This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| OPLS | |
|---|---|
| Name | OPLS |
| Developers | William L. Jorgensen, Michael J. Frisch, Martin Karplus, William D. Cornell |
| First published | 1980s |
| Latest release | OPLS-AA, OPLS-UA, OPLS3 |
| Programming languages | Fortran, C, C++ |
| License | proprietary and academic |
| Website | N/A |
OPLS
OPLS is a family of molecular mechanics force fields widely used in computational studies of organic molecules, biomacromolecules, and condensed-phase systems. It was developed to reproduce experimental thermodynamic and structural properties by parameter fitting to quantum chemistry and condensed-phase data, and it underpins many studies involving Monte Carlo method, molecular dynamics, drug discovery, and computer-aided drug design workflows. The family includes all-atom and united-atom variants and has influenced software projects and standards used at institutions such as DuPont, Merck & Co., Pfizer, National Institutes of Health, and academic groups at Yale University and Harvard University.
OPLS originated as an empirical potential designed for liquid hydrocarbons and was expanded to treat heteroatoms, peptides, and proteins for diverse simulations. It implements bonded terms for bonds, angles, and torsions plus nonbonded Lennard-Jones and Coulombic interactions to model intermolecular forces for systems ranging from methane and ethanol to adenine/thymine base pairing and folded proteins. The force field family balances portability for popular engines like GROMACS, AMBER, NAMD, and Desmond with parameter sets used in industrial settings at GlaxoSmithKline and computational chemistry groups at California Institute of Technology and University of Cambridge.
Development began in the 1980s by groups led by William L. Jorgensen at Yale University with early comparisons against data from X-ray crystallography and neutron diffraction for liquids and crystals. Subsequent expansions incorporated quantum chemical target data from Hartree–Fock and density functional theory calculations performed by researchers at IBM and Bell Labs; contemporaneous force fields include AMBER, CHARMM, and GROMOS. The OPLS-AA (all-atom) parameterization appeared in the 1990s with validation against enthalpy of vaporization and density measurements for organic liquids; later efforts produced OPLS-UA (united-atom) for hydrocarbons and OPLS3 for enhanced small-molecule coverage, with collaboration involving teams at Schrödinger, Inc. and contributors from Novartis and AstraZeneca.
The functional form of OPLS includes harmonic bond stretching and angle bending terms, Fourier-series torsional potentials, and nonbonded interactions comprising Lennard-Jones 12-6 terms plus Coulombic electrostatics with partial charges derived from quantum calculations or empirical fitting. Bonded terms are similar in formalism to those used in AMBER and CHARMM, while nonbonded parameters are tuned to reproduce condensed-phase properties measured by groups at NIST and instrumental techniques at Oak Ridge National Laboratory. Electrostatic interactions can be treated with Ewald summation variants such as particle mesh Ewald for periodic systems or reaction-field methods used in early Monte Carlo studies at Argonne National Laboratory.
Parameters in OPLS are obtained by fitting to a mixture of high-level quantum chemistry data (conformations, torsional profiles) from methods like MP2 and CCSD(T) and experimental condensed-phase observables such as vaporization enthalpies, liquid densities, and solvation free energies measured in laboratories including Los Alamos National Laboratory and Lawrence Berkeley National Laboratory. Validation studies often involve benchmarking against datasets assembled by consortia at Protein Data Bank deposition centers, small-molecule thermochemistry datasets curated by GaussView users, and hydration free energy challenges organized by the SAMPL community. Cross-validation against competing force fields such as AMBER ff99SB, CHARMM36, and GROMOS 54A7 is routine in publications from groups at University of California, San Diego and University of Oxford.
Major variants include OPLS-AA (all-atom), OPLS-UA (united-atom), and commercial extensions such as OPLS3 and OPLS3e developed with input from Schrödinger, Inc. for improved small-molecule and protein-ligand modeling. Extensions address polarizability, for example via Drude oscillator implementations explored by researchers at Rutgers University and Boston University, and coarse-grained mappings akin to methods from Martini model developers. Hybrid quantum mechanics/molecular mechanics (QM/MM) schemes combine OPLS with quantum packages like Gaussian, ORCA, and Q-Chem for reaction modeling and enzyme catalysis studies at institutions such as ETH Zurich and Max Planck Institute.
OPLS is widely applied in ligand docking, free energy perturbation (FEP) calculations for binding affinity prediction used in projects at Merck Sharp & Dohme and Bayer, molecular dynamics of membrane proteins studied at Scripps Research Institute, solvation and partition coefficient estimation relevant for ADME prediction in pharmaceutical pipelines at Johnson & Johnson, and materials modeling for organic crystals and ionic liquids examined at MIT and EPFL. Large-scale workflows integrate OPLS parameters with packages like Schrodinger Maestro, ACEMD, and OpenMM for high-throughput virtual screening and prospective drug design campaigns.
Criticisms of OPLS include limited treatment of explicit electronic polarizability in base variants, difficulties in transferability for unusual chemistries encountered by cheminformatics groups at AbbVie and Boehringer Ingelheim, and occasional discrepancies with high-level quantum reference data highlighted by studies from Columbia University and University of Toronto. Parameter coverage gaps for exotic heterocycles and metalloorganic centers have driven users to augment OPLS with bespoke parameters or hybrid approaches combining DFT calculations and reparameterization workflows implemented in community tools like ParmEd and LigParGen.
Category:Force fields