Figure 1.
Schematic view of illustrating the equivalence of conventional specificity to intrinsic specificity.
(A) The same nucleic acid (N, red) binding with multiple protein receptors (blue, to
), showing the conventional specificity as the gap in binding affinity of the nucleic acid binding to the specific protein receptor (
) in discrimination against other protein receptors. The binding affinities are represented with corresponding energy spectrum (green). (B) The same nucleic acid (N, red) binding on a large protein receptor thought as the multiple different receptors linked together (blue) with multiple binding modes (
to
), showing the intrinsic specificity as the gap in binding affinity of the native binding mode (
) in discrimination against other binding modes.
Figure 2.
(A) Evolution of the success rate and the average interfacial RMSD () as the iteration precedes. (B) The distribution of ISR values calculated with pre-optimized SPA-PN and optimized SPA-PN respectively.
Figure 3.
A typical example of protein-nucleic acid complex (PDB 1TRO).
(A) Protein-nucleic acid binding structure with protein colored in blue and nucleic acid colored in red. (B) Plot of interfacial RMSD () as a function of the fraction of native contacts (
) for 1000 docking decoys of the typical complex. (C) Energy spectrum and distribution calculated with pre-optimized SPA-PN (green) and optimized SPA-PN (magenta), the corresponding ISR values are shown and the energy of the native conformation is marked as red.
Figure 4.
Pearson correlation between the predicted affinities calculated by SPA-PN and experimental binding affinities for 30 protein-DNA complexes of testing dataset1.
The correlation coefficient () is 0.862 (statistical significance
). The predicted affinities are obtained by scaling the binding scores with a linear equation:y = 0.0045x-5.129 which is a fitting equation based on the experimental affinities.
Table 1.
Pearson correlations () between the predicted binding affinities and experimentally measured binding affinities for 30 protein-DNA complexes of the testing dataset1.
Table 2.
Success rates () of identifying the native or near-native conformations for testing dataset2 including 232 protein-DNA and 83 protein-RNA complexes.