Biological Computing & Nanotechnologies Initiative

We are striving to solve many complex biological problems that seemed unapproachable until recently, and also taking a closer look at the possibility of developing alternative models of computation. Projects fuse biotechnology and information technology — two of the most significant developments of the information age.

On the one hand, advances in computer science have played a critical role in addressing and solving problems in molecular biology, including the decoding of the human genome. The application of new computational techniques to process, analyze, and visualize emerging genomic data offers hope for solving many complex biological problems that seemed unapproachable until a few years ago. On the other hand, advances in biotechnology have raised the possibility of developing alternative models of computation.

The current strengths of the department in computational biology lie in functional genomics, analysis of protein structures, and biomolecular computation. In the area of functional genomics, research is focused on the development of principled methods for discovering genetic regulatory networks from diverse types of data. In addition, a significant amount of work is being done on the automatic classification of tissue types and automatic selection of relevant gene features from gene expression data.

In collaboration with researchers in computer science and biochemistry at Duke and other universities we are developing geometric techniques and paradigms for representing, storing, searching, analyzing, and visualizing biological structures. One of the projects in this area focuses on studying protein-protein interaction. We have defined the notion of interface surface between two or more complexed proteins and have developed an efficient algorithm for computing the interface surface. Following the ideas from Morse theory and topology we have defined elevation of points on a molecular surface. We use elevation to identify features on a molecular surface and to dock two proteins.

The area of biomolecular computation is an area of great potential because of the massive parallelism available at the molecular scale. The Duke DNA NanoTech group has demonstrated the use of DNA as smart glue to bring together electronically active nano-materials and also as templates for the formation of highly-conductive metallic nanowires. In collaboration with researchers from other universities, the group is exploring DNA self-assemblies of nano-electronic devices as a scaffold for nano-scale computational devices that could one day meet the Semiconductor Industry Association's grand challenge of replacing conventional CMOS devices. Nanoscale computers have the potential to revolutionize computing similar to the transition from vacuum tubes to integrated circuits.

The DNA NanoTech group at Duke is interested in understanding and exploiting the material properties of biomolecules (mostly, DNA and proteins) for applications in DNA computation, molecular databases, and nanofabrication. We are devising methods to encode data and algorithms into 1D, 2D, and 3D structures of DNA for execution of molecular computations and wet database searches. We are engaged in the design, implementation, and analysis of self-assembling DNA nanostructures with complex and specific 3D structures. These structures are capable of scaffolding or templating other materials with diverse functional properties to form micrometer sized objects with nanometer-scale feature resolution. For example, we have used DNA self-assembly for "bottom-up" nanofabrication of silver nanowires and arrays of gold nanoparticles. We have shown the nanowires to be highly conductive and are working to demonstrate single-electron transistors with the nanoparticles.