Figure 3

Comparison of simulation results with the experimentally determined hybridization affinities for two probe sequence motifs ( A ) and ( B ). The four small sub-figures in the top section (from top left to bottom right) show the partition function Z and the duplex binding constant K as a function of defect position x (semi-logarithmic plots), the NN-free energies Δg° of particular NN-pairs as a function of NN-pair position x _NN , and the statistical weight for complete duplex dissociation w _D as a function of defect position. Irregularities in Z(x) at the duplex ends are an artifact caused by the fact that only a single NN-pair is affected by a MM-base pair at the duplex end. The middle sub-figure shows the base pair opening probabilities (the fraction of strands in which the corresponding base pair at position x _BP is unzipped) as a function of the defect position. The spectrum of differently colored curves encodes the different defect positions x _MM (red – defect at left end; purple – defect at right duplex end). The bottom sub-figure compares the experimentally determined MM defect profile (mismatched base: A (red cross), C (green circle), G (blue star), T (cyan triangle); gray symbols correspond to PM probes) with the simulated MM defect profile θ(x) (dashed orange line). With Δg _def = 1 kcal/mol (at the simulation temperature of 325 K) and an error rate of 12 percent (per synthesis step) the calculated defect profile θ(x) matches well the experimental data.