FMO

FMO
Name	FMO
Classification	Computational chemistry method
Introduced	1970s
Notable users	Nobel Prize in Chemistry laureates, researchers at Max Planck Society, Lawrence Berkeley National Laboratory, Riken

FMO

FMO is a fragmentation-based quantum chemical method developed to enable electronic-structure calculations on large molecular systems. It partitions large systems into fragments for correlated wavefunction or density-functional treatments, allowing studies of biomolecules, materials, and supramolecular assemblies with reduced computational cost. The method has been employed by researchers at institutions such as Harvard University, University of Cambridge, Tokyo Institute of Technology, and Stanford University to investigate intermolecular interactions, protein–ligand binding, and materials interfaces.

Overview

FMO partitions a target system into interacting fragments and computes fragment and fragment-pair electronic states to reconstruct properties of the whole system. Early formalism links to methods developed at Møller–Plesset perturbation theory level and to partitioning strategies akin to those used in ONIOM and fragment molecular orbital schemes proposed at Los Alamos National Laboratory. The approach supports correlated treatments such as MP2, coupled cluster variants, and density functional theory within fragment calculations. It has been integrated into workflows at facilities like Oak Ridge National Laboratory and Argonne National Laboratory for large-scale simulations.

History and Development

FMO traces conceptual roots to fragmentation and embedding ideas advanced in the 1970s and 1980s at institutions such as Bell Labs and IBM Research. Key formal developments occurred in the 1990s and 2000s with contributions from groups at Osaka University, Hokkaido University, and RIKEN. Methodological extensions incorporated correlated methods and analytic gradients, enabling geometry optimizations and molecular dynamics used by teams at University of Tokyo and Kyoto University. Subsequent efforts integrated FMO with high-performance computing platforms at National Institute for Computational Sciences and the European Centre for Medium-Range Weather Forecasts for large-system electronic-structure studies.

Variants and Applications

Variants include two-body and many-body expansions, frozen-density embedding hybrids, and multilayer fragmentation schemes similar in spirit to techniques at California Institute of Technology and Massachusetts Institute of Technology. Applications span protein–ligand binding studies in complexes investigated at Columbia University and University of California, San Francisco; materials interfaces explored at ETH Zurich and Swiss Federal Laboratories for Materials Science and Technology; and supramolecular host–guest systems characterized by researchers at University of Oxford and University of Cambridge. FMO has been applied to interpret spectroscopic observations from facilities such as Diamond Light Source and Advanced Light Source and to assist drug-design projects in collaboration with pharmaceutical companies and institutes like National Institutes of Health.

Mechanisms and Principles

The core principle is fragmentation with systematic inclusion of inter-fragment interactions via pairwise or higher-order corrections, related to expansion strategies used in Many-Body Perturbation Theory and concepts from Density Embedding approaches pioneered at University of Illinois Urbana–Champaign. Electron correlation is treated within fragments by methods such as MP2, CCSD(T), or DFT functionals developed by groups at Beckman Institute and elsewhere. Electrostatic and exchange contributions are captured through embedding potentials analogous to those used in Polarizable Continuum Model studies at University of Geneva. The method yields energy decomposition analyses comparable to techniques from Symmetry-Adapted Perturbation Theory researchers.

Computational Methods and Software

Implementations appear in quantum chemistry packages developed by teams at institutions such as Fujitsu Laboratories, Quantum Chemistry Program Exchange, and corporate research groups at Schrödinger (company) and Molecular Sciences Institute. Software supports parallel execution on supercomputers like Fugaku, Summit (supercomputer), and clusters at CERN. Tools provide analytic gradients, fragment-based QM/MM integrations used in studies by groups at Max Planck Institute for Biophysical Chemistry, and interfaces to visualization packages from The Royal Society of Chemistry. Benchmarks compare FMO implementations against full-system calculations from packages like Gaussian (software) and NWChem.

Biological and Medical Contexts

FMO has been extensively used to study enzyme active sites, protein–ligand interactions, and nucleic acid conformations in projects at National Cancer Institute, Broad Institute, and European Molecular Biology Laboratory. Applications include binding-affinity predictions for inhibitors targeting proteins such as kinases studied at Wellcome Trust-funded centers, and analysis of antibody–antigen interfaces examined by research teams at Scripps Research. FMO-based interaction fingerprints have contributed to virtual screening pipelines used in collaborations with GlaxoSmithKline and Pfizer research groups.

Criticisms and Limitations

Limitations include fragmentation-induced errors for strongly delocalized electronic states encountered in studies of conjugated polymers at Max Planck Institute for Polymer Research and charge-transfer complexes examined at IBM Research. Accuracy depends on fragment definition, level of correlated treatment, and treatment of many-body terms—issues also highlighted by benchmarking consortia at National Institute of Standards and Technology and Centre National de la Recherche Scientifique. Computational overhead for high-order many-body corrections can approach full-system costs, a concern for projects at Los Alamos National Laboratory and centers using petascale supercomputers.

Category:Quantum chemistry methods