Bioinformatics Advance Access originally published online on May 23, 2006
Bioinformatics 2006 22(14):1803-1804; doi:10.1093/bioinformatics/btl197
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
BDT: an easy-to-use front-end application for automation of massive docking tasks and complex docking strategies with AutoDock
1 Departament de Bioquímica i Biotecnologia, C/Marcel·lí Domingo s/n Av. Països Catalans, 26 Campus de Sant Pere Sescelades, Universitat Rovira i Virgili, Tarragona 43007, Catalonia, Spain
2 Departament d'Enginyeria Informàtica i Matemàtiques Av. Països Catalans, 26 Campus de Sant Pere Sescelades, Universitat Rovira i Virgili, Tarragona 43007, Catalonia, Spain
*To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT 1 INTRODUCTION2 DESCRIPTION3 CONCLUSIONSREFERENCES |
|---|
Motivation: AutoGrid/AutoDock is one of the most popular software packages for docking, but its automation is not trivial for tasks such as (1) the virtual screening of a library of ligands against a set of possible receptors; (2) the use of receptor flexibility and (3) making a blind-docking experiment with the whole receptor surface. This is an obstacle for research teams in the fields of Chemistry and the Life Sciences who are interested in conducting this kind of experiment but do not have enough programming skills. To overcome these limitations, we have designed BDT, an easy-to-use graphic interface for AutoGrid/AutoDock.
Availability: BDT is available for free, upon request, for non-commercial research.
Supplementary information: http://www.quimica.urv.cat/~pujadas/BDT/
Contact: gerard.pujadas{at}urv.cat
|
1 INTRODUCTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DESCRIPTION3 CONCLUSIONSREFERENCES |
|---|
Bioinformatic tools such as sequence similarity searches, sequence analysis and the homology modeling of protein sequences with unknown experimental 3D structure are currently used as standard research routines in biochemical and biomedical laboratories without extensive user-training. Crucial to extending the use of such tools beyond the borders of bioinformatic groups has been the development of graphic or web-based interfaces that are on top of such sophisticated algorithms. These interfaces are usually set up with default values for the parameters that control the algorithm that are valid in most situations and that allow novice or occasional users to use the tools easily. Moreover, the graphic interfaces make it easy to customize the values of these parameters for more specific or expert uses. A nice example of these graphic interfaces is Swiss-PdbViewer/Deep View (Guex et al., 1999), which has dramatically increased the use of homology modeling beyond bioinformatic groups.
It is important to understand intermolecular interactions between small ligands and their macromolecular receptors in order to explain crucial life processes such as gene expression, the regulation of metabolic pathways and enzyme catalysis. Using docking algorithms to predict the 3D structure of macromolecule/ligand complexes is therefore of interest for a wide range of biochemical and biomedical investigations. In this context, one of the most reliable, robust and popular energy-based docking packages is AutoGrid/AutoDock (Morris et al., 1998) because it allows a very efficient docking of flexible ligands (e.g. substrates, drug candidates, inhibitors, peptides, etc.) onto receptors (e.g. enzymes, antibodies, nucleic acids, etc.) even when the receptors are also flexible (Österberg et al., 2002).
While ADT (the graphic interface for AutoGrid/AutoDock) has strongly decreased the learning-curve needed for using this package, it is also true that some docking tasks with AutoGrid/AutoDock, although possible, are far from trivial for users without strong computer skills. Examples of such tasks are (1) using receptor flexibility during docking; (2) the automatic docking of a large library of ligands onto one or more receptors [because it is necessary to set up (a) one specific AutoGrid's command file for each receptor used and (b) one specific AutoDock's command file for each ligand–receptor pair that is assayed] and (3) docking a ligand library onto one or more receptors without defining one a priori ligand-binding site on them (therefore, using the whole receptor surface) and using a distance between grid points as short as the user needs. To overcome these difficulties, we have developed BDT, a Tcl/Tk graphic front-end application that runs on top of four Fortran programs (i.e. make_grids, combine_grids, make_docks and analyze, one under each BDT window tab), which control the conditions of AutoGrid and AutoDock runs.
|
2 DESCRIPTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DESCRIPTION3 CONCLUSIONSREFERENCES |
|---|
The BDT window has four tabs: (1) MakeGrids, (2) CombineGrids, (3) MakeDocks and (4) Analyze (see the BDT website at http://www.quimica.urv.cat/~pujadas/BDT/ for a detailed description of the tabs and the algorithms of the programs under them). The tabs are sorted from left to right according to their sequence of use [from (1) to (4) in the above list]. This is because the program that runs under each tab needs some input files that result from executing the program under the previous tab. Users are therefore recommended to move to the next tap until BDT informs them by e-mail that the execution of the program under one specific tap is finished. Four buttons are common to all tabs and have the same function: (1) ‘execute’, which starts running the Fortran program under the tab; (2) ‘reset’, which replaces custom values with the default ones and (3) ‘save’, which stores the current tab parameter values in a file for later recovery with the ‘load’ button. Users also have to set in all tabs how the underlying program will communicate with them by e-mail by (1) typing their e-mail address in the corresponding windows and (2) indicating the amount of information they want to receive from the program (this option is not available for the Analyze tab). BDT also has a contextual help (activated by default) for guiding users when filling the different fields in the tabs. Default parameter values are also provided for the different fields that are useful in most common situations.
