Print optimized Version Project: MARG

Short Title: "Modelling and Animation of Regulatory Gene Networks"

Number of grant (DFG): HO 1178/10-1 

Project Leader:



  • Ming Chen, 
    Technische Fakultät,
    Universität Bielefeld,
    D-33619 Bielefeld,
    mchen (at)
  • Andreas Freier,
    Technische Fakultät,
    Universität Bielefeld,
    D-33619 Bielefeld,
    afreier (at)
  • Matthias Lange, 
    IPK Gatersleben,
    D-33619 Bielefeld,
    lange (at)
  • Jacob Köhler, 
    Technische Fakultät,
    Universität Bielefeld,
    D-33619 Bielefeld,
    jkoehler (at)


Background of project:

Although substantial progress has been achieved within the application of database and internet technology in the field of molecular biology, the problem of integrating distributed data as a preprocessing step for the analysis of genomic datasetsis still a challenging task. Actually, there are hundreds of distributed and heterogeneous databases [Galperin], where genomic information is stored. However, existing file formats, interfaces and databases change over time, which makes it a permanent problem to access and integrate the recent datasources. To overcome this, universal and integrative tools are needed.

For the modelling and analysis of metabolic, protein and gene gene networks, access to different pathway, protein interaction, gene regulatory and other databases is essential. Partially overlapping datasets have to be fused in order to integrate networks of different types. For this reason, it is a typical data integration task. 

In the broader sense, data integration is only one of the methods required for integrative modelling. At least, biochemical networks can be represented by different mathematical models. A lot of tools exist to design biochemical models interactively. The METABOLIKA simulation tool [Hofestaedt et al. ] is based on formal languages. GON [Nagasaki et al.] is a representative system with a powerful graphical interface which uses the Hybrid Petri Nets (HPN) approach.  Addicted to the dynamics of the field of application, a systematic approach is required for the automated derivation of models from distributed databases. 


General references related to project:

Galperin, M. Y. Molecular Biology Database Collection: 2004 update. 
Nucl. Acids. Res. 32(90001): D3--22, (2004).

Hofestaedt, R. and Meineke, F. Interactive Modelling and Simulation of Biochemical Networks. 
Computers in Biology and Medicine, 25(3):321--334, (1995).

Nagasaki et al. Genomic Object Net: I. A platform for modelling and simulating biopathways. 
Applied Bioinformatics 2(3): 181--184, (2003).


Aims of project:

The aim of the project was to develop a process to integratively design models of metabolic networks. Furthermore, a software platform needed to be implemented, providing generic tools to apply the developed methods to user-specific biological tasks. Our project places emphasis on two points:

  1. Data integration
    Data integration is a user-specific and problem related task. In the process of modelling and animation of regulatory networks data integration is used to acquire information about the structure of regulatory networks and the properties of biochemical substances. A universal integration service should be developed, which is capable to fuse data from distributed and heterogeneous molecular data sources. The implemented integration service should allow the configuration of pluggable datasources and user-defined schemata. 

  2. Modelling and Animation 
    A lot of different tools exist to simulate biological networks. Unfortunately, there is a lack of automated modelling capabilities, wherefore the user has to start modelling mostly from scratch. Consequently, the gap between the size of molecular databases and the size of networks modelled is growing even larger. Our aim is to combine data integration with modelling techniques to automate the modelling process. 


Description of project results:

In the center of our approach we developed a process of what we call Integrative Modelling of Biochemical Networks. As an implementation of the process, we implemented the software plattform MARGBench, which is named after this project. Figure 1 shows the projects overall architecture.

Figure 1

MARGBench includes the following four modules:

  1. BioDataServer
    Modelling starts at the fusion of data from heterogeneous data sources. BioDataServer (BDS) is an integration service based on the mediator approach. The tool (see figure 2) remotely accesses distributed datasources using specifically implemented types of wrappers. Additional wrappers can be plugged into the system supporting new types of datasources. 
    The user management of the system serves its functionality to different authorized users, allowing them to dynamically define schemata of virtual databases. When the user queries a virtual database, the query is decomposed into several subqueries. BioDataServer synchronizes the different tasks and joins the retrieved results. Standardized interfaces (JDBS/ODBC/XML) have been implemented to integrate the software into external applications.

    Figure 2

    Data fusion can be supported by semantic data integration. The fact that the fusion of data from different databases often leads to semantic conflicts, e.g. homonyms and synonyms, makes it hard to model integrated schemata. The tool SEMEDA is a Semantic Meta Database, which is directly connected to the BioDataServer. 
    It enables the user to query datasources using onthologies as well as controlled vocabularies. The system also supports collaborative editing of these elements, which makes it a powerful tool for the semantic annotation of molecular databases.

