Data Integration and VISualization (DIVIS) : from large heterogeneous datasets to interpretable visualisations in plant science

Abstract : The demand by biologists to integrate heterogeneous and large datasets from "omics" and phenotyping activites is rapidly increasing. However, methods automating this approach are still at its infancy and to our knowledge, no operational and user-friendly software yet exists. Experiments are performed independently and resulting data are cross-analysed manually and a-posteriori by scientists. For instance, the biology teams from the IRHS (Institut de Recherche en Horticulture et Semences) in Angers have been accumulating datasets of different natures (transcriptomic, biochemistry, physical measures, sensory analysis, etc.) regarding perennial, annual and biannual plants. These datasets are described using reference ontolo-gies enriched with in-house knowledge and stored in a Laboratory Information Management System (LIMS) which is developed and distributed by the IRHS Bioinformatics team. The main objective of the DIVIS (Data Integration and VISualization) project is to develop a directly usable prototype of a new data analysis tool, by combining the most promising integration and visualisation approaches that are publicly available using the heterogeneous large scale datasets stored in our LIMS.