Genetic Variability of Inbred Lines, Genotype x Environment Interaction, Yield Stability and Prediction of Three-Way Crosses of Maize (Zea mays L.) in Southern and Eastern Africa

Abstract

Maize (Zea mays L.) is one of the most important food crops in eastern and southern Africa, but the productivity of this crop is limited by constantly changing climate variables, limited availability of high-yielding and stable hybrids, and limitation of understanding to create inbred lines to produce adaptable and stable hybrids in the region. Therefore, this research was carried out to assess the genetic diversity of inbred lines, the relative advantages and heterosis of three-way crosses over their respective single crosses, the genotype by environment interaction and grain yield stability, and the prediction of three way crosses of maize in eastern and southern Africa (ESA). Two experiments (genetic variability of inbred lines, and comparative evaluation of three-way cross with single crosses) were conducted using alpha lattice designs(10x6,and 6x5,respectively) in Ethiopia and the third experiment (GEI and yield stability) was conducted using alpha lattice design (9x5) in 52 locations in ESA in 2018/19 and 2019/20. The results of the genetic variability showed that the phenotypic coefficient of variation (PCV) was greater than the genotypic coefficient of variation (GCV) for all traits. Rows per ear had the lowest PCV (13.94.07%) and GCV (8.74%), but grain yield had the highest PCV (57.09%) and GCV (22.55%). In the cluster analysis, the inbred lines were grouped into three clusters. The greatest inter cluster distance was found between Clusters II (2.69) and III (3.15) suggests that inbred lines had more genetic diversity between the clusters (cluster II and III). The overall variance was divided into three PCAs, with the first three PCAs accounting for 75.25% of the total variation with Eigen values greater than unity. The genetic advance over mean (GAM) for the variables ranged from ear diameter (cm) having the lowest value (11.29%) to ear height (cm) having the highest value (30.28%). While grain yield and rows per ear had the lowest heritability (0.16) and GAM % (11.29), higher heritability and higher GAM % were obtained for plant height (0.65, 25.78%) and ear height (0.59, 30.28 %). On the other hand, single-cross hybrids showed a highly significant (P<1%) variance for grain yield, plant height, ear height, and ear length at three locations. These single cross hybrids also showed a very significant genotype by environment interaction (P< 1%) for grain yield, plant height, ear height, and kernel per ear. In comparison analysis, three-way xx crosses outperformed single crosses in 80% of crosses in Ambo, 73% of crosses in Abala Faracho, and 67% of crosses in Melkassa. In Ambo and Melkassa, single cross 1 (76%) and single cross 7 (104%) had the highest better parent and mid-parent heterosis, respectively. The maximum better and mid-parent heterosis of three-way crossings were found in Ambo TWC 14 (52%) and TWC 24 (78%) respectively, while the highest BPH and MPH found in Melkassa were TWC1 (56%), and TWC30 (25%) respectively. In the study of GEI and grain yield stability, the most stable hybrids among the top ten hybrids, according to ASV, JLR, and GGE analysis (mean vs. stability) were hybrids H37, H19, H14, and H25 under optimal conditions, H33, H19, H15 and H14 under random stress, H17, H25, and H22 under low nitrogen, and H15, H27, H33, and H35 under drought. Additionally, using a GGE-biplot model with a discriminative and representativeness selection technique, it was determined that E1 under optimal conditions, E31 under random stress, E17LN under low nitrogen, and E41 under drought were the representative environments for the breeding program. In the comparison of three-way crosses’ prediction, the prediction value of the Jenkins method B was calculated and the chi-square test, where the variation was not significant for all the traits, demonstrated its strong agreement between the predicted and observed values. In a similar manner, the prediction for the three-way cross was calculated by using the new formula, and its chi-square test revealed a good agreement between the expected prediction values and the observed values of grain yield and ear length for all three-way crosses. Conclusively, in the study of genetic variability, the moderate heritability and GAM for the traits of ear length, ear diameter, rows per ear, and kernel per row support selection via path coefficient analysis and multivariate analysis, where kernel per row (KPR) suggested being the promising traits for a breeding program in order to select superior genotypes. Moreover, from the random crossing of these inbred lines there was a yield advantage for some of the agronomic traits of three-way crosses over their respective single crosses, which suggests future breeding on three-way crosses. The three-way hybrids that were selected through the GEI study could potentially be suggested for variety release verification in the selected environments under various managements. Lastly, with only one single cross and hundreds of third parents, the new approach for prediction might be employed for predicting three way cross performance in the future breeding activities of three-ay cross hybrids