    Figure 3


    The interactive modelling environment PROTON has been developed to reconstruct models from databases. The system is connected to the BioDataServer as well as native relational databases. It enables the user to model three aspects of regulatory networks: (1) static relationships (semantic modelling), (2) dynamic processes (mechanistic models) and (3) systems (dynamic behavoir). Each layer (see figure 4) provides several methods for its data-driven reconstruction and its analysis. 


    Figure 4

    Configurable tools have been developed to apply data integration and database queries to the acquisition of molecular objects and the investigation of complex networks of relationships between them. A view concept allows interpretation of existing datasets as networks of molecular reactions of different types. Graphical methods provide the user with support in modelling the structure of regulatory networks. Finally, a simulation core provides the animation of the dynamic behavior of the structures as systems.

  4. PathAligner
    PathAligner is designed to reconstruct/retrieve metabolic pathways and process alignments of them. It provides a simple interface to reconstruct metabolic pathways from rudimentary components such as metabolites, enzymes, genes, sequences, etc. It provides a navigation platform to investigate more related genetic/metabolic information. It also provides an alignment method to compare the similarity of metabolic pathways. PathAligner graphically presents the retrieved pathways and the alignment results (figure 5).


    Figure 5

References to publications of project results:

Chen M., Lin S. und Hofestädt R.: STCDB: Signal Transduction 
Classification Database, Nucleic Acids Research, 2004, 32(1): 456-458

Chen M. und Hofestädt R.: WEB-Based Information retrieval System for the 
Prediction of Metabolic Pathways. IEEE Transaction on Nanobioscience, 
2004, Vol.3, No. 3, 192-199

Ming Chen and Ralf Hofestaedt. 
Quantitative Petri net model of gene regulated metabolic networks in the cell.
In Silico Biology. 3, 0030 (2003).

Köhler, S. Philippi and M. Lange. 
SEMEDA: Ontology Based Semantic Integration of Biological Databases
Bioinformatics, Volume 19(18), pp. 2420-2427 (2003).

A. Freier, R. Hofestädt and M. Lange. 
Integrative analysis of gene networks using dynamic process pattern modelling.
Bioinformatics of Genome Regulation and Structure. R. Hofestaedt, N. Kolchanov (eds.), Kluwer Academic Publishers (2003).

A. Freier, R. Hofestädt and M. Lange. 
iUDB: An object-oriented system for modelling, integration and analysis of
gene controlled metabolic networks. In Silico Biology, 3 (0019), IOS Press, (2003).

A. Freier, R. Hofestädt, M. Lange and U. Scholz. 
Information Fusion and Metabolic Network Control 
In J. Collado-Vides and R. Hofestädt, Herausgeber, Gene Regulation and Metabolism - Post-Genomic Computational Approaches. 
Cambridge, MA: MIT Press, (2002).

A. Freier, R. Hofestädt, M. Lange, U. Scholz und A. Stephanik. 
BioDataServer: A SQL-based service for the online integration of life science data.
In Silico Biology, 2(0005), (2002).

Ming Chen, Andreas Freier, Jacob Köhler, Alexander Rüegg. 
The Biology Petri Net Markup Language.
In: Jörg Desel, Mathias Weske (Hrsg.), Lecture Notes in Informatics - proceedings of Promise'2002, 
Oct. 9-11, Potsdam, Germany, Vol. 21: 150-161 (2002).

Ming Chen. 
Modelling and Simulation of Metabolic Networks: Petri Nets Approach and Perspective
In the proceeding of "ESM 2002, 16th European Simulation Multiconference" , 2002, June 3-5, Darmstadt, Germany, 441-444.

A. Freier, R. Hofestädt und M. Lange.
The electronical infrastructure for the analysis of metabolic networks.
In Abstracts' Book of the Workshop on CORBA and XML: Towards a Bioinformatics Integrated Network Evironment, NETTAB 2001 - Network Tools and Applications in Biology, S. 38-43, (2001).

R. Hofestädt, K. Lautenbach und M. Lange, Herausgeber.
Modellierung und Simulation Metabolischer Netzwerke: DFG-Workshop im Rahmen des DFG-Schwerpunktes Informatikmethoden zur Analyse und Interpretation großer genomischer Datenmengen, Magdeburg, 19. - 20. Mai 2000, Preprint Nr. 10. Fakultät für Informatik, Universität Magdeburg, (2000).


Organized Workshops (DFG supported):

  1. "Metabolic Pathways"
    Hofestädt R und K.Lautenbach, Magdeburg (2000) - under /Events/2000

  2. "Integrative Bioinformatics"
    Hofestädt R., Kolchanov N. und T. Dandekar, Bielefeld (2003) - under /Events/2003


Links related to project:

Print optimized Version