Abstract only

Pearl millet is a multipurpose grain/fodder crop of the semi-arid tropics, feeding many of the world?s poorest and most undernourished people. They depend on cereal- and legume-based diets, as they have limited access to meat and dairy products, fruits and vegetables. Genetic variation among adapted pearl millet inbreds and hybrids suggests it will be possible to improve grain micronutrient concentrations by selective breeding. We mapped QTLs for grain [Fe] and [Zn] using replicated samples of 106 pearl millet RILs derived from ICMB 841-P3 x 863B-P2, which segregate for these and many other traits. Skeletonmapping with 104 SSR markers detected seven linkage groups covering 1557 cM (Haldane). Self-pollinated grain samples were collected from plots of RILs, their parents and additional controls in a 120-entry, 3-replication alpha-design field experiment sown on an Alfisol at ICRISAT-Patancheru late in the 2009 rainy season. Atomic Absorption Spectroscopy determined mineral composition of ground grain samples. Plot data for Fe and Zn concentrations, 50% flowering time (FT), plant height (PH), panicle length (PL), and thousand grain mass (TGM) were analyzed by Residual Maximum Likelihood. Repeatabilities for all traits exceeded 0.7. Among RILs, grain mineral concentration ranges were 34.3-99.5 ppm for [Fe], and 36.3-90.3 ppm for [Zn]. Similarly, ranges for FT (36.9-53.2 d), PH (79-130 cm), PL (14-22 cm), and TGM (6.13-11.70 g) were substantial. Correlation between [Fe] and [Zn] was significantly positive (+0.86**), but both mineral concentrations exhibited significant negative correlations with TGM (-0.4**). RIL BLUPs employed for QTL detection (by Composite Interval Mapping using PlabQTL) detected seven putative QTLs for grain mineral concentration ? five for [Fe] (LG2, LG3, LG5, LG6 and LG7) explaining 6, 23, 10, 14 and 11% of observed phenotypic variation, respectively (39% adjusted R2), with two overlapping those for [Zn] (LG3 and LG6) explaining 34 and 13% of observed phenotypic variation, respectively (40% adjusted R2). Four favorable alleles were from Iniari landrace-derived 863B and one (LG6) was from ICMB 841-P3. Adjusted R2 values for putative QTLs for FT (7), PH (4), TGM (3), and PL (2) were 54, 43, 40 and 21%, respectively. The LG3 QTL region will be validated by marker-assisted backcrossing.

Global Wheat Program

English

INT2983

CIMMYT Staff Publications Collection