MOLECULAR NETWORKS

The current paradigm of molecular biology is shifting towards the interpretation of data produced by high-throughput methods. The new data sources allow one to study system-wide properties in molecular terms. We are developing novel, generalized knowledge representation schemes for the study of the topological properties of molecular networks such as the regulatory networks in living cells and the similarity space of protein structures.

Structural descriptions in molecular sciences consist of entities such as atoms, residues, domains as well as (binary) relationships such as chemical bonds, sequential and spatial vicinity etc. The representation of genomic data follows the same rule, but the resulting molecular network descriptions (MND) greatly vary in size and contain a large number of elements, many of which are uncertain (Figure 1A). Similarly to molecular structures, MNDs can be studied in terms of patterns and similarities, but their most intriguing property is the analogy with networks found in other areas such as electrical and social networks, the Internet, etc. MNDs allow one to ask questions about stability, robustness, error and attack tolerance, which are relevant to biological or biochemical systems. Our group is interested in applying this general approach to protein folding, protein similarity analysis as well as the transcriptional regulatory networks of living cells.

Proteins, described either as sequences or as 3D structures, form a similarity space in which domain types appear as tightly connected clusters reminiscent of small world networks. The distribution of the random similarities that are found between unrelated groups, on the other hand, follows a scale free distribution, found in many other biological and non-biological networks (Figure 1B). Our group is involved in collecting domain sequence data. Results of a similarity search can be considered as adding a subnetwork to the existing network of similarities, and protein domain prediction schemes can be based upon the comparative analysis of the underlying similarity networks.

SYSTEMS BIOLOGY: GENERIC MODELS OF BIOLOGICAL SYSTEMS

Generic models of biological systems strive to capture the logical structure of biological processes. In the field of visual pattern recognition we develop simplified models of neural activities that can capture the differences between continuous/discontinuous, open and closed contours that are well distinguished by human vision (collaboration with the Cognitive Science Group of HAS-BME and with the Cognitive Science Department of BME, Budapest). In the field of bacterial communication we develop models that include chemical communication between bacterial cells – the chemical language – as well as competition between cells, which will allow us to describe the development of local structures in bacterial communities (collaboration with ICGEB, Trieste). Our long-term goal is to apply the principles of generic models in the interpretation of bioinformatics, i.e. in biological datamining.

PROTEIN FAMILY ANNOTATION

The identification of the Tribbles family members’ role as regulators of signal processing systems and physiological processes, including development, together with their potential involvement in diabetes and cancer has generated considerable interest in these proteins. Tribbles have been reported to regulate the activation of numerous intracellular signaling pathways, with roles extending from mitosis and cell activation to apoptosis and modulation of gene expression. Within the framework of an international collaboration with the Cardiovascular Research Unit of the University of Sheffield (UK) we are involved in the bioinformatic investigation of the Tribbles family using phylogenetics and functional protein pattern identification methods. Our current research activity is focused on the bioinformatic analysis of the syntheny, and the chromosomal regions where the genes are located.

GENOME BIOINFORMATICS

The Bioinformatics Group of BRC has special expertise in large-scale bioinformatics data management systems that have become an integrating force in systems biology, by providing common platforms and databases for different high-throughput experimental technologies. One of the major focus points of the ongoing research activity is the bioinformatic evaluation of experimental data from new transcriptome profiling approaches using next generation sequencing technology. We participate in an international collaborative project with the aim of identifying marker genes involved in the immune response of Zebrafish in different pathological conditions.

BIOINFORMATICS INFRASTRUCTURE RESOURCES

Most biologists today need some kind of assistance in accessing and analyzing the data available through the Internet and in keeping up with the new analysis methods. Bioinformatics is relatively new in the BRC, and our group has been involved in establishing and maintaining the biocomputing infrastructure. We have developed BRC BioNet, an HTTP-based intranet resource, which is accessible from any of the over four hundred PCs and workstations in the laboratories. BRC BioNet's tools include:

• EMBOSS bioinformatic softwaresuit (~100 programs)
• PISE graphical interface for EMBOSS package
• Library information and online access to scientific journals
• Medline alerting service
• Citation database
• A repository of local teaching materials
• Scientific Toolbox (periodic table, fact sheets about isotopes, vectors, restriction sites, etc.)
• A collection of more than a thousand WWW links of biological interest

We also provide institute-wide access to various molecular modeling tools including the SYBYL program package of Tripos Inc. covering the scientific discipline of molecular mechanics, molecular dynamic simulations, homology modeling, ligand docking and threading.