Synergistic approach to probing the dynamics and mechanics of patchy soft matter

Md Mozakker H. Shojib †, Asier C. Monasterio †, Emanuele Locatelli, Pascal Friederich, Christopher Ness, Iliya D. Stoev* · 2026

Read Synergistic approach to probing the dynamics and mechanics of patchy soft matter by Md Mozakker H. Shojib †, Asier C. Monasterio †, Emanuele Locatelli, Pascal Friederich, Christopher Ness, Iliya D. Stoev* (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-29967.txt.

Introduction

Md Mozakker H. Shojib and five colleagues blend simulation, lab rheology, and machine learning to map tiny tweaks steering flow in soft matter. Their test case is DNA-based fluids from nucleic acid nanostars, whose response shifts with base sequence. Coarse simulations, checked against lab data, feed a Gaussian process plus active learning loop roaming the design space with forecasts. The loop runs again and again, guiding design and faster find of soft suspensions. Three lanes together thus tame a parameter space too wide for any one lane alone.

Catalog record

CREATORMd Mozakker H. Shojib †, Asier C. Monasterio †, Emanuele Locatelli, Pascal Friederich, Christopher Ness, Iliya D. Stoev*
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54050
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.29967

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Synergistic approach to probing the dynamics and mechanics of patchy soft matter Md Mozakker H. Shojib †, Asier C. Monasterio †, Emanuele Locatelli, Pascal Friederich, Christopher Ness, Iliya D. Stoev* Abstract Tailoring microscopic details to tune bulk rheology is a key paradigm in soft matter physics, yet the vast parameter space associated with constituent interactions precludes a fully systematic approach. To address this, we have designed a synergistic strategy to explore the parameter space that comprises simulations, experimental rheology, and machine learning. As a case study, we choose DNA-based self-assembled fluids whose viscoelastic response can be fine-tuned by manipulating the base sequencing of the constituent nucleic acid nanostars. We use coarse-grained simulations, benchmarked against experimental data, to obtain the rheology of the DNA fluids, which feeds forward to a framework of Gaussian Process Regression and active learning. The latter is then used to explore the rheological design space with high predictive precision. The pipeline is designed to be deployed iteratively for the rational design and accelerated discovery of generic soft matter suspensions. Keywords bead-spring model, structure-property relations, material properties, machine learning, in-silico rheology 1 Introduction A central goal in soft condensed matter is to establish quantitative structure-property relationships that link microscopic design variables to macroscopic material performance. This challenge is particularly prominent in self-assembling colloidal suspensions and polymeric networks; a well-known and experimentally accessible platform within this broader class are DNA-based bulk fluids. Here, sequence-encoded complementarity follows strict Watson-Crick base-pairing rules and enables highly selective, reversible bonding. This sequence programmability allows fine-tuning of the geometry and valency of DNA nanomotifs, which opens new opportunities for controlling gel formation, designing responsive architectures and even constructing information-carrying materials [ 8 , 3 , 6 , 7 , 21 , 5 , 10 , 16 , 1 , 4 , 19 ] . Beyond finding targeted applications in data storage and programmable matter, these systems serve as foundational model patchy networks in which multi-valency and tunable bond lifetimes can be effectively isolated and studied systematically [ 34 , 29 ] . In general, the presence of reversible bonds and finite valency leads to a viscoelastic response that depends strongly on connectivity and relaxation dynamics [ 23 , 29 , 26 , 34 ] . Developing predictive models that capture how interaction strength, particle geometry, stiffness and number density control the rheological fingerprint is essential for the rational design of soft materials with tailored thermodynamic and mechanical properties [ 10 ] . An additional challenge is that macroscopic observables, such as melting curves and linear viscoelastic moduli, emerge from many-body assembly pathways that are difficult to resolve or access directly in experiments. This motivates the wide use of computational models that connect microscopic design choices to bulk behavior at the scale of tens to thousands of building blocks. High-resolution nucleotide-level models ( e.g. , oxDNA) are powerful but are computationally too demanding for systematic studies of network formation and mechanics in large volumes over long times [ 34 , 24 , 25 ] . Higher-level coarse-grained descriptions, such