This Perspective surveys the central methodological challenges in developing ML/MM frameworks, including the generation of high-quality reference data and the treatment of multiscale coupling.
Abstract
The integration of QM/MM methods with molecular dynamics (MD) simulations has become a powerful tool for elucidating enzymatic reaction mechanisms at atomistic resolution, providing valuable insights for biocatalyst design and drug development. However, the high computational cost of QM methods, amplified by the extensive configurational sampling inherent in MD, remains a key bottleneck in studying complex enzymatic processes. Hybrid machine learning/molecular mechanics (ML/MM) methods, which replace the quantum calculations with machine-learned interatomic potentials, offer a promising alternative toward near-QM/MM accuracy at substantially reduced computational cost. This Perspective surveys the central methodological challenges in developing ML/MM frameworks, including the generation of high-quality reference data and the treatment of multiscale coupling. Emerging applications in biosystems are highlighted, and future directions are outlined toward accurate, efficient, and broadly transferable ML/MM models to support next-generation workflows for biomolecular simulations.
Recent developments in ML potentials, ML collective variables, and committor-based sampling are presented as innovative methods that have been able to address some of the current challenges in accuracy, sampling efficiency, and the identification of low-dimensional representations of reaction coordinates.
Rui P. P. Neves, J. Coimbra, Pedro Paiva et al.· Chemical Science· 0 citations
Molecular dynamics (MD) simulations have become an increasingly important component of modern medicinal chemistry and structure-based drug discovery, providing atomistic insight into protein-ligand interactions that extends beyond static experimental structures and docking models. By explicitly accounting for conformational flexibility, solvent effects, and time-dependent behaviour, MD simulations enable the refinement of binding poses, the identification of transient and allosteric sites, and the quantitative estimation of binding thermodynamics and kinetics, the latter increasingly accessible through Markov state models (MSMs) and milestoning approaches that reconstruct long-timescale behaviour from ensembles of short trajectories. In this mini-review, we provide a practical overview of classical atomistic MD methodologies commonly used in medicinal chemistry, including force-field-based simulations, enhanced sampling techniques, and free-energy calculation methods such as alchemical and end-point approaches. Emphasis is placed on the strengths and limitations of each technique, with particular attention to their appropriate use across different stages of the drug discovery pipeline. We further discuss best practices for system preparation, simulation protocol design, convergence assessment, and reproducibility, highlighting common pitfalls that can lead to overinterpretation of simulation results. Selected examples illustrate how MD simulations have informed medicinal chemistry decisions in lead identification and optimisation. Finally, we briefly outline emerging directions, including the integration of machine learning, ensemble-based approaches, and next-generation force fields, which are expected to further expand the role of MD simulations in medicinal chemistry.
S. S. Çınaroğlu· Mini-Reviews in Medical Chem...· 0 citations
A physics-based framework for the relationship between the electrostatic energy and the electrostatic field generated by the MM environment, MIREANN exhibits powerful transferability in reproducing the free energy barriers predicted by QM/MM-MD with errors less than 0.5 kcal mol–1 for various enzyme variants without the need to retrain the MLP.
Xinhu Sha, Xuehui Guo, Zi-Yu Chen et al.· Journal of Chemical Theory a...· 0 citations
Electronic rearrangements associated with bond forming/breaking in catalytic enzymes require quantum mechanical (QM) treatment beyond classical molecular mechanics (MM). Hybrid QM/MM methods enable tractable simulations but require system-specific setup and are sensitive to the QM region choice and treatment of the QM/MM interface. We demonstrate quantum-accurate treatment of all-atom, complete enzymes in explicit solvent comprising up to 54k atoms and 1 microsecond of total simulation time using the machine-learned interatomic potential (MLIP) eSEN-omol. We reproduce experimental barrier trends for Claisen rearrangement in chorismate mutase, resolve critical intermediate states in PETase catalyzed polymer depolymerization, and distinguish mechanistic alternatives for metal-activated phosphoryl transfer in nucleoside diphosphate kinase. We realize 1000x speedups relative to typical QM/MM calculations without system-specific tuning. These results establish MLIPs as a practical route to QM-accurate simulations of enzyme catalysis.
Meng Gao, Armin Shayesteh Zadeh, Aniruddha Seal et al.· 0 citations
An integrated workflow for MLP/ΔMLP-assisted QM/MM simulations that achieves ab initio (ai) or density functional theory (DFT)-level accuracy in predicting enzyme reaction thermodynamics and incorporates an iterative model refinement strategy that systematically improves predictive performance through successive rounds of sampling, high-level labeling, and retraining.
Abdul Raafik Arattu Thodika, S. Panda, Xiao-Liang Pan et al.· Journal of Chemical Theory a...· 0 citations
Molecular dynamics simulation (MD) has become an important molecular-level approach for addressing the increasing global demand for efficient and sustainable water treatment technologies at the atomic scale. However, conventional methods of molecular dynamics simulation remain limited by transferability of empirical force fields, unreliable property prediction, inaccessible mechanistic pathways, and inefficient, labor-intensive methodologies. Here, we discuss artificial intelligence (AI)-enhanced molecular dynamics simulation (AI-MD) as a set of computational strategies that can support broader material screening, property prediction, mechanistic analysis, and more efficient simulation workflows. The review provides a systematic overview of fundamental concepts, theoretical foundations, and the evolutionary trajectory of artificial intelligence assisted molecular dynamics. Then, the rational construction of water treatment materials across multiple scales is discussed, encompassing structures from zero-dimensional single-atom catalysts to three-dimensional porous and composite materials and enabled by AI-driven structural optimization. Furthermore, the role of AI models in supporting predictive performance modeling is evaluated, through high-dimensional descriptor engineering and quantitative structure-activity relationship analysis. The mechanistic resolution of interfacial transport and reactive events is also elucidated, capturing the fundamental physics governing water purification processes. Methodological strategies for enhancing simulation efficiency and the bridging role of AI in multiscale modeling are further discussed to address current spatial and temporal limitations. Overall, this review provides a molecular-level perspective on AI-MD-enabled water treatment, linking interfacial structure, transport behavior, and reactive mechanisms to predictive and automated material design. These insights can support the development of more efficient and environmentally sustainable water-treatment technologies.
Jia Yan, Di Wu, Yixuan Cai et al.· Environmental Research· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
