Water stress is recognized as the most severe abiotic stress for agricultural productivity. Root traits play a key role in tolerance to water stress but have largely been neglected in selection schemes. In order to identify the maize genetic bases of the root adaptive responses to water deficit (WD), we used a MAGIC mapping population of 400 lines based on the intercrossing of 16 genotypes. The fine phenotyping of the different genotypes was performed under contrasting water supply on the French root phenotyping platform (4PMI). On the 16 founder genotypes, in addition of phenotyping, we sampled different root tissues daily over 7 days after irrigation arrest and performed RNAseq. On the basis of these 448 transcriptomes, we identified 6945 differentially expressed genes between axial and lateral roots and in response to WD and inferred a regulatory gene network to identify transcription factors (TF). Using a hierarchical clustering, we split the network in 35 clusters homogeneous in their expression pattern. Fine analysis of individual cluster pointed out, without prior knowledge, already known FTs responding to WD and identified new candidates. Functional validation of Arabidopisis orthologues has been initiated and many genotypes have an altered root developmental response to in vitro osmotic stress. In parallel, the phenotyping and a transcriptomic analysis by RNAseq of the genotypes of the mapping population under optimal conditions and water deficit enabled a GWAS and an eQTL analysis. Both approaches identified polymorphisms in genes of interest and identified SNPs colocating near transcription factors also identified by the gene network approach. Taken together all the data identified candidate genes and alleles potentially controlling adaptive root development that can be interesting target for breeding. This work was supported by the European Research Council (ERC) (HyArchi to CM; grant agreement No 788553)