Abstract
LISE is a web server for a novel method for predicting small molecule binding sites on proteins. It differs from a number of servers currently available for such predictions in two aspects. First, rather than relying on knowledge of similar protein structures, identification of surface cavities or estimation of binding energy, LISE computes a score by counting geometric motifs extracted from sub-structures of interaction networks connecting protein and ligand atoms. These network motifs take into account spatial and physicochemical properties of ligand-interacting protein surface atoms. Second, LISE has now been more thoroughly tested, as, in addition to the evaluation we previously reported using two commonly used small benchmark test sets and targets of two community-based experiments on ligand-binding site predictions, we now report an evaluation using a large non-redundant data set containing >2000 protein–ligand complexes. This unprecedented test, the largest ever reported to our knowledge, demonstrates LISE’s overall accuracy and robustness. Furthermore, we have identified some hard to predict protein classes and provided an estimate of the performance that can be expected from a state-of-the-art binding site prediction server, such as LISE, on a proteome scale. The server is freely available at http://lise.ibms.sinica.edu.tw.
Highlights
Once a protein’s 3D structure becomes available, our knowledge about the protein and our ability to use the knowledge gained can be greatly enhanced
We recently reported a new ligand-binding site prediction method, LISE [11], in which the prediction is made based on triplets of protein surface atoms, called protein triangles, that are statistically enriched at ligand-binding sites [12]
For the much larger non-redundant set of 2073 protein–ligand complexes retrieved from the FINDSITE website [3], LISE’s Top1 and Top3 success rates were 72.7 and 85.6%, respectively (Table 1), both showing a decrease of $10% compared with those using the small data sets
Summary
Once a protein’s 3D structure becomes available, our knowledge about the protein and our ability to use the knowledge gained can be greatly enhanced. One approach for obtaining knowledge from structures is the prediction of binding sites for small molecule ligands that is needed in a variety of structure-based investigations, notably virtual docking in computer-aided drug design Because of this need, an increasing number of web servers and databases have been established for predicting/archiving small molecule-binding sites. They include those identifying the largest cavities on protein surfaces, those computing energetically favoured binding regions, and those using other computational methodologies [1,2,3,4,5,6,7,8,9,10]. This is because most of the binding site prediction methods have been tested on two small benchmark data sets containing, respectively, 210 and 48 pairs of ligand-bound and -unbound protein structures [6,7,8,9,10]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call