Our team gathers statiscians and bioinformaticians and our project aims at developing statistical models and bioinformatics approaches to improve the functional and relational annotation of Arabidopsis genes and to transfer this knowledge from the model plant to plants of agronomic interest. Our project is made in three parts
Prediction of gene networks involved in plant stress responses
The objective is to take advantage of the transcription re-programming that occurs in stressed plants in order to assign orphan-of-function genes to plant responses to environment and to predict relational annotations between genes involved in stress adaptation. A recent results is GEM2Net a new module of CATdb, which presents the analysis of 18 stress categories, in which 17 264 genes are involved and organized within 681 co-expression clusters. Functions were associated with these clusters by integrating various resources (GO, subcellular localization of proteins, Hormone Families, Transcription Factor Families and a refined stress-related gene list associated to publications) and gene networks were displayed by exploiting protein–protein and transcription factors-targets interactions.
Our objective is to pursue this work by integrating new sources of data in GEM2Net, by applying the proposed methodology to specific datasets in collaboration with other teams and by exploiting this co-expression analysis to improve the functional annotation of the genes.
Statistical methods for omics analysis
This project aims to develop original statistical methods oriented towards high-throughput technologies used in molecular biology. These methods are developed in close collaboration with the transcriptome platform of IPS2.
Skills are in model-based methods, linear models, algorithms for the analysis of omics data ...
This project allows us to investigate genome activity (gene expression and control mechanisms). Main topics are (i) Evaluation of normalization and differential analysis methods for microarray and HTS data, (ii) development of mixture models for co-expression network and protein-DNA interaction and (iii) statistical methods for regulation networks.
The team is involved in several integrative projects merging biological ressources, bioinformatics predictions and statistical analyses. Databases and graphical interfaces are relevant approaches to organize and visualize data integration efforts. For several years now, we develop and maintain two databases FLAGdb++ and CATdb which help to improve insight in the biological roles of plant genes and give us an ideal medium to release the results of our projects to a broader community. FLAGdb++ is a database on the structural and functional annotations of 6 plant genomes and CATdb is a database dedicated to data generated by the transcriptome platform of IPS2.