Titolo della tesi: Understanding the thermodynamics and kinetics of DNA- and RNA-based self-assembly with coarse-grained models
The self-assembly of nucleic acids, particularly DNA and RNA, plays a fundamental role in nanotechnology, synthetic biology, and biomolecular engineering. Understanding the thermodynamic and kinetic principles governing these processes is crucial for the rational design of functional nanostructures. This thesis investigates nucleic acid-based self-assembly using coarse-grained modeling, with a particular focus on phase separation phenomena, sequence-dependent interactions, and emergent structural motifs.
As part of my doctoral training, I was required to undertake an industrial research internship, reinforcing the applied dimension of my work and strengthening the link between theoretical modeling and practical applications. This requirement motivated me to focus on computational approaches that are not only relevant for fundamental research but also applicable in industrial settings. In this context, we explored the design, simulation, and analysis of DNA- and RNA-based nanostructures, leveraging coarse-grained molecular dynamics models to investigate their stability, folding pathways, and kinetic properties.
One of the key challenges in modeling nucleic acid self-assembly is the trade-off between resolution and computational feasibility. All-atom molecular dynamics (MD) simulations provide atomic-level precision and detailed energetic insights, making them invaluable for studying local interactions, conformational flexibility, and solvation effects. However, their high computational cost severely limits their ability to capture long-timescale processes and large system sizes, particularly in complex nanostructures. Coarse-grained models, by contrast, reduce molecular complexity while retaining essential biophysical properties, enabling the study of larger systems over timescales that would otherwise be inaccessible.
A central contribution of this thesis is the development and application of ANNaMo, a novel coarse-grained model designed to capture essential thermodynamic and kinetic features of DNA and RNA self-assembly. By parameterizing ANNaMo against melting temperature data and benchmarking it against oxDNA, we validated its predictive accuracy in reproducing key structural and energetic properties. The model was then used to study the formation and stability of nucleic acid motifs such as DNA hairpins, RNA pseudoknots, and double helices, as well as toehold-mediated strand displacement (TMSD), a fundamental reaction in DNA computing and dynamic nanotechnology. Additionally, we investigated entropy-driven phase separation in transient polymer networks, drawing analogies between DNA-based systems and synthetic associative polymers, revealing how multivalent interactions and sequence programmability influence phase behavior.
To complement my theoretical research, I carried out an industrial internship at Nanogami GmbH, where I applied computational methods to problems relevant to DNA nanotechnology and molecular design. My work focused on three main areas:
1- Developing a plugin for caDNAno that streamlines the conversion of DNA origami designs into oxDNA-compatible simulation files, facilitating more efficient computational workflows.
2- Performing large-scale oxDNA simulations to analyze the structural stability of DNA origami constructs and provide theoretical predictions that could assist in design optimization.
3- Implementing a convolutional neural network (CNN) for automated detection of DNA origami in atomic force microscopy (AFM) images, improving the efficiency of structural analysis in nanostructure characterization.
This experience broadened my expertise in computational DNA nanotechnology and demonstrated the importance of industry-academia collaboration, where simulations and theoretical models contribute to advancing molecular design strategies.
The findings of this thesis contribute to both fundamental and applied aspects of nucleic acid self-assembly. The development of ANNaMo represents a significant step forward in the coarse-grained simulation of DNA and RNA, enabling the study of structural transitions and phase behavior at a level of detail not previously accessible.
The future directions emerging from this work include refining ANNaMo by improving its representation of backbone flexibility to better capture structural constraints in nucleic acid folding, addressing discrepancies in bonding behavior in phase separation models, and further optimizing the computational
efficiency of oxDNA simulations for large systems. Additionally, extending ANNaMo to more complex multi-stranded systems, such as DNA origami, could open new possibilities for studying folding pathways and optimizing structural designs. These advancements would enable a more comprehensive understanding of DNA-based self-assembly, with potential applications in nanotechnology, synthetic biology, and biomolecular engineering.