Overview
Research
Gallery
Download
Feedback
Home


Research

One of the grand challenges in computational biology is the molecular docking problem. The molecular docking problem is to determine how molecules interact with other molecules and plays a key role in understanding how cells function. Its solution will help scientists find better ways of changing cell functions with new and more specialized drugs designed specifically for that purpose [Figure 1]. Given two independently determined molecular structures, the molecular docking problem predicts the bound association, or best fit between them, while allowing for conformational changes of the individual molecules during construction of a molecular complex. DockingShop is an integrated environment that permits interactive molecular docking by navigating a ligand or protein to an estimated binding site of a receptor with real-time graphical feedback of scoring factors as visual guides. Our program can be used to create initial configurations for a protein docking prediction process. Its output --the structure of a protein-ligand or protein-protein complex-- may serve as an input for a protein docking algorithm, or an optimization process. This tool provides molecular graphics interfaces for structure modeling, interactive manipulation, navigation, optimization, and dynamic visualization to aid users steer the prediction process using their biological knowledge.



Figure1. DockingShop Rendering of HIV-1 Protease in Complex with the Cyclic Sulfamide Inhibitor Aha006 (PDB code 1AJV).

Features

Hydrogen Bonds Visualization - Hydrogen bonds play an important role in molecular docking because they help to stabilize and strengthen a bound complex. DockingShop visualizes inter- and intra-molecular hydrogen bonds. Moreover, it provides visual guides to facilitate the formation of hydrogen bonds through molecular manipulations. The program renders a bond site, i.e., the midpoint of a hypothetical hydrogen bond, for each charged backbone group, showing a potential bond's midpoint position and orientation. Forming bonds is thus reduced to aligning the midpoints and orientations of two differently charged backbone groups.
Molecular Overlap Visualization - DockingShop calculates and visualizes atom collisions in real-time during interactive docking to assist users in evaluating the overlap, thus helping to achieve the desired molecular interactions when a protein or ligand is close to a binding pocket. Penetrations are penalized and scored. These collision detection approaches can also help users reject a solution when the inter- and intra-molecular penetrations exceed a tolerance threshold during visual assessments.
Energy Visualization - DockingShop provides an interface layer that allows dynamically loading of energy computation modules so that users can couple different energy functions from commercially or internally developed packages. Our program has two schemes for the energy visualization, atom-based and volume-based. In the former, per-atom values are visualized by mapping colors to atoms' van-der-Waals spheres. Volume rendering of energy produces a cleaner and more appealing representation than per-atom rendering because it reduces occlusion and visual clutter caused by van-der-Waals spheres. DockingShop goes beyond visualization to include numerical analysis of the energy. This feature assists the users to relate calculated quantities to molecular motions and to measure the effect of molecular interactions and structure alignments.
DockingShop: A Tool for Interactive Molecular Docking Scoring Function - DockingShop integrates biological knowledge into the docking process by allowing users to alter the docking process in response to visualization of computational parameters and scoring functions. This approach helps users to accelerate the search stage by providing them with an efficient method for a reliable discrimination between correct solutions and false positives. The computational parameters can be either visualized or analyzed in numerical form.
Flexibility of Molecular Structure - DockingShop's manipulations are based on those in ProteinShop. They are based on an inverse kinematics (IK) algorithm that transforms parts of a protein with respect to other parts by rotating the backbone dihedral angles, without changing any bond lengths. During refinement, users can also finely adjust the protein structure (receptor) by changing the dihedral angles of a selected residue within the range of the Ramachandran plot. DockingShop permits another type of refinement by substituting the side chain of a selected residue on the receptor protein using a rotamer library. DockingShop also permits to perform mutations, which allow users to change a residue for another while keeping the backbone fixed.
Interactive Docking - An important capability of DockingShop is its six-degree-of-freedom manipulation of single and multiple molecules (like proteins, ligands and water). Users can alter the overall view of a molecular complex through the main window. The orientation and position of each individual molecule inside the complex are controllable by a pilot window. Instead of an exhaustive conformational space search that takes days to complete, users can navigate one molecule to the binding site of another molecule in a very short period of time when binding sites are known. Users can bring two molecules close together and see them bind in real-time.
Binding Site Prediction -DockingShop allows users to use Pocket [1], which is based on an analytical method for detecting pockets in proteins, to decide the binding sites.


Development

The design of DockingShop follows a modular development concept. DockingShop is written in C++ and relies on the OpenGL library for three dimensional graphics rendering and is portable across most Unix platforms like Linux, SGI, and Mac. DockingShop uses FLTK, an interpreted interface layer, for the graphical user interface. Like ProteinShop, DockingShop provides support for modeling and visualization of molecules and biological data, constraints-based manipulation of molecular structure based on inverse kinematics algorithms, and optimization of an energy function. However, the essential elements of computational biology include computational methods to predict structure, properties, and behavior of those molecules along with molecular modeling and visualization. To that end, we are developing interface layers to bridge between computational biologists and emerging computational methods. Such interfaces allow developers to plug in their own codes to build powerful new applications for computational biology research.




 Figure 2. DockingShop’s user interface in Visualization mode.  Structure of complex of synthetic HIV-1 protease (HIV-1 PR) complex with a substrate-based inhibitor N-Acetyl-Thr-Ile-Nle-psi(CH2-NH)-Nle-Gln-Arg amide (MVT-101) ) at 2.3 A resolution.(PDB code 4HVP).



Presentations and Demonstrations

"DockingShop - a Tool for Interactive Protein Docking."     PDF
Ting-Cheng Lu, Nelson Max, Jinhui Ding, Wes Bethel, and Silvia Crivelli. submitted to IEEE Visualization 2005

Publications


References
1. http://www.cs.ucdavis.edu/~koehl/ProShape/