Abstract:Inspired by biology's most sophisticated computer, the brain, neural networks constitute a profound reformulation of computational principles. Remarkably, analogous high-dimensional, highly-interconnected computational architectures also arise within information-processing molecular systems inside living cells, such as signal transduction cascades and genetic regulatory networks. Might neuromorphic collective modes be found more broadly in other physical and chemical processes, even those that ostensibly play non-information-processing roles such as protein synthesis, metabolism, or structural self-assembly? Here we examine nucleation during self-assembly of multicomponent structures, showing that high-dimensional patterns of concentrations can be discriminated and classified in a manner similar to neural network computation. Specifically, we design a set of 917 DNA tiles that can self-assemble in three alternative ways such that competitive nucleation depends sensitively on the extent of co-localization of high-concentration tiles within the three structures. The system was trained in-silico to classify a set of 18 grayscale 30 x 30 pixel images into three categories. Experimentally, fluorescence and atomic force microscopy monitoring during and after a 150-hour anneal established that all trained images were correctly classified, while a test set of image variations probed the robustness of the results. While slow compared to prior biochemical neural networks, our approach is surprisingly compact, robust, and scalable. This success suggests that ubiquitous physical phenomena, such as nucleation, may hold powerful information processing capabilities when scaled up as high-dimensional multicomponent systems.

What problem does this paper attempt to address?

The problem this paper attempts to address is the exploration of pattern recognition capabilities in nucleation dynamics during non-equilibrium self-assembly processes. Specifically, the authors designed a system consisting of 917 DNA tiles that can self-assemble in three different ways to form three distinct structures. In this way, they studied how high-dimensional concentration patterns influence the selectivity of the nucleation process, thereby achieving complex pattern recognition similar to neural network computation. ### Main Research Questions: 1. **Impact of High-Dimensional Concentration Patterns on the Nucleation Process**: Investigate how the nucleation process selects different self-assembly pathways under different concentration patterns. 2. **Pattern Recognition Capability**: Verify whether the system can recognize and classify high-dimensional concentration patterns through nucleation dynamics, similar to the computational capabilities of neural networks. 3. **System Design and Experimental Validation**: Design a system containing 917 DNA tiles and experimentally validate its performance in pattern recognition. ### Research Background: - **Biological Inspiration**: Inspired by the most complex computer in biology—the brain, neural networks have redefined computational principles. - **Similar Architectures in Molecular Systems**: Similar high-dimensional, highly interconnected computational architectures are also found in molecular systems for information processing within living cells, such as signal transduction cascades and gene regulatory networks. - **Collective Patterns in Physical and Chemical Processes**: The authors explore whether similar collective patterns can be found in other physical and chemical processes, even if these processes do not ostensibly involve information processing. ### Experimental Methods: - **DNA Tile Design**: Designed 917 DNA tiles that can self-assemble into three different structures, with the nucleation process of each structure depending on the co-localization of high-concentration tiles. - **Training and Testing**: Trained the system through computer simulations to classify 18 grayscale 30×30 pixel images. Fluorescence and atomic force microscopy measurements were used in experiments to verify the correct classification of all training images and to test the system's robustness. ### Main Findings: - **Pattern Recognition Capability**: The system can recognize and classify high-dimensional concentration patterns through nucleation dynamics, similar to the computational capabilities of neural networks. - **Robustness and Scalability**: Although slower, the method is compact, robust, and scalable. - **Information Processing Capability of Physical Phenomena**: The study suggests that ubiquitous physical phenomena (such as nucleation) may have powerful information processing capabilities in high-dimensional multi-component systems. ### Conclusion: This study demonstrates the pattern recognition capabilities of nucleation dynamics in non-equilibrium self-assembly processes, providing new perspectives for understanding physically embedded computation in biology and offering new tools for designing autonomous molecular systems (such as artificial cells).

Pattern recognition in the nucleation kinetics of non-equilibrium self-assembly

Associative pattern recognition through macro-molecular self-assembly

Programmable Control of Nucleation for Algorithmic Self-Assembly

Neural network computation with DNA strand displacement cascades

Single-molecule Automata: Harnessing Kinetic-Thermodynamic Discrepancy for Temporal Pattern Recognition

Pattern Recognition Computation in a Spiking Neural Network with Temporal Encoding and Learning

Event Tracking: A systematic method for analyzing nucleation and growth in hierarchical self-assembly

A Reusable Non-Complementary-DNA-Based Neural Network

Diverse and robust molecular algorithms using reprogrammable DNA self-assembly

How do particles with complex interactions self-assemble?

Temporal pattern recognition through analog molecular computation

Coordinated Spatial Pattern Formation in Biomolecular Communication Networks

Self-Assembly of a Designed Nucleoprotein Architecture through Multimodal Interactions

The shape of things

Self-assembly of nanocrystal checkerboard patterns via non-specific interactions

Regulating DNA Self-Assembly Dynamics with Controlled Nucleation

Discovery of interpretable patterning rules by integrating mechanistic modeling and deep learning

Atomic Scale Dynamics Drive Brain-like Avalanches in Percolating Nanostructured Networks

Conformational Flexibility Facilitates Self-Assembly Of Complex Dna Nanostructures

Molecular convolutional neural networks with DNA regulatory circuits

Programming molecular systems to emulate a learning spiking neuron