File S1 - Optimizing Scoring Function of Protein-Nucleic Acid Interactions with Both Affinity and Specificity

Supporting figures and tables. Figure S1 The development of SPA-PN contains three stages: The preparation of database, optimization of the scoring function, Testing and application of SPA-PN. Figure S2 Typical atom-pair interaction potentials of SPA-PN. (A and B) Two of most frequently occurred atom pairs. (C and D) Atom pairs related to hydrogen bond. (E and F) Atom pairs involving phosphorus atom. Table S1 Training dataset for the development of SPA-PN. Table S2 Experimental determined affinities and SPA-PN predicted affinities for 30 protein-DNA complexes of the testing dataset1, the calculated affinities were obtained by scaling the binding scores with linear fitting equations (SPA-PN: y = 0.0045*x-5.129, Affinity-PN:y = 0.0044x-5.080, Pre-optimized: 0.0043x-5.443, Rosettadock: y = 0.0053x-7.24) based on the experimental affinities. Table S3 PDB codes of the testing dataset2. Table S4 15 Atom types used for calculating the atom pair potentials based on the SYBYL definition of atom type. The atom types can be converted from PDB files by the software OpenBabel. Table S5 95 effective types of atom pairs for the protein-nucleic acid interactions with the cutoff of total occurrences larger than 600 in the training dataset. Table S6 The high accuracy quality of CAPRI assessment criteria was taken to define the near-native conformation.
(DOC)