Quantitative assessment of SNP discrimination for computational molecular beacons

The ability to discriminate nucleic acid sequences is necessary for a wide variety of applications: high throughput screening, distinguishing genetically modified organisms (GMOs), molecular computing, differentiating biological markers, fingerprinting a specific sensor response for complex systems, etc. Hybridization-based target recognition and discrimination is central to the operation of nucleic acid microsensor systems. Therefore developing a quantitative correlation between mishybridization events and sensor output is critical to the accurate interpretation of results. Additionally, knowledge of such correlation can be used to design intelligent sensor systems that incorporate mishybridization noise into system design. Using experimental data produced by introducing single mutations (single nucleotide polymorphisms, SNPs) in the probe sequence of computational catalytic molecular beacons (deoxyribozyme gates) [Stojanovic & Stefanovic, 2003], we investigate correlations between free energy of the target-probe complex and the measured fluorescence of the deoxyribozyme gate. Experimental data for forty-five SNP-containing probe sequences are compiled and compared against the true probe sequence to determine the relationship between position, type of mutation, and the fluorescence level of the molecular beacon. The sequence set accounts for every possible SNP for a fifteen-base probe. Experiments are conducted using a 55 mul detection volume containing a modified YESi_A(E6) deoxyribozyme molecular beacon (100 nM) [Stojanovic et al., 2001], TAMRA substrate (1 muM) and input sequences (2 muM). Using free energy as a first-approximation of the energetic interactions that occur during target-probe recognition, we generate empirical data for each target-probe pair using a nucleic acid hybridization thermodynamics server called HyTher (http://ozone2.chem.wayne.edu/). HyTher uses empirical fits of experimentally measured data to generate hybridization ther- - modynamic predictions for nucleic acid sequence pairs. Empirical data for all target-probe combinations are correlated with experimental fluorescence measurements to determine a quantitative link between target-probe hybridization free energy and molecular beacon fluorescence for each SNP-containing probe. We investigate Bayesian-based classification approaches as well as combinatorial design based methods for identifying and classifying mismatch patterns that produce similar fluorescence levels