Rother, Kristian;Froemmel ,C. - Building a Protein Structure Workbench with PyMOL

ECCB 2002 Poster sorted by: Author | Number

Next | Previous poster (in order of the view you have selected)

Title: Building a Protein Structure Workbench with PyMOL	P136
Rother, Kristian; Froemmel ,C. kristian.rother@berlin.de Protein Structure Theory Group, Charite, Humboldt University Berlin

Method integration has been recognized as a major challenge in today's bioinformatics research [1]. A huge amount of software for all kinds of applications is freely available, but they lack a common framework. For Sequence-based analyses, the Bio* (BioPerl, BioJava, BioPython) project has received attention as the major method integration project. For structure-based studies, the most essential component of a work environment is a molecular display and modeling tool.

PyMOL [2] is an open-source molecular modeling program written in the Python language. For achieving a high picture quality and performance, basic parts of the software were implemented in C.
As all the user interaction can be handled through a Python programming interface, extending the PyMOL platform is very straightforward. Python code is inherently quick to be written and easy to maintain. Thus, connecing the PyMOL environment to other high-performance software using Python code as a bridge is an efficient way to integrate different methods.

We have written several extensions for PyMOL in order to make standard structure analysis programs available from within the PyMOL environment. Most plugins bridge the gap between the PyMOL environment and an application that has been installed somewhere on the local machine, others enhance PyMOL's own capabilities. Activating the extensions does not require any alteration of the original PyMOL source code or recompilation.

The new functions are: Calculation of secondary structures using DSSP [3], display of intramolecular interfaces [2], calculation of hydrogen bonds using HBExplore [4], superposition of atomic structures [5], prediction of active sites and grooves on the surface using PASS [6], retrieving structures from the PDB [7], Calculating the solvent-accessible surface [8], removing junk atoms from PDB files, calling frequently used scripts and easy creation of smooth animations.

Within PyMOL, these new functions are available from the graphical user menu, the command line, command scripts and python programs. They can be flexibly used by human users as well as from further extensions.
We would like to demonstrate that extensions of the PyMOL system are possible even with minimal knowledge about Python or the PyMOL internals. We provide a template for developing own PyMOL plugins.
The integration of custom applications into a common modeling environment appears to be especially attractive for teaching purposes.

However, when a large number of new functions is added this way, the user will be easily confused by the possibilities he has. It will be necessary to structure extensions for PyMOL, as soon as they are accumulating. Basically, this problem is also present when many different methods are brought together without a framework.
We consider it as another disadvantage of our method that it requires some manual interaction until everything is installed. Many of the applications have to be installed manually because of licensing reasons.
Both problems could be solved by establishing an interface for automatic registration of methods that can be applied to molecular structures. At the moment, no such architecture is available for the PyMOL system. A service registry was to suit to applications on the local machine, the local network and the internet.

Further plans include extending PyMOL to use more standard applications, e.g. packing calculations. We also look forward to provide an interface to standard web service protocols like XML-RPC or SOAP. Inclusion of the extensions presented here into the PyMOL source code is intended by the authors in the near future.

The extensions for PyMOL are available on: http://www.rubor.de/bioinf .

[1] Stein L. Related Articles Creating a bioinformatics nation. Nature. 2002 May 9;417(6885):119-20.
[2] DeLano, W.L. The PyMOL Molecular Graphics System (2002) DeLano Scientific, San Carlos, CA, USA. , http://www.pymol.org.
[3] Kabsch W. and Sander C., Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers. 1983 Dec;22(12):2577-637. http://www.cmbi.kun.nl/swift/dssp/.
[4] HBexplore - A New Tool for Identifying and Analyzing Hydrogen Bonding Patterns in Biological Macromolecules, Lindauer K., Bendic C. and Suehnel J., Comput. Appl. Biosci. 1996, 12, 281-289, http://www.imb-jena.de/www_bioc/hbx/hbx.html.
[5] Algorithm for superposition of arbitrary atomic arrangements. Goede, A., Preissner, R. and Froemmel, C. (submitted).
[6] Fast Prediction and Visualization of Protein Binding Pockets With PASS, G. Patrick Brady, Jr. and Pieter F.W. Stouten, Journal of Computer-Aided Molecular Design, 14: 383-401, 2000, http://www.delanet.com/~bradygp/pass/.
[7] The Protein Data Bank: a computer-based archival file for macromolecular structures. F.C.Bernstein, T.F.Koetzle, G.J.B.Williams, E.F.Meyer Jr, M.D.Brice, J.R.Rodgers, O.Kennard, T.Shimanouchi, M.Tasumi J. Mol. Biol. 112 pp. 535-542 (1977) http://www.pdb.org/.
[8] Tsai J, Taylor R, Chothia C, Gerstein M. Related Articles The packing density in proteins: standard radii and volumes. J Mol Biol. 1999 Jul 2;290(1):253-66, http://bioinfo.mbb.yale.edu/geometry/.