Genetic analysis and improvement of pearl millet for rust resistance and grain yield in Uganda.
Pearl millet is a sustainable food security crop for people living in areas with extreme drought and heat conditions. Like in many countries where it is grown, in Uganda the crop grows in semi-arid zones characterised by low average annual rainfall and hot conditions. Despite saving many from hunger, not much is known about the crop. Studies were therefore initiated to generate information on the production characteristics of pearl millet in Uganda, genetic improvement and to assess gene action for grain yield and rust resistance. A survey and experiments were conducted in two predominantly pearl millet growing areas with four objectives: 1) to establish production determinants of the pearl millet cropping system with related uses, constraints and desirable traits, 2) to determine the response to S1 progeny recurrent selection for rust resistance and grain yield in two local populations, 3) to study the inheritance and gene action for grain yield and rust resistance in improved pearl millet germplasm, and 4) to determine the stability of improved pearl millet lines and crosses for grain yield and rust resistance. The first objective was achieved by conducting a participatory rural appraisal in four districts (Kumi and Katakwi in the East, Kitgum and Lamwo in the North), where data was collected from 140 households. The second objective was achieved by subjecting two commonly grown rust susceptible populations (Omoda from East and Lam from North) to two cycles of phenotypic S1 progeny recurrent selection and the cycles evaluated in randomised complete block design with three replications, three locations and one season. Objective three was achieved by crossing six rust resistant male parents with ten rust susceptible female lines in a North Carolina design II mating design. The parents and crosses were evaluated in four environments in a 4 x 19 alpha-lattice design. Additionally, data from the same experiment were used to achieve objective four. The survey findings indicated that pearl millet was mainly grown for food and income. The production environment was low input, where farmers used family labour, planted unimproved genotypes and used neither chemicals nor manure to enhance productivity. Majority of the households had minimal access to credit, agricultural trainings or extension services. The households identified the most desirable variety traits that enhanced yield as stay green, tall, high tillering, high yield, early maturity and ergot resistant. The constraints that reduced yield were ergot and rust susceptibility, short genotypes, low yielding, low tillering, late maturity, and sterile panicles. Lack of markets, low prices and price fluctuation were the most important market constraints. Farmers also lacked knowledge about rust. Regression analysis showed that area planted, age of spouse and experience in pearl millet cultivation were the most important factors enhancing grain yield, while age of the household head, amount of seed planted and distance to the market negatively affected grain yield. The findings from the recurrent selection suggested a possibility to improve grain yield and rust resistance of locally adapted populations through two cycles of phenotypic S1 progeny recurrent selection. The Lam population responded faster than the Omoda population leading to respective net genetic gains of 72% and 36%, respectively. The effect was an increase in grain yield from 611 kg ha-1 to 1047 kg ha-1 in Lam population and 693 kg ha-1 to 943 kg ha-1 in Omoda population. The genetic gain for rust resistance was -55% and -71% achieved in populations Lam and Omoda, respectively, leading to improvement in rust resistance from 30% to 14% infection in Lam and 57% to 17% infection in Omoda. Results of genetic analysis indicated predominance of additive gene action for grain yield and rust severity at 50% physiological maturity while non-additive gene action was predominant for area under disease progress curve. Better-parent heterosis was significantly high for all the traits. For better-parent heterosis the top ranked crosses had heterosis of 12%-28%. One cross (ITMV8001 x SDMV96053) performed exceptionally well with better-parent heterosis of 93%. The better-parent heterosis for rust severity at 50% physiological maturity was higher than the better-parent heterosis for area under disease progress curve for all the crosses. The genotype by environment (GxE) interaction was important for grain yield, rust severity at 50% physiological maturity and area under disease progress curve. The GGE biplot identified the crosses ICMV3771 x SDMV96053 and Shibe x Okollo as the winners for grain yield and rust resistance, respectively. These crosses will be advanced in the pearl millet improvement programme in Uganda.