Our research focuses on the representation, analysis, and visualisation of biological networks in their spatial and temporal embedding under consideration of related multimodal and multidimensional data. Goal of these research activities is to support the knowledge generation process in the life sciences, in particular in plant bioinformatics and plant systems biology. For an integrative, systems biology directed approach it is not sufficient to consider the biological entities (e.g. DNA, RNA, proteins, metabolites) alone but is necessary to study their interactions and link the data from experimental groups to relevant networks (e.g. regulatory, interactomics, and metabolic networks). These networks are embedded in spatial (from cellular compartments to tissues and organs) and temporal (from enzymatic reactions to developmental processes to evolution of networks) environments. Therefore we research and develop algorithms, methods and systems in areas such as databases and information systems (to represent networks as well as high-throughput data), data integration (to combine complex data and networks), information and scientific visualisation (to help explore the data/networks in 2D to 4D), network analysis (to support hypothesis building), and modeling/simulation (to evaluate processes and support in silico predictions). We are interested in theoretical and practical problems from computer science and bioinformatics as well as in the underlying biological questions.

• Representation and Analysis of Experimental High-throughput Data
• Managing Metabolic Pathway Data
• Visualisation and Analysis of Networks containing Experimental Data
• Motifs in Networks
• Complex Phylogenetic Trees
• Centralities in Biological Networks
• Analysis and Simulation of Metabolic and Regulatory Processes
• Editing, processing and visualization of KGML pathway files
• Interactive Visualisation of Biological Networks

To deal with experimental data we are developing DBE (Data integration and analysis for Biological Experiments), a comprehensive information system for the analysis and visualisation of experimental data. It consists of five parts: (1) DBE-Gravisto, a network analysis and graph visualisation system, (2) the DBE-Website as the interface for the system, (3) the Excel-Importer application for the data import, (4) the DBE-Pictures application for the up- and download of binary (e. g. image) files, and (5) the DBE-Database for consistent data storage. The web-based interface allows an easy access to the different components of the system from computers in laboratories as well as in offices. The start page of the DBE system is here.

Publications: [BHK04]

In recent years, more and more attempts are being made to understand biological subjects at a systems level. A major resource for these approaches are biological databases, a variety of which has been designed to store manifold information. The use of these databases is often hampered by the fact that they are designed for special application areas and thus lack universality. Furthermore, the plant research field is often neglected. To circumvent these problems we have designed Meta-All, a software that allows to store and access information about metabolic pathways, including reaction kinetics, detailed locations, environmental circumstances and taxonomic information. Meta-All is a collaborative project of three IPK research groups, further information about the system can be found here.

Publications: [WGK06]

The aim of this project is the development of a software tool for visual exploration and statistical analysis of complex biochemical data sets. For this task it is necessary to integrate and process data from different areas of genome, proteome and metabolome research and to present the results in a user-friendly way. The emphasis is on the linkage of experimental data, for example containing metabolite concentrations and expression profiles, with metabolic and regulatory networks. The new system called VANTED (Visualisation and Analysis of NeTworks containing Experimental Data) makes it possible to load and edit graphs, which may for example represent biological pathways. It is possible to map experimental data sets onto the graph elements and visualise time series data or data of different genotypes or plants in the context of a network. The tool is available from the VANTED Homepage.

Publications: [JKS06a]

Motifs (patterns) in networks may represent building blocks of functional modules that carry out a relatively distinct function in biological networks. With MAVisto (Motif Analysis and VISualisation TOolkit) we are developing a framework for the analysis of motifs in networks. MAVisto is not restricted to the analysis of biological networks, but can be used to discover network properties in all kinds of networks. It contains an algorithm to recognise multiple appearances of motifs in the target network under different frequency concepts, a motif preserving layout algorithm, and navigation techniques to explore the underlying structure of the network given by the motifs. The tool is available from the MAVisto Homepage.

Publications: [SS05a,SS05b]

Complex phylogenetic trees (e.g. trees of metabolic pathways) represent multiple aspects of similarity and hypothetical evolution in a single, yet complex structure that is difficult to understand and interpret. We developed methods for analysing and visualising sets of related networks and complex phylogenetic trees. Based on the WilmaScope system we built a visualisation system that facilitates analysis of such structures by presenting multiple coordinated perspectives simultaneously. Several algorithmic problems such as the optimal leaf ordering in stacks of phylogenetic trees, similarity-preserving orderings of related networks and locally optimal levelling in hierarchical drawings have been considered and new algorithms for these problems have been developed.

Publications: [BDS04b,DS04]

Several mathematical methods support the analysis of network structured data and help to uncover important properties of networks. The ranking of network elements (vertices and edges) using centralities is such a method and we have developed CentiBiN (CENTralities In BIological Networks) to assist the centrality based analysis of networks. CentiBiN is a tool for the computation and comparison of different network centralities. It is used in our ongoing work on evaluating centrality measures for specific biological networks. The tool is available from the CentiBiN Homepage.

Publications: [JKS06d,KS04]

Usually biological models are incomplete and the simulation and prediction of biological processes is difficult. To support the creation, incremental development, simulation and evaluation of biochemical pathway models we built SyBME (SYstem Biology Modelling Environment). Based on the pathway topology retrieved from a database or edited by a user, several models with different parameter sets can be simulated using simulation systems such as Gepasi or Jarnac. The results of different simulations can then be used to compare parameterised models and further to iteratively improve the model.

Publications: [JKS06b]

KGML-ED allows the dynamic exploration and editing of KEGG Pathway diagrams. It is a graphical network editor, that provides read- and write-support for the KGML (KEGG Markup Language) file format. Pathway files are loaded and transformed into a graph network which may be modified to fulfill user-specific needs (e.g. it is possible, to delete or add network elements, change labels and colors). Novel network exploration approaches are supported by the system as well.

Biological processes form large and complex networks. The automatic and interactive visualisation of these networks and the combination of such techniques with network analysis approaches can aid biologists in gaining new insights into the processes in cells. We are developing new visualisation algorithms, navigation, animation and interaction techniques to produce high quality visualisations of pathway and network diagrams, to compare networks visually, to support the exploration of complex structures such as typed protein-protein interaction networks, network-based phylogenetic trees and clustered networks, and to visualise network-related data. Besides theoretical questions we are interested in the practical usability of the developed methods. We have implemented several methods and have already conducted some user-studies.

Publications: [DHK06,KKS05,BDS04b,BFP04,DRS04,FS04,BDS03,FS03,S03b]