AI Designs Intricate DNA Nanostructures Using Origami Technique

Researchers from Seoul National University and Hanyang University have developed a generative AI model named Generative SNUPI to create complex DNA origami structures. This AI significantly streamlines the previously tedious process of designing DNA strands that fold into specific three-dimensional shapes, such as the Mona Lisa or dog faces, at the nanoscale. Unlike traditional methods requiring extensive manual expertise and algorithm tweaking, Generative SNUPI can generate functional DNA designs directly from a target shape by considering the chemical rules of DNA folding. The AI utilizes a diffusion model, similar to those used in image generation platforms like DALL-E, to translate a desired shape into a DNA sequence. The resulting DNA sequences, when synthesized, self-assemble into the intended nanostructures through the binding of staple strands to a scaffold strand, guided by DNA's base-pairing rules. While the technique has been around for two decades with potential applications in nanorobotics and therapeutics, the design process has been a bottleneck due to its time and cost. The new AI platform accelerates this by automating the design phase, allowing users to move from concept to physical assembly more rapidly. Early designs sometimes failed to hold their shape, but the team added a step to predict structural stability. Future research aims to enable the design of more flexible and dynamically reconfigurable DNA structures, crucial for advanced applications like drug delivery and immunotherapy.