アブストラクト
| Date | June 19, 2006 |
| Speaker | Claire Gaugain, Centre de Bio-informatique de Bordeaux Universite de Bordeaux 2 |
| Title | Integration and representation of biological data |
| Abstract | Functional genomics approaches should help us to better understand
how the cell works by highlighting the relationships between molecular
mechanisms and cellular functions. They produce large amount of highly
heterogeneous data that need to be integrated. This integration has
become a major challenge in bioinformatics. An example, which typically
illustrates this need for integration, is the interpretation of
expression profiles obtained using microarray experiments. Once
clustering methods have been used to define groups of potentially
co-regulated genes, one would like to find out if there is any known
information (such as those contained in the annotations) which already
bring theses genes together in a significant manner.
In this aim, various methods and tools have been developed. Among them, BlastSets (http://cbi.labri.fr/outils/BlastSets/) proposes a general strategy which consists in building, and collecting in a database, sets of genes attached to a biological information (i.e. set of genes that are annotated as being involved in a given biological process). The composition of these sets can then be compared with the composition of any query set (such as a group of coregulated genes) in order to identify correlation. Our results show that the effectiveness of this approach strongly relies on the choice of the method used to build sets that capture biological information. |