as based on a bead-spring model, retain only the minimal physical ingredients required for recapitulating experimental bulk behavior, thereby offering generality and universality of the obtained insights [ 27 ] . While coarse-grained simulations accelerate design predictions, an important question is whether such simplified descriptions can accurately capture experimentally observed behavior in actual DNA hydrogels. For example, experimental systems composed of Y-shaped nanostars connected by complementary linkers exhibit characteristic viscoelastic spectra and tunable mechanical responses that depend on sticky-end binding strength and structural flexibility [ 5 ] . Therefore, establishing a correspondence between simulation parameters and experimentally measurable rheological properties is crucial to boost the predictive power of the model, allowing it to guide experimental effort and accelerate material design. In this work, we introduce an efficient approach for the exploration of the parameter space of self-assembling fluids and, in general, soft matter systems, focusing specifically on the case of DNA hydrogels. This approach is based on the synergistic combination of coarse-grained modeling, Brownian dynamics simulations and machine learning (ML), with final validation against experimental data. First, a suitable computational model should be identified. Here, we choose to adapt a minimal bead-spring and patchy-particle description of DNA-based Y-shaped building blocks; this model has already been employed to recapitulate the elasticity of all-DNA networks [ 34 ] . We incorporate here the building blocks’ internal flexibility to enable extensions to mixed-valency architectures and systematic exploration of interaction strength, stiffness, number density and composition. To accelerate exploration of the large design space and reduce the computational cost of exhaustive parameter scans, we combine simulation outputs with machine-learning surrogate models, specifically Gaussian Process Regression (GPR), capable of interpolating and predicting rheological observables across non-sampled regions [ 10 , 13 , 35 ] . By comparing predicted results to experimental data, we demonstrate the capacity of the method to identify the appropriate region of the parameter space, where the model yields rheological curves that qualitatively match the experimental ground truth. Overall, our goal is to provide a scalable and transferable modeling workflow that links motif-level geometry and interaction rules to emergent bulk structure and mechanics, thereby facilitating rational design of attractive soft-matter networks with tailored functionality. 2 Model and methods We discuss here in some detail the three elements that comprise our synergistic approach. To give more context, we also introduce the experimental system under investigation, serving as a case study. We consider a binary mixture of DNA nanomotifs, comprised of Y-shaped nanostars and linear linker oligonucleotides [ 10 ] . By design, both linker and Y-shaped molecules have dangling ends, made of single-stranded DNA; sequence complementarity allows the hybridization of two dangling ends, commonly referred to as “sticky” ends, between a linker and a Y-shape only. Following the terminology used in patchy systems, the Y-shapes have a valency of 3 and the linkers have a valency of 2; this design enables selective Y-linker binding, leading to the formation of a network with specific architecture. Further, the valency asymmetry defines a geometric stoichiometric optimum, i.e. , a Y:L (Y-shape-to-linker ratio) of 2:3 ( \mathrm{ratio}=0.667 ), ensuring maximum complementary pairing [ 32 ] . 2.1 Coarse-grained model We employ a minimal coarse-grained bead-spring framework to investigate the assembly, melting and viscoelastic response of mixed-valency nanostar-linker networks. Following Ref. [ 33 ] , each unit is represented as a set of “structural” and “patchy” beads, where the latter model the DNA sticky ends. The units of length and energy of the model are \sigma , the excluded volume size of a bead, and the thermal energy k_{B}T , respectively. Y-shaped units are composed by a central bead connected to three arms; each arm is formed by Y_{\mathrm{arm}} beads ( Y_{\mathrm{arm}} = 2 or 3 in this study) and terminates with a single patch. Linkers are instead modeled as linear chains of 3 beads with patches at both termini ( cf. , Figure 1 A). Intra-molecular connections, involving structural and patch beads that define the building-block geometry, are modeled as harmonic interactions: V_{\mathrm{bond}}(r)=K_{\mathrm{bond}}(r-r_{0})^{2}, (1) where r is the distance between two bonded monomers, r_{0}=0.96\sigma is the equilibrium bond length and \beta K_{\mathrm{bond}}=1000 , with \beta=1/(k_{B}T) , is the bond stiffness [ 34 ]

Read the complete text