EMI-CD - European modelling initiative combating complex diseases
Genome research has seen fundamental technical breakthroughs in recent years such as the sequencing of the human genome and the genome of other species serving as experimental model systems. The main sociological and economical impact of genome research is the molecular understanding of major human diseases and the development of new therapies and medicals for the combat of these diseases. However, despite the fact that there was a nearly three-fold increase of pharmaceutical investment in R&D in the time period 1992-2001 from 11.5-30.5 billion USD, the number of newly filed molecular entities has been fairly constant as pointed out by several pharmaceutical researchers recently (BIO 2003, PhRMA report 2001).
One possible reason for this development might be the fact that analytical methods and tools are not yet significantly installed in the drug development process. While bioinformatics methods are well incorporated in the first part of this process (drug target discovery), this is not the case for the later stages. In particular, the simulation and modelling of biological processes such as disease-relevant signaling pathways and metabolic processes are under-developed in drug target validation. Nevertheless, computational methods are needed here. In contrast to the early 90s where target discovery was a main problem, nowadays the number of potential drug targets has increased to a large extent leading to an unfeasible number of targets and to excessive costs in drug development. For example, the R&D costs per drug have increased from 95 million USD in 1982 to almost 880 million USD in 2000.
A fundamental challenge is thus, to search through this exhaustive set of targets and separate feasible from unfeasible ones. Here, in silico experiments can be the basis for a successful screening within the drug discovery process and the entire drug development process should be accompanied by bioinformatics and systems biology approaches especially by the introduction of simulation techniques and experimental design in all phases of the process. Furthermore, the need for integration rules and methods is fundamental in current functional genomics research Multiple databases exist, a variety of experimental techniques have produced gene and proteome expression data from various tissues and samples and important disease-relevant pathways have been investigated. Information on promoter regions and transcription factors is available for a lot of genes as well as sequence information. This information - although extremely helpful - cannot be utilized in a sufficient way because of the lack of integrative analysis tools.
In our project we attempt to develop a software platform that is able to meet some of the above requirements. The software platform bases on three layers and will connect and implement several modules necessary for the in silico modelling process. In the first layer information is gathered on the biological objects under analysis and experimental measurements on these objects are integrated. An analysis layer will translate this knowledge into biological networks. Using probabilistic learning methods (e.g. Bayesian networks) these networks will be expanded in light of all available data on the objects. In a simulation and modelling layer these network hypotheses will be evaluated and predictions of experiments will be produced which have a direct feed-back to the forthcoming experimental design and experimental verifications.
For more information, questions, etc. contact: Ralf Herwig
Last modified 2005-09-13 02:08 PM