A Workbench for Computational Analysis of Protein Sequence and Structure on the Internet


Computational analysis of protein sequences with statistical models, machine learning techniques or empirical rules leads to prediction of structural and functional features of the corresponding molecules. Taking into account the recent explosion of biological information in terms of complete genome sequences the use of computational tools of analysis is considered essential. Individual tools are available through the Internet, and their performance depends on their liability and simplicity of use. We present a workspace that establishes data trafficking towards a collection of tools (developed also by our team), communication between them and simple representation of the results of sequence and structure analysis. At the sequence level, the following are now offered: Analysis of periodical features, structural classification with neural networks, prediction of secondary structure, prediction and topology of transmembrane regions and multiple sequence alignment. At the structure level, the creation and representation of protein structures and three-dimensional fit is possible. The end-user is offered with a user-friendly interface and a simple representation of the results. The system’s open and decentralized design permits the addition of new individual modules that may reside anywhere on the Internet. Our future plans include the development of modules for fold recognition, receptor prediction and sequence complexity analysis. Interconnection between the system, the individual tools and the user is established through the HTTP protocol, the only requirement for its usage being Internet access (web site http://athina.biol.uoa.gr/DAM-Bio ).

22nd Conference of the Hellenic Society for Biological Sciences