Molecular modeling and simulation are now essential tools for comprehending and engineering intricate molecular systems in chemistry, biology, and materials science. However, realistic environments, such as crowded biomolecular assemblies, heterogeneous interfaces, and multicomponent mixtures, pose substantial challenges for sampling, accuracy, and interpretation. At the same time, the rapid growth of machine learning and data-driven methods is opening new possibilities for accelerating simulations, improving force fields and coarse-grained models, and extracting insight from high-dimensional trajectory data. Bringing together classical physics-based approaches with modern AI tools offers a unique opportunity to bridge scales, close accuracy and efficiency gaps, and enable more predictive models of molecular function in realistic conditions.

The goal of this Research Topic is to showcase and stimulate advances at the interface of classical molecular modeling and simulation and data-driven, machine-learning-based approaches for complex molecular systems in realistic environments. We are interested in contributions that address core limitations of current methods, such as insufficient sampling, limited model accuracy, and the difficulty of analyzing large, multidimensional simulation datasets. Methodologies based on neural networks and deep learning are particularly welcome, especially when combined with strategies that improve the interpretability and explainability of model predictions. Relevant directions include, but are not limited to, ML-enhanced force fields and potentials, accelerated sampling schemes, data-driven collective variables, automated analysis of trajectories, and workflows that couple experiments, simulations, and AI.

We particularly encourage contributions targeting complex biomolecular systems in realistic intracellular and mesoscale environments, where heterogeneity, crowding, and active, non-equilibrium processes create major modeling bottlenecks. This includes open problems such as multiscale modeling of chromatin structure and (re)organization (genome architecture and nuclear organization), and self-assembly, remodeling, or disruption of cellular machinery and intracellular structures/organelles under physiological and pathological conditions.

Additionally, we welcome biomolecular studies that connect method development to mechanistic questions in complex cellular contexts, where interpretability/explainability and physically grounded generalization are crucial. By gathering work from environmental, soft-matter, and biomolecular/therapeutic applications under a shared methodological umbrella, this Research Topic aims to highlight cross-cutting ideas, identify common challenges, and point to promising routes for more robust, interpretable, and design-oriented simulations.

Themes of interest include, but are not limited to:
• Molecular modeling and simulation of complex molecular systems in realistic environments, including condensed-phase, interfacial, and multicomponent systems.
• Studies that develop or leverage data-driven and machine-learning methods to improve sampling efficiency, model accuracy, analysis of high-dimensional simulation data, or rational design workflows.
• Methodological developments based on neural networks and deep learning, especially when coupled with strategies that enhance the interpretability and explainability of model predictions.
• Applications to contaminant transport and remediation materials, environmental interfaces, and functional materials for separation or catalysis.
•Applications to soft-matter and biomolecular assemblies, including self-assembled, supramolecular, and out-of-equilibrium systems.
• Biomolecular modeling in complex intracellular environments, including crowding/heterogeneity and active, non-equilibrium effects.
• Chromatin structure, dynamics, and (re)organization across scales, including data-driven multiscale models of genome architecture and nuclear organization.
• Self-assembly, remodeling, and disruption of cellular machinery and intracellular structures/organelles in physiological and pathological settings.
• Applications to biosensing platforms and bio-inspired molecular data-storage or information-processing systems.
• Applications to therapeutically relevant molecular targets, drug discovery pipelines, and computational structural biology.

Are you interested in contributing a manuscript to this Research Topic?

Molecular Modeling and Simulation of Complex Molecular Systems with Physics-Based and Data-Driven Approaches