Project: BC-GeneExMiner (team project)
BC-GenExMiner is a public webtool developped by the team for statistical mining of gene expression data measured by transcriptome screening of breast tumors. This web tool is intended for researchers and oncologists involved in basic research or transfer projects in cancerology. It allows 1) to overcome difficulties related to the lack of expertise in statistical analysis (tests, normalization...), in the manipulation of large databases and in the production of figures, and 2) to obtain robust results thanks to the use of various and/or large cohorts of patients.
BC-GenExMiner allows:
- to validate results on isolated or grouped cohorts, sorted according to clinical-pathological and molecular criteria ;
- to refine knowledge about the prognostic informativeness of a gene or its expression according to various clinical-pathological and molecular parameters;
- to prioritize genes of interest;
- to make discoveries intuitively or by serendipity.
The gene expression measurements contained in the database were produced by hybridization on DNA chips (microarrays) and by high-throughput sequencing (RNA-seq). A careful screening of the cohortes was performed before inclusion. bc-GenExMiner is composed of three analysis modules: a "Correlation" module, an "Expression" module and a "Prognosis" module, each allowing various types of analyses. These are detailed in this flow chart. The results are displayed through tables and several types of figures: heatmap and correlation scatterplot, whisker box, bee swarm plot, violin plot, rain cloud plot and Kaplan-Meier curve.