Computation and programmability at the nano-bio interface

Abstract

The manipulation of physical reality on the molecular level and construction of devices operating on the nanoscale has been the focal point of nanotechnology. In particular, nanotechnology based on DNA and RNA has a potential to nd applications in the eld of Synthetic Biology thanks to the inherent compatibility of nucleic acids with biological systems. Sca olded DNA origami, proposed by P. Rothemund, is one of the leading and most successful methods in which nanostructures are realised through rational programming of short 'staple' oligomers which fold a long single-stranded DNA called the 'sca old' strand into a variety of desired shapes. DNA origami already has many applications; including intelligent drug delivery, miniaturisation of logic circuits and computation in vivo. However, one of the factors that are limiting the complexity, applicability and scalability of this approach is the source of the sca old which commonly originates from viruses or phages. Furthermore, developing a robust and orthogonal interface between DNA nanotechnology and biological parts remains a signi cant challenge. The rst part of this thesis tackles these issues by challenging the fundamental as- sumption in the eld, namely that a viral sequence is to be used as the DNA origami sca old. A method is introduced for de novo generation of long synthetic sequences based on De Bruijn sequence, which has been previously proposed in combinatorics. The thesis presents a collection of algorithms which allow the construction of custom- made sequences that are uniquely addressable and biologically orthogonal (i.e. they do not code for any known biological function). Synthetic sca olds generated by these algorithms are computationally analysed and compared with their natural counter- parts with respect to: repetition in sequence, secondary structure and thermodynamic addressability. This also aids the design of wet lab experiments pursuing justi cation and veri cation of this novel approach by empirical evidence. The second part of this thesis discusses the possibility of applying evolutionary op- timisation to synthetic DNA sequences under constraints dictated by the biological interface. A multi-strand system is introduced based on an alternative approach to DNA self-assembly, which relies on strand-displacement cascades, for molecular data storage. The thesis demonstrates how a genetic algorithm can be used to generate viable solutions to this sequence optimisation problem which favours the target self- assembly con guration. Additionally, the kinetics of strand-displacement reactions are analysed with existing coarse-grained DNA models (oxDNA). This thesis is motivated by the application of scienti c computing to problems which lie on the boundary of Computer Science and the elds of DNA Nanotechnology, DNA Computing and Synthetic Biology, and thus I endeavour to the best of my ability to establish this work within the context of these disciplines.