On the chemical self-assembly of nucleic acid mixtures

One of the 25 "most compelling puzzles and questions facing scientists" in the next 25 years was put forth in the 125th anniversary issue of the journal Science: "How far can we push chemical self-assembly?" This work characterizes the chemical self-assembly of nucleic acid mixtures in various important contexts. I examined the chemical self-assembly of nucleic acids with the goal of producing a wholly synthetic gene. I provided computational models of multiplex DNA assembly and a distributed computational infrastructure for DNA melting temperature calculations. This work enabled the development of computational and biological methods which produce synthetic genes encoding nearly any natural or unnatural protein constructed from entirely synthetic deoxyoligonucleotides specifically designed for reliable chemical self-assembly. This combined approach established an accurate gene synthesis methodology capable of simplifying numerous experiments in genomic, systems and synthetic biology.I investigated the chemical self-assembly of nucleic acids with the goal of understanding and predicting oligonucleotide directed mutagenesis failures in protein engineering experiments. I computationally identified and experimentally validated off-target assembly, termed cross-hybridization, as a factor which confounds oligonucleotide-directed mutagenesis, especially of large combinatorial libraries. This work established the existence of cross-hybridization problems in mutagenesis and provided the foundation for engineering plasmids and/or altering mixture compositions in order to improve success rates and enhance protein engineering efforts.I engineered the chemical self-assembly of nucleic acids by programming the folding a single-stranded molecule of DNA into a specific shape 3D shape – a mega-dalton DNA nanostructure shaped as an irregular tetrahedron, roughly the size of the ribosome. As the targeted shape is capable of tessellating three-dimensional space, the nanostructures developed in this work are useful as potential a building blocks for future construction of arbitrary three-dimensional DNA-nanostructures.Lastly, I studied chemical self-assembly of nucleic acids in retro-virus development. I describe an efficient computational score for the identification of retroviral genome dimerization sites. The score correlated well with known retroviral dimerization sites. This work lays the foundation for the potential rational exploration and discovery of previously unknown retroviral dimerization sites.