Theoretical Investigations and Microarray Applications of DNA multiplex hybridization reactions

Abstract

General equilibrium and kinetic models with numerical solutions describing DNA multiplex hybridization reactions are developed and presented. Theoretical investigations on example multiplex systems were carried out in order to test the sensitivity of the model system to a variety of experimental and model parameters. A spotted microarray platform was calibrated to yield quantitative results, and the theoretical model motivated the design of an effective kinetic assay to discriminate SNPs in a multiplex fashion. Effects of duplex length, %G+C and relative position of the SNP on duplex hybridization and SNP resolution were determined. The theoretical model of multiplex hybridization accurately predicted observed results, and implicated target concentration as a critical variable in multiplex SNP detection. Hybridization intensities of 30 short duplex DNAs, measured on microarrays, were compared with solution state thermodynamic stabilities obtained by differential scanning calorimetry (DSC). A linear relationship between intensity and stability was found for perfect match and most mismatch duplexes. Examination of outliers suggested that duplex length and tandem mismatch position could be important factors contributing to observed deviations from linearity. Thermodynamic parameters for 32 duplexes containing 2X2 and 4X4 mismatches involving g/a and c/t pairs, together with 10 corresponding perfect matches, were evaluated in 85 mM, 300 mM, and 1M Na+ buffers using DSC. Negative standard heat capacity changes were found for duplexes containing ga/ag tandem mismatches and a logarithmic dependence on [Na+] was observed for all evaluated thermodynamic parameters. A detailed statistical comparison with nearest-neighbor (NN) model predictions identified gross inaccuracies in ΔH°, ΔS°, and ΔG° predictions for mismatches at all sodium ion concentrations, and perfect matches at 85 and 300 mM Na+. Tm predictions, however, were much more accurate. The purine/pyrimidine order and G+C content of flanking Watson-Crick (W/C) base-pairs was shown to influence the stability of the intervening tandem mismatch, and contextual effects beyond the adjacent flanking base pairs (W/C or not) were shown to be negligible. Since dependence of tandem mismatch stability on flanking bases precludes their incorporation into a doublet model, a model-independent perturbation approach is instead proposed for predictions of thermodynamic parameters relative to a corresponding duplex.