2.2 Protein preparation
PPI and PPepI dataset entries were imported in Schrödinger Maestro [38] and prepared as discussed here. Each of the entries in both the benchmarking dataset were prepared using a Protein Preparation Wizard utility available in Maestro [38]. For each PDB, protein hydrogens were stereochemically added and appropriate ionization states (at pH 7.0) for the acidic and basic amino acid residues were maintained. Missing side-chain atoms were added using a rotamer library of Xiang and Honig [41] and missing backbone atoms were fixed using a loop modeling procedure implemented in Prime [38]. Subsequently, the structure optimization, which utilizes maximization of hydrogen bonding, was carried out to i) identify the most probable positions for freely rotating –OH and –SH hydrogens, ii) to assign the correct protonation states of charged residues, iii) tautomers of His, and iv) Chi ‘flip’ of Asn, Gln and His residues. Finally, to relieve any steric clashes within atoms, the structures were energy minimized using OPLS2005 force field until the heavy-atom displacement converged to root mean square deviation (RMSD) of 0.30 Å. Many entries were manually examined and processed, where it was not possible to prepare using the default option.