Skip to main content
Fig. 3 | Journal of Cheminformatics

Fig. 3

From: EDock: blind protein–ligand docking by replica-exchange monte carlo simulation

Fig. 3

The ligand-receptor atomic distance profile heuristic. To derive the distance profile, query receptor and ligand structures are separately matched with the template-ligand complexes, where the receptor structure match is copied from the alignment generated by the first step of EDock, and the ligand structure match is made by LS-align [29]. The average distance \(\mu_{ij}\) is then calculated by \(\mu_{ij} = \left( {1/N} \right)\mathop \sum \nolimits_{t = 1}^{N} h_{ij} \left( t \right)\), where \(h_{ij} \left( t \right)\) is the distance between ith atom of the query receptor and the ligand atom (at tth template) that corresponds to the jth atom of the query ligand, and \(N\) is the number of template complexes which have the ligand aligned with the jth atom of the query ligand in the LS-align alignment. In EDock, we only consider template ligands whose LS-score is > 0.7 with the query ligand. The variance is calculated as \(\sigma_{ij}^{2} = \left( {1/N} \right)\mathop \sum \nolimits_{t = 1}^{N} (h_{ij} \left( t \right) - \mu_{ij} )^{2}\)

Back to article page