Sorghum (Sorghum bicolor [L.] Moench) is the fifth most important cereal crop worldwide and second after maize (Zea mays L.) in Kenya. It is an important food security crop in arid and semi-arid lands, where its production potential is hampered by drought. Drought tolerance can be measured by a plant's ability to resist premature senescence, often described as stay-green. This study was carried out with the objective of identifying novel stay-green trait among wild and landrace genotypes of sorghum. Forty-four sorghum genotypes that included 16 improved, nine landraces, and 17 wild relatives of sorghum alongside known stay-green sources, B35 and E36-1, were evaluated under well-watered and water-stressed conditions in an alpha-lattice design of three replications. Data was collected on plant height (PHT), flag leaf area (FLA), panicle weight (PWT), 100-seed weight (HSW), relative chlorophyll content (RCC), number of green leaves at maturity (GLAM), days to 50% flowering (DFL), and grain yield (YLD). Genetic diversity was determined using diversity arrays technology (DArT) sequencing and quality control (QC) markers were generated using a java script. Lodoka, a landrace, was the most drought-tolerant genotype, recorded the highest numbers of RCC and GLAM, and outperformed B35 and E36-1 in yield under water-stress and well-watered conditions. The RCC was highly correlated with GLAM (r = .71) and with yield-related traits, HSW (r = .85), PWT (r = .82), and YLD (r = .78). All traits revealed high heritability (broad-sense) ranging from 60.14 to 98.4% for RCC and DFL, respectively. These results confirm earlier reports that wild relatives and landraces are a good source of drought tolerance alleles.

Striga hermonthica is the most important parasitic weed in sub-Saharan Africa and remains one of the most devastating biotic factors affecting sorghum production in the western regions of Kenya. Farmers have traditionally managed Striga using cultural methods, but the most effective and practical solution to poor smallholder farmers is to develop Striga-resistant varieties. This study was undertaken with the aim of identifying new sources of resistance to Striga in comparison with the conventional sources as standard checks. We evaluated 64 sorghum genotypes consisting of wild relatives, landraces, improved varieties, and fourth filial generation (F4) progenies in both a field trial and a pot trial. Data were collected for days to 50% flowering (DTF), dry panicle weight (DPW, g), plant height (PH, cm), yield (YLD, t ha−1), 100-grain weight (HGW, g), overall disease score (ODS), overall pest score (OPS), area under Striga number progress curve (ASNPC), maximum above-ground Striga (NSmax), and number of Striga-forming capsules (NSFC) at relevant stages. Genetic diversity and hybridity confirmation was determined using Diversity Arrays Technology sequencing (DArT-seq). Residual heterosis for HGW and NSmax was calculated as the percent increase or decrease in performance of F4 crossover midparent (MP). The top 10 best yielding genotypes were predominantly F4 crosses in both experiments, all of which yielded better than resistant checks, except FRAMIDA in the field trial and HAKIKA in the pot trial. Five F4 progenies (ICSVIII IN × E36-1, LANDIWHITE × B35, B35 × E36-1, F6YQ212 × B35, and ICSVIII IN × LODOKA) recorded some of the highest HGW in both trials revealing their stability in good performance. Three genotypes (F6YQ212, GBK045827, and F6YQ212xB35) and one check (SRN39) were among the most resistant to Striga in both trials. SNPs generated from DArT-seq grouped the genotypes into three major clusters, with all resistant checks grouping in the same cluster except N13. We identified more resistant and high-yielding genotypes than the conventional checks, especially among the F4 crosses, which should be promoted for adoption by farmers. Future studies will need to look for more diverse sources of Striga resistance and pyramid different mechanisms of resistance into farmer-preferred varieties to enhance the durability of Striga resistance in the fields of farmers.

Recurring drought stress cycles and widespread striga (Striga hermonthica (Del.) Benth) infestations are two of the major constraints of sorghum production in sub-Saharan Africa (SSA) where they cause a crop loss of about 60 billion US dollars and affect a population of about 100 M people annually. Plant breeders continue to employ conventional and molecular crop breeding strategies in the search for durable genetic resistance/tolerance mechanisms or for germplasm with genes against these two constraints. Crop wild relatives and landraces remain valuable resources of resistance/tolerance genes and have been utilized in the past to improve tolerance to drought stress and resistance to striga. The aim of this study was to assess the stability of performance of 64 sorghum wild relatives, landraces and progenies from some generation of crosses under striga infested and drought stress conditions in agroecological environments endemic for these two stresses. The performance of the genotypes under drought stress was assessed in well-watered and in water stressed conditions at the Kenya Agricultural Livestock Research Organization (KALRO) Kiboko Research Centre whereas the same set was evaluated under striga artificially infested field and potted trials at the KALRO Alupe Research Centre during 2018/2019 rainy seasons. Genotypes, B35 × ICSV III N, Macia, N13, ICSV 111 IN, F6YQ212 × B35, SRN39, GENO47293, ICSV 111 IN × B35, IS9830, Framida, GENO 45827, F6YQ212, B35 × AKUOR ACHOT were found to maintain stable high yields in both striga and drought conditions. The results here, showed that Genotype, Genotype × Environment (GGE) interaction partitioned genotypes in two of the four mega-environments according to their stability and mean grain yield (GY) and identified representative genotypes of the two traits that could be exploited to develop superior sorghum varieties adapted to drought and striga prone environments.