Periodic Reporting for period 4 - DYNAPOL (Modeling approaches toward bioinspired dynamic materials)
Période du rapport: 2024-05-01 au 2025-10-31
Nature uses the principle of self-assembly to build materials with fascinating properties, such as the ability to self-heal, dynamically adapt, reconfigure, or respond to specific stimuli in controlled way. Microtubules, protein filaments, cellular membranes, for example, are supramolecular polymers made of fundamental molecular building-blocks, or monomers (e.g. proteins, lipids, etc.), which recognize each other and self-assemble into ordered structures in determined conditions with unique fidelity and precision. The dynamic character of such assemblies makes it possible to embed them in complex systems where they are in continuous communication with the external environment, expressing complex functionalities exceeding those of traditional technological materials such as, e.g. metals or covalent polymers.
This ERC project “Modeling Approaches Toward Bioinspired Dynamic Materials (DYNAPOL)” aimed at answering key, fundamental questions, such as:
- Is it possible to impart similar bioinspired behaviors into synthetic materials build via self-assembly?
- And if yes, how?
The ability to conceive and design new types of materials has determined the development of humanity from the stone, bronze and iron ages to our current world, dominated by electronic materials and semiconductors. The DYNAPOL project explored new routes and fundamental principles to design new types of artificial materials with fascinating bioinspired properties, reminiscent to those of living systems. The obtained results span from new fundamental knowledge in materials science to new paradigms and advances in many fields, from biomedicine, to the chemical industry, to the design of new types of materials with advanced properties.
The DYNAPOL project combined multiscale molecular modeling, advanced computer simulations, advanced data analysis and machine learning to learn how to control the bioinspired dynamic properties of supramolecular assemblies. Multiscale molecular models allowed studying the supramolecular structure of a large variety of different self-assembled materials on multiple scales (Objective 1). Advanced computational simulation approaches made it possible to study the intrinsic dynamics of these materials at very high (submolecular) resolution, unveiling the key factors controlling it (Objective 2). In silico experiments allowed to study at high resolution bioinspired properties such as e.g. the ability of different types of supramolecular materials to adapt or reconfigure dynamically in response to specific stimuli, understanding the origin of their responsive behaviors (Objective 3). Machine learning approaches finally allowed us to extract crucial physical/chemical information from the large amount of data extracted from the simulations, identifying key features in the monomers’ structure, interactions and dynamical communication networks controlling higher-scale structure, dynamics, and properties of the supramolecular materials that these generate (Objective 4).
Overall, this research produced unprecedented insights and fundamental models for designing new types of artificial materials with controllable dynamic bioinspired properties.
This ERC project “Modeling Approaches Toward Bioinspired Dynamic Materials (DYNAPOL)” aimed at answering key, fundamental questions, such as:
- Is it possible to impart similar bioinspired behaviors into synthetic materials build via self-assembly?
- And if yes, how?
The ability to conceive and design new types of materials has determined the development of humanity from the stone, bronze and iron ages to our current world, dominated by electronic materials and semiconductors. The DYNAPOL project explored new routes and fundamental principles to design new types of artificial materials with fascinating bioinspired properties, reminiscent to those of living systems. The obtained results span from new fundamental knowledge in materials science to new paradigms and advances in many fields, from biomedicine, to the chemical industry, to the design of new types of materials with advanced properties.
The DYNAPOL project combined multiscale molecular modeling, advanced computer simulations, advanced data analysis and machine learning to learn how to control the bioinspired dynamic properties of supramolecular assemblies. Multiscale molecular models allowed studying the supramolecular structure of a large variety of different self-assembled materials on multiple scales (Objective 1). Advanced computational simulation approaches made it possible to study the intrinsic dynamics of these materials at very high (submolecular) resolution, unveiling the key factors controlling it (Objective 2). In silico experiments allowed to study at high resolution bioinspired properties such as e.g. the ability of different types of supramolecular materials to adapt or reconfigure dynamically in response to specific stimuli, understanding the origin of their responsive behaviors (Objective 3). Machine learning approaches finally allowed us to extract crucial physical/chemical information from the large amount of data extracted from the simulations, identifying key features in the monomers’ structure, interactions and dynamical communication networks controlling higher-scale structure, dynamics, and properties of the supramolecular materials that these generate (Objective 4).
