Breeding of finger millet for agronomic and nutritional traits and drought tolerance in Ethiopia.

Abstract

The changing climate, human population pressure, and lifestyle changes require new, improved, and adaptive plant genetic resources for value-added product development. Finger millet (Eleusine coracana (L.) Gaertn., 2n=4x=36, AABB) is a climate-resilient and highly nutritious small grain crop widely grown in the semi-arid tropics, including Ethiopia. It has multiple uses, including for food, feed, and the preparation of alcoholic and non-alcoholic beverages. However, finger millet is an under-utilized and under-researched crop with a mean yield of < 2.4 t/ha in Ethiopia despite a potential productivity of up to 6-8 t/ha. The yield gap is attributed to biotic and abiotic stresses, a lack of access to improved seeds and production inputs, and poor agronomic management practices. Recurrent and severe drought conditions are amongst the major production constraints of finger millet, limiting grain yield and quality. Therefore, the objectives of this study were: i) to systematically assess finger millet production systems in Ethiopia by identifying key production constraints, opportunities, and farmer-preferred traits and to inform ideotype and variety design for targeted breeding programs; ii) to evaluate and characterise genetically diverse Ethiopian finger millet accessions for agronomic performance, drought response, and grain nutritional compositions under contrasting moisture regimes, and to identify superior genotypes for use in breeding and/or production; iii) to analyse the population structure of the Ethiopian finger millet landraces and to identify robust marker-trait associations and quantitative trait nucleotides (QTNs) and haplotypes linked to key agronomic and nutritional and iv) to determine the combining ability, heterosis, and gene action of selected finger millet parents for economic traits in hybrid combinations to guide the selection of best families and new breeding populations. Separate but complementary research activities were conducted to attain the aforementioned objectives. In the first part of the study, a participatory rural appraisal (PRA) study was conducted in six selected districts of the Southern Nation Nationalities People Region (SNNPR) and Oromia Region in Ethiopia, using 240 and 180 participant farmers through a semi-structured questionnaire and focus group discussion, respectively. The result showed, finger millet was the most important crop in the study areas grown mainly for a combination of uses, including for food, feed and cash (reported by 38.8% of respondent farmers), food and feed (14.5%), food and cash (13.7%), food (11.5%) and food, cash, feed and construction material (9.7%). The main constraints to finger millet production in the study areas were drought stress (reported by 41.3% respondents), lack of improved varieties (12.9%), lack of financial resources (11.3%), small land holdings (10.8%), limited access to seed (10.0%), a shortage of fertilizers (5.4%), poor soil fertility (4.6%), shortage of draught power (1.3%), labour shortages (1.3%) and high labour costs (1.3%). The most important farmer-preferred traits in a finger millet variety were high grain yield, compact head shape, ‘enjera’-making quality, high marketability and early maturity, resolved through principal component analysis. The aforementioned production constraints and farmer-preferred traits are strategic drivers for enhancing finger millet productivity and should be incorporated into Ethiopia’s finger millet breeding and technology development. In the second study, 448 Ethiopian finger millet accessions were evaluated under field conditions for agronomic performance, drought responses and nutrient composition. The study was conducted across three moisture regimes: non-stressed (Arsinegelle: denoted as AN), moderately droughtprone (Maitsebri: SH), and severely drought-stressed (Meiso: MI). Significant variations were recorded among genotype, environment, and genotype × environment interaction for all assessed traits. Grain yield (GY) declined by ~60% at MI site. The test genotypes varied considerably at MI for days to 50% flowering (DTF) (78.5–107.5 days), days to maturity (DTM) (101.0–149.5 days), GY (0.5-3.2 t/ha), iron (Fe) (53.6 – 81.0 ppm) and zinc (Zn) (67.8-83.1 ppm) contents suggesting considerable genetic variation for selection. Black‐seeded genotypes maintained higher GY and Fe/Zn under drought, while red‐seeded types flowered and matured earlier and produced larger seed size. Broad‐sense heritability was > 60% for drought tolerance (Drt) score, DTF, and DTM, but <30% for plant height, starch content and GY. GY was negatively correlated with DTF, DTM, Drt and stay green traits, but positively correlated with seed weight across environments. Principal component analysis explained >67% of the variance for some traits across test sites, and hierarchical clustering grouped genotypes into four clusters. Eight accessions G141 (219819), G423 (AxumColl37), G297 (238326), G247 (237449), G171 (229726), G204 (234176), G294 (238323), and G46 (211473) were identified as promising candidates for breeding and further validation through multi-environment testing and variety evaluation procedures prior to potential release. In the third study, 448 finger millet genotypes were evaluated across three contrasting environments to characterise their agronomic performance, drought response, and grain micronutrient compositions. A subset of 391 genotypes from this panel was skim-sequenced, generating 24,112 high-quality Single Nucleotide Polymorphisms (SNPs) that were employed for population structure, association mapping and haplotype analyses. Genome-wide association study (GWAS) resolved 70 marker-trait associations, including 15 major QTNs with more than 30% phenotypic variance explained (PVE) for all traits except stay-green score and grain yield. Pleiotropic major QTNs were identified for 50% days to flowering/days to maturity and Fe/Zn on chromosomes 9B and 2B, respectively. Haplotype analysis of major QTNs identified 54 significant haplotype blocks and 2 additional haplotypes, many of which originated from landraces collected in Tigray and Amhara regions. These findings provide invaluable insights for targeted breeding to enhance finger millet resilience, nutritional profile, and grain yield. The final part of the study assessed the combining ability, heterosis, and gene action of selected finger millet parents for economic traits in hybrid combinations. A 7 × 10 line × tester mating design was used, and the resultant 70 F1 hybrids and 17 parents were evaluated under well-watered (NST) and drought-stress (ST) conditions across greenhouse and field environments. Significant variations (p < 0.001) were observed among lines, testers and their interactions for the assessed traits, indicating both additive and non-additive gene action underlying inheritance of the assessed traits. Additive genetic effect predominantly conditioned the inheritance of DTF, ear length (EL), DTM under ST conditions, whereas non-additive effects affected the number of tillers (NT) and DTM under NST conditions. Broad-sense heritability for GY was higher under ST (0.51) than NST (0.13) conditions, necessitating multiple testing environments for drought tolerance evaluation and selection. Lines such as 235700, 234187 and Addis-01 and testers such as Gute, 24394, Bako-09, Mereb-1, and 215982 exhibited higher general combining ability effects for GY in a desirable trend. Crosses Wama/Kako-1, 234187/24394, Wama/215982, 235700/213835, Addis-01/213835, Addis- 01/Gute, Necho/Gute, and Necho/Kako-1 exhibited higher specific combining ability effects and heterosis for GY under ST conditions. The selected parents and the top crosses are recommended for breeding and selection of new generation and drought-adapted finger millet genotypes. Overall, the present study identified the major production constraints, traits of economic value of finger millet, identified unique genetic resources and genes and developed new breeding populations with better agronomic performance, nutrient compositions and drought tolerance. The new breeds have integrated farmers’ preferred traits and will be evaluated across multiple sites for selection and release in the marginal production areas in Ethiopia and similar agro-ecologies.