Dr. Jenny J. Yang's Lab

Click here to edit subtitle







* PC Windows/DOS Verison

* User's guide to GG-windows verison



Calcium in proteins is surrounded by oxygen atoms. The bond-strength contributed by each oxygen has been expressed as a valence function. Based on the function, a high-performed algorithm for detecting calcium-binding sites has been developed by embedding a protein in a very fine grid of points. Although it can predict calcium-binding sites at very high accuracy, the principle of detecting binding sites uses exhaustive searching through all local grid points, which leads to intensive computation and thus does not suits for automated screening through protein databases. The key point of improving the computational complexity relies on the knowledge that calcium-binding sites are geometrically predetermined. Despite the long belief that such a geometric structure cannot be existed, the results of this paper show that circumcenters of ligand oxygen atoms play a critic role in detecting binding sites. The program of detecting calcium-binding sites contains three parts: finding the sites where there are four oxygen atoms bonded closely together in a graph-theoretic way; choosing the sites whose geometric circumcenters satisfying certain threshold conditions, eliminating redundant predictions. All three algorithms are extremely fast. They are linear to the number of oxygen atoms, the number of sites and the number of all atoms other than oxygen, and the square of the number of sites, respectively. The high prediction accuracy has also been achieved using this approach.


Hits   HTML Hit Counters   since 3/27/2008
Last Modified: 6/29/2016 For more information, contact: [email protected]