Overall, this research produced unprecedented insights and fundamental models for designing new types of artificial materials with controllable dynamic bioinspired properties.
The main results achieved in the DYNAPOL project are summarized here:
OBJ 1:
• Multiscale simulations of a variety of supramolecular systems, supramolecular polymers, micelles, catenanes, superlattices, host–guest systems, etc. (Nature 2020, 583, 400; Nat. Chem. 2021, 13, 940; Nat. Chem. 2022, 14, 507; Nat. Commun. 2022, 13, 248)
• Development of SwarmCG, a software for automatic optimization of molecular models & force fields to fit simulation & experimental observables (ACS Omega 2020, 5, 32823; JCP 2022, 156, 024801; JCIM 2023, 63, 3827)
OBJ 2:
• Submolecular resolution study of monomer exchange & dynamical rearrangements in various supramolecular systems: from self-assembled materials (ACS Nano 2021, 15, 14229; JACS 2020, 142, 7606), to rotaxanes, shuttles, catenanes (Chem. Sci. 2023, 14, 6716; ACIE 2023, 62, e202309393)
• Study of supramolecular systems as complex systems & characterization of their internal dynamical networks (Nat. Commun. 2022, 13, 2162; Nat. Commun. 2025, 16, 5030)
OBJ 3:
• Simulation of stimuli-responsive host–guest systems, studying the effect of crowding & dynamics on their responsiveness (JACS 2020, 142, 9792; Chem. Sci. 2022, 13, 11232)
• Modeling of bioinspired assemblies with programmable autonomous properties (ACS Nano 2021, 15, 16149)
• Simulations of complex responsive assemblies revealing how structural/dynamical heterogeneities lead to non-trivial emergent properties (ACS Nano 2023, 17, 275; Nat. Commun. 2025, 16, 5030)
OBJ 4:
• Classification of supramolecular assemblies based on their internal order/disorder via machine learning (ML) & abstract data-driven metrics (Commun. Chem. 2022, 5, 82; JPCB 2020, 124, 589)
• Development of descriptors & ML methods (LENS, timeSOAP, Onion) to detect defects, track fluctuations in dynamical self-organizing systems (PNAS 2023, 120, e2300565120; JCP 2023, 158, 214302; PNAS 2024, 121, e2403771121; PNAS Nexus 2025, 4, pgaf038) and quantify the information contained in them (MINE: MLST 2025, DOI:10.1088/2632-2153/ae2dbb; JCP 2025, 162, 234110)
• Development of DYNSIGHT: a software for advanced analysis in complex dynamical systems (arXiv 2025, 2510.23493)
The results of DYNAPOL have been exploited & disseminated via (i) high-impact publications in important scientific journals; (ii) release of open & documented software (SwarmCG, DYNSIGHT) available to a broad community; (iii) scientific collaborations with leading groups in self-assembly & supramolecular chemistry; (iv) invited talk at conferences & important institutions, further consolidated by the PI election as Chair of the next 2027 Gordon Research Conference on “Self-assembly and Supramolecular Chemistry”
OBJ 1:
• Multiscale simulations of a variety of supramolecular systems, supramolecular polymers, micelles, catenanes, superlattices, host–guest systems, etc. (Nature 2020, 583, 400; Nat. Chem. 2021, 13, 940; Nat. Chem. 2022, 14, 507; Nat. Commun. 2022, 13, 248)
• Development of SwarmCG, a software for automatic optimization of molecular models & force fields to fit simulation & experimental observables (ACS Omega 2020, 5, 32823; JCP 2022, 156, 024801; JCIM 2023, 63, 3827)
OBJ 2:
• Submolecular resolution study of monomer exchange & dynamical rearrangements in various supramolecular systems: from self-assembled materials (ACS Nano 2021, 15, 14229; JACS 2020, 142, 7606), to rotaxanes, shuttles, catenanes (Chem. Sci. 2023, 14, 6716; ACIE 2023, 62, e202309393)
• Study of supramolecular systems as complex systems & characterization of their internal dynamical networks (Nat. Commun. 2022, 13, 2162; Nat. Commun. 2025, 16, 5030)
OBJ 3:
• Simulation of stimuli-responsive host–guest systems, studying the effect of crowding & dynamics on their responsiveness (JACS 2020, 142, 9792; Chem. Sci. 2022, 13, 11232)
• Modeling of bioinspired assemblies with programmable autonomous properties (ACS Nano 2021, 15, 16149)
• Simulations of complex responsive assemblies revealing how structural/dynamical heterogeneities lead to non-trivial emergent properties (ACS Nano 2023, 17, 275; Nat. Commun. 2025, 16, 5030)
OBJ 4:
