Soft matter research has long been characterized by intrinsic complexity, arising from multiscale structure–property relationships, strong coupling between physical and chemical processes, and the need to reconcile theoretical abstraction with experimental observability. Computational tools have therefore become indispensable for navigating this complexity. From early algorithmic implementations developed for narrowly defined scientific problems, scientific software has evolved into sophisticated and often community-driven ecosystems. Despite these advances, persistent challenges remain in aligning rapidly evolving scientific demands with robust, generalizable, and sustainable software solutions.

This Research Topic brings together contributions that reflect the central role of scientific software in contemporary soft matter research. Rather than treating software as a secondary technical component, the articles collected here emphasize its function as an enabling research infrastructure—one that shapes how data is generated, curated, analyzed, and ultimately transformed into knowledge. Across diverse applications and methodological perspectives, the contributions highlight how progress in soft matter research increasingly depends on the integration of sound software engineering practices with domain-specific scientific insight.

Developing transferable coarse-grained (CG) models is a major challenge in molecular simulations, as conventional potentials are often state-point and system composition dependent. While neural network (NN) potentials offer a promising path to transferability, their development is hampered by training instabilities and the complex, often heuristic, selection of model parameters. This work presents a systematic methodology to address these issues, focusing on the rational selection of descriptors for NN-based CG potentials. We propose a systematic workflow for parameterizing Behler–Parrinello symmetry functions (G₂) by analyzing their resolving power and response to changes in system conditions, such as concentration. This allows for the selection of an optimized set of descriptors that captures essential structural features, handles short-range repulsions efficiently, and includes descriptors sensitive to the thermodynamic state to ensure transferability. Furthermore, we adopt a network extension technique that enables iterative improvement of model accuracy by expanding the network architecture and descriptor set without discarding prior training. The methodology is demonstrated on a CG model of methanol–water mixtures, with the NN trained to reproduce radial distribution functions derived from atomistic simulations across a wide range of methanol concentrations (10%–100%). The resulting NN potential, built using the systematic approach, demonstrates significantly improved accuracy and transferability, particularly at high concentrations, outperforming models developed previously with empirically chosen parameters. Our findings provide practical guidelines and a robust workflow for developing accurate and transferable NN potentials for CG simulations, paving the way for more reliable mesoscale modeling of complex systems.

Eukaryotic DNA is packaged in the cell nucleus into chromatin, composed of arrays of DNA-histone protein octamer complexes, the nucleosomes. Over the past decade, it has become clear that chromatin structure in vivo is not a hierarchy of well-organized folded nucleosome fibers but displays considerable conformational variability and heterogeneity. In vitro and in vivo studies, as well as computational modeling, have revealed that attractive nucleosome-nucleosome interaction with an essential role of nucleosome stacking defines chromatin compaction. The internal structure of compacted nucleosome arrays is regulated by the flexible and dynamic histone N-terminal tails. Since DNA is a highly negatively charged polyelectrolyte, electrostatic forces make a decisive contribution to chromatin formation and require the histones, particularly histone tails, to carry a significant positive charge. This also results in an essential role of mobile cations of the cytoplasm (K+, Na+, Mg2+) in regulating electrostatic interactions. Building on a previously successfully established bottom-up coarse-grained (CG) nucleosome model, we have developed a CG nucleosome array (chromatin fiber) model with the explicit presence of mobile ions and studied its conformational variability as a function of Na+ and Mg2+ ion concentration. With progressively elevated ion concentrations, we identified four main conformational states of nucleosome arrays characterized as extended, flexible, nucleosome-clutched, and globular fibers.

The double-helical DNA of large eukaryotic genomes is tightly compacted within the tiny cell nucleus as a DNA–protein complex, chromatin. The universal elements of chromatin, nucleosome core particles (NCPs, 147 base pairs of DNA wrapped around an octamer of histone proteins), are connected by linker DNA of variable lengths into nucleosome arrays, which fold into various and dynamic higher-order structures. Since DNA is a highly negatively charged polyelectrolyte, electrostatic interactions of DNA with positively charged histones, other charged nuclear proteins, as well as with monovalent and multivalent cations, contributes decisively to the formation and folding of nucleosome arrays. The dimensions and timescales of cellular chromatin states and transformations necessitate a multiscale coarse-graining (CG) approach to understand their properties through computational modeling. In this review, we highlight the importance of electrostatics for NCP interactions and nucleosome fiber folding in vitro and in vivo, and argue that the inclusion of explicit ions is indispensable for accurate CG modeling of chromatin structure and dynamics. A summary of the existing CG mapping and force field setups is provided. A brief account of CG modeling studies in which salt dependency is approximated by the Debye–Hückel treatment is given. The primary focus is on the presentation of results from papers that include explicit monovalent and multivalent ionic species in CG simulations of nucleosomes and nucleosome arrays. Finally, we underline perspectives and challenges for future multiscale computational modeling of chromatin.

Generating hydrogen peroxide (H₂O₂) within aqueous or eco-friendly solvent systems presents significant challenges due to the complex reaction dynamics and the need for highly selective and stable catalysts. Herein, we report the production of H₂O₂ with an exceeding rate of one millimole per liter per hour in the absence of catalysts, achieved by agitating aerated ethanol-aqueous solutions with ultrasound. This result is attributed to the water charge transfer, which induces charged water molecules to react with dissolved oxygen and ethanol, respectively. In addition, the diffusion of the superoxide radical is faster in ethanol aqueous solutions than in pure water or in ethanol alone, contributing to the high rate of H₂O₂ generation. Our technology provides new insights into sonochemistry and establishes a green synthetic system for H₂O₂ production.

