Visual Clustering Approach for Docking Results from Vina and AutoDock

摘要

AutoDock Tools allows the analysis of docking files and is used to represent clustering conformations, yet it analyses only one docking file at a time and the method applied to represent the clustering complicates the visualization of clustering conformations. The creation of a plugin called PyDRA for the molecular visualizer PyMOL resolves that problem and allows to simultaneously process more than one docking file for the two types of file format from AutoDock 4.2 and Vina 1.1 (dlg and pdbqt). Moreover, this plugin facilitates the visualization of conformations through two clustering methods. The first method is a K-RMSD algorithm, which is based on the clustering through RMSD and enables the interactive visualization groups through a treemap. And the other one is based on a hierarchical clustering algorithm, using an algorithm of average distances which generates a dendrogram that offers the possibility to explore sequentially the groups that illustrate best the docking. The results obtained with the visualization methods implemented showed that the treemap, due to the implemented colour bar, facilitates to identify the clusters that have a greater affinity to the protein at a glance, and to determine which of the clusters hold a greater number of elements, on the other hand, the dendrogram shows a detailed analyses of the hierarchical clustering, which also enables the user to distinguish the clustering regardless the size of the window, as well as to differentiate each cluster and conformation in order to gain insight of docking results of Autdock and Vina. The fact that both visualizations are connected to PyMOL increases its ability of discernment.