• Classification of supramolecular assemblies based on their internal order/disorder via machine learning (ML) & abstract data-driven metrics (Commun. Chem. 2022, 5, 82; JPCB 2020, 124, 589)
• Development of descriptors & ML methods (LENS, timeSOAP, Onion) to detect defects, track fluctuations in dynamical self-organizing systems (PNAS 2023, 120, e2300565120; JCP 2023, 158, 214302; PNAS 2024, 121, e2403771121; PNAS Nexus 2025, 4, pgaf038) and quantify the information contained in them (MINE: MLST 2025, DOI:10.1088/2632-2153/ae2dbb; JCP 2025, 162, 234110)
• Development of DYNSIGHT: a software for advanced analysis in complex dynamical systems (arXiv 2025, 2510.23493)
The results of DYNAPOL have been exploited & disseminated via (i) high-impact publications in important scientific journals; (ii) release of open & documented software (SwarmCG, DYNSIGHT) available to a broad community; (iii) scientific collaborations with leading groups in self-assembly & supramolecular chemistry; (iv) invited talk at conferences & important institutions, further consolidated by the PI election as Chair of the next 2027 Gordon Research Conference on “Self-assembly and Supramolecular Chemistry”
The beyond-the-state-of-the-art character of the work performed in DYNAPOL is summarized below:
OBJ 1: The multiscale molecular simulations conducted in DYNAPOL allowed to study at high-resolution supramolecular systems of unprecedented complexity, providing data useful to rationalize their behaviors. The development of SwarmCG, an increasingly used algorithm to optimize molecular models & force fields, went even beyond the originally planned results.
OBJ 2: The use of advanced simulations for studying dynamics at submolecular resolution in supramolecular materials, rotaxanes & catenanes provided unprecedented details of key factors that control dynamics in self-assembled systems. The study of self-assembling systems as complex systems opened new perspectives for understanding their behavior & properties, going beyond what originally planned in the project.
OBJ 3: The high-resolution study of stimuli-responsive behaviors in a variety of complex supramolecular systems provided deep details of how adaptive and reactive properties emerge in them. This provided unprecedented insights on how to program autonomous dynamical/responsive behaviors in supramolecular assemblies.
OBJ 4: The descriptors, analysis & ML tools developed in the project provided new tools to detect defects & classify/correlate dominant fluctuations in self-organizing systems. The generality of these methods (e.g. LENS, ONION, MINE, DYNSIGHT) makes them applicable to virtually any system, opening new unprecedented ways for studying self-organizing systems & learning how to control their properties.
OBJ 1: The multiscale molecular simulations conducted in DYNAPOL allowed to study at high-resolution supramolecular systems of unprecedented complexity, providing data useful to rationalize their behaviors. The development of SwarmCG, an increasingly used algorithm to optimize molecular models & force fields, went even beyond the originally planned results.
OBJ 2: The use of advanced simulations for studying dynamics at submolecular resolution in supramolecular materials, rotaxanes & catenanes provided unprecedented details of key factors that control dynamics in self-assembled systems. The study of self-assembling systems as complex systems opened new perspectives for understanding their behavior & properties, going beyond what originally planned in the project.
OBJ 3: The high-resolution study of stimuli-responsive behaviors in a variety of complex supramolecular systems provided deep details of how adaptive and reactive properties emerge in them. This provided unprecedented insights on how to program autonomous dynamical/responsive behaviors in supramolecular assemblies.
OBJ 4: The descriptors, analysis & ML tools developed in the project provided new tools to detect defects & classify/correlate dominant fluctuations in self-organizing systems. The generality of these methods (e.g. LENS, ONION, MINE, DYNSIGHT) makes them applicable to virtually any system, opening new unprecedented ways for studying self-organizing systems & learning how to control their properties.