The increased occurrence and severity of drought stress have led to a high yield decline in rice in recent years in drought-affected areas. Drought research at the International Rice Research Institute (IRRI) over the past decade has concentrated on direct selection for grain yield under drought. This approach has led to the successful development and release of 17 high-yielding drought-tolerant rice varieties in South Asia, Southeast Asia, and Africa. In addition to this, 14 quantitative trait loci (QTLs) showing a large effect against high-yielding drought-susceptible popular varieties were identified using grain yield as a selection criterion. Six of these (qDTY_1.1, qDTY_2.2, qDTY_3.1, qDTY_3.2, qDTY_6.1, and qDTY_12.1) showed an effect against two or more high-yielding genetic backgrounds in both the lowland and upland ecosystem, indicating their usefulness in increasing the grain yield of rice under drought. The yield of popular rice varieties IR64 and Vandana has been successfully improved through a well-planned marker-assisted backcross breeding approach, and QTL introgression in several other popular varieties is in progress. The identification of large-effect QTLs for grain yield under drought and the higher yield increase under drought obtained through the use of these QTLs (which has not been reported in other cereals) indicate that rice, because of its continuous cultivation in two diverse ecosystems (upland, drought tolerant, and lowland, drought susceptible), has benefited from the existence of larger genetic variability than in other cereals. This can be successfully exploited using marker-assisted breeding.

Background: Drought is one of the most important abiotic stresses causing drastic reductions in yield in rainfed rice environments. The suitability of grain yield (GY) under drought as a selection criterion has been reported in the past few years. Most of the quantitative trait loci (QTLs) for GY under drought in rice reported so far has been in the background of low-yielding susceptible varieties. Such QTLs have not shown a similar effect in multiple high- yielding drought-susceptible varieties, thus limiting their use in marker-assisted selection. Genetic control of GY under reproductive-stage drought stress (RS) in elite genetic backgrounds was studied in three F_3:4 mapping populations derived from crosses of N22, a drought-tolerant aus cultivar, with Swarna, IR64, and MTU1010, three high-yielding popular mega-varieties, with the aim to identify QTLs for GY under RS that show a consistent effect in multiple elite genetic backgrounds. Three populations were phenotyped under RS in the dry seasons (DS) of 2009 and 2010 at IRRI. For genotyping, whole-genome scans for N22/MTU1010 and bulked segregant analysis for N22/Swarna and N22/IR64 were employed using SSR markers. Results: A major QTL for GY under RS, qDTY _1.1 , was identified on rice chromosome 1 flanked by RM11943 and RM431 in all three populations. In combined analysis over two years, qDTY _1.1 showed an additive effect of 29.3%, 24.3%, and 16.1% of mean yield in N22/Swarna, N22/IR64, and N22/MTU1010, respectively, under RS. qDTY _1.1 also showed a positive effect on GY in non-stress (NS) situations in N22/Swarna, N22/IR64 over both years, and N22/MTU1010 in DS2009. Conclusions: This is the first reported QTL in rice with a major and consistent effect in multiple elite genetic backgrounds under both RS and NS situations. Consistency of the QTL effect across different genetic backgrounds makes it a suitable candidate for use in marker-assisted breeding.