We are designing new data mining techniques on gene expression data, more precisely inductive querying techniques that extract a priori interesting bi-sets, i.e., sets of objects (or biological situations) and associated sets of attributes (or genes). The so-called (formal) concepts are important special cases of a priori interesting bi-sets in derived boolean expression matrices, e.g., matrices that encode over-expression of genes. It has been shown recently that the extraction of every concept is often possible from typical gene expression data because the number of biological situations is generally quite small (a few tens). In specic applications, we have been able to extract every concept and it can lead to millions of concepts. Obviously , post-processing these huge volumes of patterns for the discovery of biologically relevant information is challenging. It is useful since the added-value of transcription module discovery is very high and formal concepts can be seen as putative transcription modules. We describe our on going research on concept post-processing by means of classication and visualization. It has been applied to a real-life gene expression data set with a promising feedback from end-users.