polarizability, and scaled-charge models. Force-field families maintained and distributed through CHARMM efforts include parameter sets derived by Alexander D. MacKerell Jr. and teams, with validation against experimental observables from laboratories including Columbia University and University of Pennsylvania. CHARMM supports mixed quantum mechanics/molecular mechanics (QM/MM) schemes leveraging methods developed in collaboration with researchers such as William A. Goddard III and Mark S. Gordon, enabling comparison to spectroscopic data from groups at California Institute of Technology and University of Illinois Urbana-Champaign.

Software Architecture and Modules

CHARMM’s codebase is organized into modular components implemented in Fortran and C/C++, enabling extensibility by academic and industrial groups like those at Rutgers University and Harvard University. Core modules provide energy evaluation, integrators (Verlet, velocity-Verlet, Langevin), constraint solvers (SHAKE, RATTLE), and PME electrostatics developed alongside methods by teams at Lawrence Livermore National Laboratory and Oak Ridge National Laboratory. Specialized modules support alchemical transformations, replica-exchange, generalized Born implicit solvent, and enhanced-sampling algorithms influenced by work at University of California, San Francisco and University of Pittsburgh. Interoperability layers facilitate coupling to visualization and analysis packages such as VMD, Chimera, and PyMOL and to multiscale frameworks used at Argonne National Laboratory and Los Alamos National Laboratory.

Applications and Use Cases

CHARMM is applied across structural biology, biophysics, and materials science in studies conducted by investigators at Harvard Medical School, Stanford University, Massachusetts Institute of Technology, and pharmaceutical companies including Pfizer and Merck & Co.. Representative applications include protein folding and misfolding studies tied to research at Scripps Research Institute, membrane simulation work in labs at University of California, San Diego, nucleic-acid conformational dynamics examined by teams at University of Cambridge, and ligand-binding free-energy calculations used in drug-design pipelines at Novartis. CHARMM has also been used for carbohydrate modeling in collaborations with Johns Hopkins University and for coarse-grained and multiscale simulations integrated with methods from Max Planck Institute groups.

Validation and Benchmarking

Validation of CHARMM force fields and algorithms is performed against experimental data produced by synchrotron facilities like Diamond Light Source and European Synchrotron Radiation Facility, NMR studies from groups at Bruker-linked labs, calorimetry datasets from National Institutes of Health-funded centers, and high-level quantum-chemical benchmarks from computational chemistry groups such as Gaussian, Inc. collaborators. Benchmarking efforts compare performance and accuracy with packages including NAMD, GROMACS, and AMBER across hardware platforms developed by Intel, NVIDIA, and supercomputing centers like Oak Ridge National Laboratory. Community-driven validation suites and blind challenges involving consortia such as the Drug Design Data Resource and collaborations with European Molecular Biology Laboratory provide cross-validation of thermodynamic and kinetic predictions.

Licensing and Distribution

CHARMM is distributed under academic and commercial licensing models administered by entities connected to the original developer consortium and university technology-transfer offices at Rutgers University and Harvard University. Academic licenses permit use by investigators at institutions such as Yale University and University of Oxford under site-license terms, while commercial licenses are available to companies like GlaxoSmithKline and contract research organizations. Source-code access and module licensing policies are coordinated with repositories and distribution partners used by research groups at National Center for Supercomputing Applications and hardware vendors like NVIDIA for optimized builds.

Category:Molecular dynamics software