Résumé
Pearl millet is an important staple food for farming communities across semi-arid tropical systems of South Asia and Sub-Saharan Africa where production suffers uncertain precipitation. This work is undertaken under the premise that maximizing grain yield under water-limited conditions depends on both maximizing water use and ensuring water availability for the grain filling period. Here we discuss the phenotyping methods targeting the variability in plant water use strategies which determine the crop production success in water-limited environments. A fine-mapping population of pearl millet, segregating within the previously identified drought tolerance quantitative trait locus (QTL) on chromosome 2 (LG02), was tested across different experimental environments (pot culture, high-throughput phenotyping platform (LeasyScan), Lysimeter, and Field). Recombinants were then analyzed for traits at different levels of plant organization, ranging from water-use traits (transpiration rate, leaf area, plant organ dry weights, etc.) to crop production and agronomic traits (grain yield, tiller number, harvest index, etc.) The linkages between traits across the experimental systems were analyzed, using principal component analysis (PCA) and QTL co-localization approach. The functional relevance of the phenotyping systems was traced by PCA analysis. Furthermore, we found four regions within the LG02-QTL underlying substantial co-mapping of water-use related and agronomic traits. These regions were identified across the experimental systems and justified linkages between water-use traits were phenotyped at lower level of plant organization to the agronomic traits assessed in the field. Therefore, the phenotyping systems at ICRISAT are validated and well set to accelerate crop breeding for drought adaptations.