DNA by Design: De novo Computational Framework for DNA Sequence Design and Nanotechnology

Abstract

Chemical analysis of metalized DNA has made it quite clear that traditional models of DNA thermodynamics are insufficient to predict and control self-assembly in the context of orthogonally-paired nucleotides. Recently, there has been an increase in reports of Watson-Crick assembly of DNA wires and nanostructures [1–4]. The ability to add or remove pairing rules between nucleobases toward non-Watson-Crick, or orthogonal, self-assembly alters the fundamental language of DNA assembly: this change in behavior necessitates an accompanying shift in computational design. We begin by exploring the state-of-the-art in DNA modeling, and include both sequence analysis and sequence design practices. We then start from first principles and establish a mathematical basis for heterostructure and ‘nmer’ analysis in connected DNA networks that operates without assumptions about nucleobase parity. A generalized search algorithm is then constructed in Matlab and implemented using evolutionary techniques. We then discuss DNA nanostructure design criteria, operation efficiency in differentially-connected networks, and the application of computationally-aided sequence design for nanotechnological applications. We design a double crossover DNA motif with a silver base pair modification as a test case, and demonstrate successful model implementation. In sum, we present a novel computational framework for geometry-informed optimization of DNA networks. This tool is meant to enable design of both linear and nonlinear polynucleotide assemblies with inherent modularity for base parity, metalation, or more exotic nucleotide substitutions that may arise from advances in synthetic biology, nanomaterials and nanomedicine.