From the MakeGrids tab, users can (1) include or exclude the flexibility of the receptors in the calculations and (2) search for the ligand-binding site in all the receptor surface (‘Area around the whole receptor surface’ option) or in a user-defined portion of it (‘Area around one specific point’ option) using, in both cases, a grid-point distance as short as they like regardless of the dimensions of the 3D space that is searched. Users can then combine the flexibility and ligand-binding site location strategies according to their needs (e.g. they can do the docking in a specific part of a flexible receptor). Thus, for each protein, this tab can automatically deal with either a single PDB file or with a set of PDB files corresponding to different snapshots of its conformations [such as those that (1) can be found when a specific protein has been crystallized in a set of different conditions and the resulting structures deposited in the PDB (http://www.pdb.org, Berman et al., 2000); (2) can be readily obtained from FlexWeb tools (http://flexweb.asu.edu/, Zavodszky et al., 2004) or (3) can be retrieved from the MODEL database (http://mmb.pcb.ub.es/MODEL)]. Here the user does not need to do anything special to take into account receptor flexibility. Receptor flexibility is automatically assumed for those receptors provided to this tab by the user's list that has more than one PDBQS file (where PDBQS corresponds to the input format for the receptors in AutoGrid) in the selected PDBQS directory [it is assumed that PDBQS filenames that start with the same four-character code (usually a PDB code) correspond to different conformations of the same receptor]. Therefore, this tab is used to control where AutoGrid has to run and its output files can be used either with the CombineGrids or the MakeDocks tabs (depending on whether the receptor's flexibility is considered).
The CombineGrids tab is used to incorporate the receptor's mobility in docking calculations based on the work of Österberg et al. (2002). Briefly, this method combines all the grid maps from the different receptor conformations and the same probe to obtain a single grid-map file. In this file, the energy of each point is obtained from a weighted average of the energies of the same point in all the original conformational-dependent grid maps (where the corresponding weight is calculated using either a clamped grid or a Boltzmann assumption based on the interaction energy). The resulting grid maps can be readily used by the MakeDocks tab.
The MakeDocks tab enables easy selection of the receptors, ligands and conditions used by AutoDock during docking. Depending on the origin of the grid maps used (obtained from either the CombineGrids tab or the MakeGrids tab), the user will or will not consider receptor flexibility during docking.
The Analyze tab can be used to analyze the docking results (e.g. by comparing the docking of different ligands onto the same receptor, etc.). For each studied receptor, Analyze e-mails the user a PDB file and two RasMol (Sayle and Milner-White, 1995) scripts to use with it. The PDB file contains the coordinates for (1) the self receptor; (2) the box where the possibility of ligand binding is studied and (3) the cluster representatives for docking solutions of the assayed ligands. The RasMol scripts make it easier to compare the docking results of the different ligands on the same receptor. The first script (i.e. script_01) colors the ligands according to different intervals of affinity for the receptor irrespective of their identity (i.e. from hot to cold colors for decreasing affinity) and is therefore useful for identifying where the most important ligand-binding sites in the receptor structure are located. The second one (i.e. script_02) colors the ligand according to its identity and is therefore useful for comparing the specificity of the different ligands for each binding site.
|
3 CONCLUSIONS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DESCRIPTION3 CONCLUSIONSREFERENCES |
|---|
With just a few clicks of the mouse, BDT enables sophisticated docking strategies to be carried out, not only for research but also for teaching. Therefore, BDT contributes significantly to the progress of bioinformatics and biomedical research.
|
Acknowledgments |
|---|
We thank the authors of AutoGrid/Autodock for providing us with version 3.0.5 of their software and, especially, Dr Garrett Morris and Dr Ruth Huey for their help. We also thank Kevin Costello of our University's Language Service for correcting the manuscript. This study was supported by grant number CO3/O8 from the Fondo de Investigación Sanitaria (FIS) and AGL2005-04889 from the Comisión Interministerial de Ciencia y Tecnología (CICYT) of the Spanish Government. Montserrat Vaqué is the recipient of a fellowship from grant number CO3/O8.
Conflict of Interest: none declared.
|
FOOTNOTES |
|---|
Associate Editor: Martin Bishop
Received on March 27, 2006; revised on May 11, 2006; accepted on May 17, 2006
|
REFERENCES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DESCRIPTION3 CONCLUSIONSREFERENCES |
|---|
Berman, H.M., et al. (2000) The Protein Data Bank. Nucleic Acids Res, . 28, 235–242
Guex, N., et al. (1999) Protein modelling for all. Trends Biochem. Sci, . 24, 364–367[CrossRef][Web of Science][Medline].
Morris, G.M., et al. (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J. Comput. Chem, . 19, 1639–1662[CrossRef][Web of Science].
Österberg, F., et al. (2002) Automated docking to multiple target structures: incorporation of protein mobility and structural water heterogeneity in AutoDock. Proteins, 46, 34–40[CrossRef][Web of Science][Medline].
Sayle, R. and Milner-White, E.J. (1995) RasMol: biomolecular graphics for all. Trends Biochem. Sci, . 20, 333–379[CrossRef][Web of Science][Medline].
Zavodszky, M.I., et al. (2004) Modeling correlated main-chain motions in proteins for flexible molecular recognition. Proteins, 57, 243–261[CrossRef][Web of Science][Medline].
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||