AutoDock, a flexible docking software, is widely used in academic research and currently the most cited docking software (1). It involves complicated steps that require careful parameterizations. AutoDockTools (ADT), the Python based user interface developed by NBCR, is designed for setting up AutoDock experiments in an interactive fashion through menu driven commands. To reduce the confusion introduced by the lack of apparent ordering in menu commands, we have created a prototype of the computer-aid drug discovery pipeline (Figure 5), fully leveraging the previous development effort of Vision, PMV, and ADT, aka, MGLTools. The component based programming model adopted by the MGLTools developers, a suite of tools for molecular visualization, docking simulations and scientific workflow management (2), has encouraged code reuse. Advanced users are able to leverage the visual programming ability of Vision to implement software for docking (3), and for identification of ligand binding cavities on proteins (4).
Figure 1. Rich desktop clients such as Vision may leverage distributed applications exposed as Opal based web services.
Vision is a visual programming environment with workflow support. Through the use of the Opal toolkit (5), which includes a dashboard for job monitoring, and service listings, and an Opal GUI component that enables the automatic interface generation in different clients (6), Vision is now able to access distributed resources through web service calls. The prototype pipeline accesses NBCR services such as protein electrostatics calculations using APBS (7) and virtual screening experiments using AutoDock (8).
DejaVu is the 3D visualization engine based upon OpenGL for the MGLTools. One of the recent features is the availability of a tile renderer, which enables the rendering of images of arbitrary size while using the hardware acceleration of the onboard graphics card. This rendering mode is compatible with all other rendering options in DejaVu, including anti-aliasing, stereo, Non Photo Realistic, etc (Figure 6). The Non Photo Realistic Rendering mode allow the rendering of silhouette lines. The availability of tile rendering, and NPR mode would make cavities on proteins much more visible, and make it easier to set up grid boxes in the case of AutoDock 4.
While different users may be familiar with different workflow programs, such as, Kepler, Taverna, Pipeline pilot, NBCR has made key applications as Opal based web services. These services are accessible by different workflow programs. An Opal actor for Kepler is currently available through the Kepler CVS repository. For virtual screening workflows that may require extended execution time, a new batch execution mode for Vision will allow the entire workflow to be run and accessible as a service. This type of Software as a Service (SaaS) model for application delivery is an important step towards advanced cyberinfrastructure for biomedicine. |
Figure 2. New MGLTools feature highlights: 1) tile-rendered image; each tile has a different background. 2) 3D labels. 3) Coarse molecular surface. 4) isocontours and slicing planes for 3D volumetric data. 5) Blue tongue virus capsid (960 proteins, ~3 million atoms). 6) Electrostatic potential mapped on molecular surface and volume rendered. 7) new tree widget showing the hierarchy of molecules, chains, residues and atoms and allowing to show/hide/color various representations (i.e. CPK, Sticks and Balls, surface, labels, secondary structure, etc.). 8) 2D plots in 3D scenes. 9) ribon diagram of DNA. 10) Matplotlib plots created in Vision.
____________
References:
1. Sousa, S. F., Fernandes, P. A. & Ramos, M. J. (2006). Protein-ligand docking: current status and future challenges. Proteins 65, 15-26.
2. Sanner, M. F., Stoffler, D. & Olson, A. J. (2002). 10th International Python Conference, Feb 4-7, 2002, Alexandria, VA.
3. Zhao, Y. & Sanner, M. F. (2008). Protein-ligand docking with multiple flexible side chains. J Comput Aided Mol Des 22, 673-9.
4. Harris, R., Olson, A. J. & Goodsell, D. S. (2008). Automated prediction of ligand-binding sites in proteins. Proteins 70, 1506-17.
5. Krishnan, S., Clementi, L., Ding, Z., Arzberger, P. W. & Li, W. W. (2008). Leveraging the Power of the Grid with Opal: A Guide to Biomedical Application Developers and Users. In Handbook of Research on Computational Grid Technologies for Life Sciences, Biomedicine and Healthcare (Cannataro, M., ed.), pp. In Press, Milan.
6. Clementi, L., Ding, Z., Krishnan, S., Wei, X., Arzberger, P. W. & Li, W. W. (2007). Grid Computing Environment 2007 (GCE 07), Reno, Nevada.
7. Baker, N. A., Sept, D., Joseph, S., Holst, M. J. & McCammon, J. A. (2001). Electrostatics of nanosystems: application to microtubules and the ribosome. Proc Natl Acad Sci U S A 98, 10037-41.
8. Morris, G. M. (1998). Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. Journal of Computational Chemistry 19, 1639-1662.
|
