Artificial intelligence strategies for optimisation of Nucleic Acid Origami experiments

Abstract

Nanotechnology based on DNA and RNA as building blocks has been advancing with an increasing number of applications using devices at the nanoscale, finding use cases across many fields of research such as Photonics, Molecular Payloads and Enzymatic Reactions. There is a large potential in the field of Synthetic Biology to apply such structures due to the use of nucleic acid building blocks to create customised compatible structures. The use of Scaffolded Nucleic Acid Origami is a relatively cheap and fast method of producing a "scaffold" strand that combines with shorter sequence oligomer "staples" to form nanostructures, rationally programming these to produce designed structures. The structures produced are DNA/RNA origami folded nanostructures enabled by the addressability and complementary pairing of bases in oligonucleotide strands. The Scaffolded Nucleic Acid Origami technique has shown promise but still has challenges for widespread use. These include the yield of nucleic acid origami nanostructures, the scalability of the technique and the complexity of the Nucleic Acid Origami structures created. This thesis aims to address these challenges in three parts. First, we curate a database of nucleic acid origami structures from literature to gain insights and store information for future projects. Second, we use machine learning to discover patterns and design suitable lab protocols for origami production. Finally, we create a multi-objective optimisation algorithm to generate optimal scaffold and staple sequences for improved yield of nucleic acid origami nanostructures