Book of Abstracts: Albany 2005
Validation of DNA Sequences for DNA-Based Computing Using SYBR Green Fluorescence and Thermodynamic Calculations
We are developing a computing basis that employs new DNA-based algorithmic paradigms with the ultimate goal of using DNA to solve computationally hard mathematical problems. An important first step is the creation and validation of DNA sequences to be used in coding. It is critical that DNA strands behave as predicted from their sequence; otherwise computing errors can result. Here we describe the generation of two sets (a DNA code) of single-stranded DNA sequences designed such that each code strand should hybridize only with its reverse-complement and not with any other strand in the set. The code design draws on principles from nearest neighbor studies and from the thermodynamics of base stacking (1). We tested every possible combination of strands for their potential to mispair by measuring fluorescence over varying temperatures in the presence of SYBR Green I, a dye whose fluorescence increases exponentially when bound to double-stranded DNA. Since the strands are designed to be linked together ultimately, we also tested hybridization of strand junctions. Most of the strands and their junction sequences showed suitable preference for their direct complement. One exception was found in strands that contained four or more successive guanine residues. These strands were eliminated from the code. As a complementary approach, the thermodynamics of hybridization were also examined using the program PairFold from RNASoft (2).
References and Footnotes
Wendy Knapp Pogozelski1
1Department of Chemistry