The Art of Compiling Protein Binding Site Ensembles

Abstract

In the published version of this Communication, figures 1and 2contain errors. Several elements of the structure diagrams are not depicted as submitted by the authors. All dashedl ines have been converted to solid lines duringp roduction of the paper leadingt oi ncorrectly illustrated hydrogen bonds and delocalized electrons. Figure 1. Complex of aT etrahydroquinazoline Antifolate with Di-hydrofolate Reductase[21] after processing with Protoss. Figure 2. Ricin Ac omplex with Neopterin [22] after processing with Protoss.

Ranking Enzyme Structures in the PDB by Bound Ligand Similarity to Biological Substrates

“…For example, ligand-homology modeling ( Drwal and Griffith, 2013 ) uses binding site alignment and ligand transposition ( Konc et al., 2015 ) as the basis to score and validate protein-ligand interactions ( Najmanovich et al., 2008 , Shin et al., 2011 , Evangelidis et al., 2012 , Konc et al., 2012 , Kurbatova et al., 2013 , Zhou and Skolnick, 2013 , Heo et al., 2014 , Konc and Janežič, 2014 , Cleves and Jain, 2015 , Roy et al., 2015 ). Docking methods have also been enhanced by using the location of bound ligands to supplement scoring functions ( Stanton et al., 2015 , Anighoro and Bajorath, 2016 ) and to enable false positives to be pruned from virtual screening ( Bietz et al., 2016 ). Large-scale computational methods that identify 3D binding pharmacophores ( Meslamani et al., 2012 ) or represent ligand-protein interactions as networks ( Kalinina et al., 2011 , Martínez-Jiménez and Marti-Renom, 2015 , Kasahara and Kinoshita, 2016 ) are also likely to be enhanced with knowledge about the biological relevance of the ligands on which they are based, potentially improving the prediction of ligand-protein interactions ( Kinnings and Jackson, 2011 ) and the performance of machine learning methods to classify actives from decoys ( Chupakhin et al., 2013 ).…”