Nanoscale structures can be manufactured using a top-down or bottom-up approach. Top-down techniques fabricate a desired structure by imposing an external source of order (such as specific patterns of light in photolithography). This contrasts with bottom-up approaches, which exploit molecular self-assembly. Particle interactions are programmed so that intended structures form spontaneously. Bottom-up molecular self-assembly promises affordable large-scale manufacture of nanoscale structures with regularity controlled by molecular thermodynamics rather than instrument precision.

DNA nanotechnology is a versatile self-assembly technique. Favourable DNA-DNA interactions are mediated by Watson-Crick base pairs which form only between complementary nucleotides. Mutual affinity and equilibrium structure can be programmed by an appropriate choice of base sequence, analogous to protein tertiary and quaternary structure emerging from linear amino acid chains. DNA strand displacement reactions allow the design of thermodynamic pathways with controllable kinetics. Meanwhile, the DNA origami technique permits the self-assembly of arbitrary three dimensional structures.

Strand displacement pathways have promising applications in clinical and industrial sensors, translating and amplifying signals into a detectable response. Meanwhile, DNA origami have been applied to biophysical measurements, as fiducials for instrument calibration, and as therapeutics for disease. As the field matures, the need for effective models for these devices increases, to enable faster design iteration and to contextualise results. This thesis reports on several novel applications of DNA nanotechnology, where coarse-grained molecular simulation was used either for in silico prototyping, or experimental rationalisation. While each individual project has a relevant application, the underlying goal is to validate and identify the limits of modelling for both nanoscale sensors, and higher-order structures.

Nanoscale sensors convert information about the presence of a molecule to a readout measurable by an instrument. While these synthetic pathways conventionally take place in solution, information relays in cells often occur on membranes, which I have mimicked here, incentivised by molecular modelling that suggested that analyte responsive DNA strand displacement is accelerated by the geometrical restraints of the membrane of a liposome. A similar principle underlies a subsequent project: to design a proximity-sensitive super-resolution microscopy assay, prototyped by coarse-grained simulation.

Just as DNA structure can be programmed from sequence, so can higher order structures be directed from monomeric mutual affinity. I have used molecular simulation to rationalise the flexibility dependent phase of amphiphilic DNA nanostars. Conversely, one of the goals in structural DNA nanotechnology is the manufacture of materials with unusual mechanical properties. Here, I report the formation of a DNA origami with negative Poisson's ratio: when tension is applied in one direction, it expands in the orthogonal direction also.

Throughout this work, the importance of fast, automated approaches to identifying configurational free energies has been highlighted. To accelerate acquisition of free energy landscapes to inform nanostructure design, I have implemented metadynamics in a popular coarse-grained DNA molecular simulation package. I have demonstrated its application in predicting the energetic cost in bending a mechanically compliant DNA origami, and the relative conformer stability of Holliday junctions. It is hoped that this technique will enable rapid in silico prototyping of DNA nanostructures.