The solubility of molecular crystals is of interest in many areas of chemistry, of which pharmaceutical applications are of particular importance. Predicting solubility using atomistic first-principle methods, that compare the chemical potential of the solid and the solvated phase, has become more common, but remains challenging due to the difficulty in modeling both the crystal form, including its polymorphs, and the interactions with the solvent. Here we aim to compute the solubilities of three active pharmaceutical ingredients of increasing size and complexity: paracetamol, carbamazepine and indomethacin. The known anhydrous polymorphs of each compound are considered in the calculation of the free energy of the solid form and the solvation is explored both in water and in some organic solvents (ethanol, methanol and acetonitrile). The compounds are categorized as poorly soluble or well-soluble based on the comparison of the solid form free energy and the solvation free energy of a single molecule at infinite dilution. Poor solubility then prompts the use of a solubility estimation based on excess free energies, while for well-soluble molecules their chemical potential has to be calculated as a function of concentration. This is done using the recently developed S⁰ method. While promising results are obtained for paracetamol, carbamazepine and indomethacin predictions systematically underestimate the solubility. This can be ascribed to incomplete descriptions of intermolecular interactions by the force field, but the order of stability of the solid forms and systematic nature of the deviations point toward additional problems originating in the accuracy of the method used to calculate the solid free energies.

Hydrated anatase (101) titanium dioxide surfaces with oxygen vacancies have been studied using a combination of classical and ab initio molecular dynamics simulations. The reactivity of surface oxygen vacancies was investigated using ab initio calculations, showing that water molecules quickly adsorb to oxygen vacancy sites upon hydration. The oxygen vacancy then quickly reacts with the adsorbed water, forming a protonated bridging oxygen atom at the vacancy site and at a neighboring oxygen bridge. Ab initio simulations also revealed that this occurs via a short-lived hydronium ion intermediate. It was investigated how this reaction affects the structure and dynamics of water near the anatase surface. Classical molecular dynamics simulations of surfaces with and without oxygen vacancies showed that vacancies disrupt the second solvation shell, consisting of water molecules hydrogen bonded to the surface, thereby changing the local water density and diffusion as well as the binding modes for hydrogen bonding. Our findings support the hydroxylation of oxygen vacancies on anatase (101) surfaces, rather than stabilization by molecular adsorption or subsurface diffusion. The work gives new atomistic insight into water structure and surface chemistry on the catalytically relevant anatase (101) titanium dioxide surface.

Titanium binding peptides are useful tools for material functionalization in both biomedical and nanotechnology applications because of their ability to attach selectively to titanium surfaces. In this work, we investigate the adsorption behavior of a series of 360 six amino acids long peptides obtained by permutations of titanium binding peptide residues, RKLPDA, on hydroxylated anatase TiO₂ (101) surfaces using extensive atomistic molecular dynamics (MD) simulations, with the purpose identifying sequences with stronger adsorption affinity to titanium. Our results show that small changes in amino acid order can significantly affect both binding strength and structural conformations. Peptides with arginine at the N-terminus and lysine or aspartic acid near the C-terminus tended to exhibit more stable adsorption. The clustering and radial distribution function (RDF) analyzes revealed different binding modes and key atomic interactions, with nitrogen-containing groups and, in some cases, Na⁺ ions playing a significant role in the anchoring of peptides to the surface. These findings suggest a detailed sequence-level understanding of peptide-TiO₂ interactions and can guide the design of improved peptides for titanium functionalization.

We report here an experimental-computational study of hydrated TiO₂ anatase nanoparticles interacting with glycine, where we obtain quantitative agreement of the measured adsorption free energies. Ab initio simulations are performed within the tight binding and density functional theory in combination with enhanced free-energy sampling techniques, which exploit the thermodynamic integration of the unbiased mean forces collected on-the-fly along the molecular dynamics trajectories. The experiments adopt a new and efficient setup for electrochemical impedance spectroscopy measurements based on portable screen-printed gold electrodes, which allows fast and in situ signal assessment. The measured adsorption free energy is −30 kJ/mol (both from experiment and calculation), with preferential interaction of the charged  group which strongly adsorbs on the TiO₂ bridging oxygens. This highlights the importance of the terminal amino groups in the adsorption mechanism of amino acids on hydrated metal oxides. The excellent agreement between computation and experiment for this amino acid opens the doors to the exploration of the interaction free energies for other moderately complex bionano systems.

Adsorption free energies of 32 small biomolecules (amino acids side chains, fragments of lipids, and sugar molecules) on 33 different nanomaterials, computed by the molecular dynamics - metadynamics methodology, have been analyzed using statistical machine learning approaches. Multiple unsupervised learning algorithms (principal component analysis, agglomerative clustering, and K-means) as well as supervised linear and nonlinear regression algorithms (linear regression, AdaBoost ensemble learning, artificial neural network) have been applied. As a result, a small set of biomolecules has been identified, knowledge of adsorption free energies of which to a specific nanomaterial can be used to predict, within the developed machine learning model, adsorption free energies of other biomolecules. Furthermore, the methodology of grouping of nanomaterials according to their interactions with biomolecules has been presented.