BAPPL

Bappl+ is a improved methodology for predicting the binding affinities of protein-ligand and metalloprotein-ligand complexes. It computes binding affinity based on the most important energetic contributors such as electrostatics, van der Waals, hydrophobicity and entropy of protein and ligand. For metalloprotein-ligand complexes, It uses the explicitly-derived quantum-optimized charges for various metal ions (Zn, Mn, Mg, Ca and Fe). It uses the Random Forest to derive the final score. This methodology (implemeted in web server) is widely tested and developed on the PBDbind datasets of 2007, 2013 and 2016 releases. Bappl+ achieves a strong pearson correlations with respect to experimental affinities in all the core datasets with low standard deviations and works better then most of the renowned state-of-the-art scoring functions.