Toward an integrated tool for high throughput functional analysis of plant genomes

Gene detection and structural annotations of plant model genomes is mainly done by bioinformatics groups running large sequences databases (essentially MIPS and TIGR). The raw material is the assembled genome sequences produced by standardized methods. Functional annotation is a more comple x process based on bioinformatic approaches, many heterogeneous data traditionally produced by biology laboratories and more recently by high throughput (HTP) genomic laboratories [1]. It is of critical importance to rapidly integrate the data issued from both laboratories. Local databases, managed by bioinformaticians immersed into a genomicresource-producing laboratory, are a good way to rapidly integrate new data generated either by biological or bioinformatics methods. Tools dedicated to the functional annotation of large eukaryotic genomes are still experimental and different options may be explored [2]. We describe, here, the current development of such a tool dedicated to plants and driven by the central idea of a coordinated functioning of the whole genome. Rules governing this global functioning of plant genomes are still largely unknown but progressively discovered. They should be used, as soon as possible, to explain both punctual and HTP results of functional analyses. We evaluate the interest to use a topological representation of genomes components onto the chromosomes to integrate those data with the knowledge on genome organization, compartmentation, variation and evolution. Integration of data for HTP functional analysis should respond to two major aims: a) Offering a genome wide representation for family of sequences for which regulatory function and expression have been experimentally demonstrated. For each sequence bearing information it would be useful to produce a statistical and weighted prediction of their information including their topological organization and an analysis of the different combinations they exhibit into the whole genome or in functional sub-groups. In a second step, the same representation may be done for regulatory sequences discovered by applying prediction algorithms to the whole genome space. b) Offering the genome context, based onto HTP experiments, of experimentally proven functional annotations. Biologists that want to consider their data into the context of HTP functional data – transcriptome, proteome, metabolome and interactome are facing a challenge particularly difficult due to the rapidly evolving and not firmly established HTP methods in general and data analysis in particular. Even more basic is the necessity for biologists to know the state of annotations and their rapid evolution, not only for the organism they are working on but also for all other organisms.