You are here: Home > For Researchers > Projects > Genetic Network Inference - Development of a generic methodology for genetic network inference

Genetic Network Inference - Development of a generic methodology for genetic network inference

From 10-01-2001 to 30-09-2006

Description

Bioinformatics research has focused mainly on the application of mathematical methods on high-dimensional data generated by high-throughput molecular and biological techniques. Part of those tools lack the necessary biological relevance. The future of bioinformatics lies in the design of more biology oriented algorithms. This interdiscplinary research project can be situated in this perspective. Concretely, the selection, implementation and optimization of existing mathematical methods for the identification and modeling of regulatory interactions is studied. Thanks to the accomplishments in high-throughput screening it is now possible to measure simultaneously the expression level of a large set of genes. This flood of data has to be processed with the right datamining and modeling techniques. Genetic network inference is a great challenge both from a biological and systems identification viewpoint. The development of a mathematical framework for inference implies robust preprocessing, optimized feature extraction, selection of the most appropriate model class and training procedure, and interpretation of the results. SCD will profit from the finetuning and optimization of existing methods. From a biological point of view, genetic network inference forms the basis of fundamental and applied research. Since we like to develop a generic method to model regulatory networks in procaryotes and eucaryotes, we opt to study the glycolyzation pathway in yeast and the Type III secretion system (TTSS) in Salmonella thypimurium.