The genetics of farmer preferred sorghum traits in south western highlands of Uganda
Sorghum (Sorghum bicolor L. Moench) is a staple food and commercial crop for over 95% of the households in the highlands of southwestern Uganda. However, the farmer based selection traits for sorghum in the zone have not been identified and yet farmer involvement in plant breeding programmes is necessary in enhancing adoption of new improved varieties. The diversity in the sorghum populations in the area has also not been established yet diversity in plant genetic resources provides opportunity for plant breeders to develop new and improved cultivars with desirable characteristics. In addition, heritability and gene action for yield and yield traits for the sorghum germplasm in the highlands of southwestern Uganda have not been studied though such information is essential in guiding genetic improvement of various quantitative characters such as yield. As a process of breeding farmer acceptable genotypes with appropriate yield potential, a study on the existing sorghum genotypes in the zone was conducted with the objectives of assessing the basis of farmer selected traits, determining genetic diversity and interrelationships for yield and yield traits of the genotypes and the heritability and gene action for the trait associated with yield. The study to assess farmer preferred traits was conducted using a survey involving 163 farmers purposively selected from the target population based on their indigenous knowledge in sorghum growing, during sorghum growing seasons (July – November) of 2013 and 2014. Genetic diversity studies involving 47 sorghum landraces collected from the zone were done based on 11 qualitative and 13 quantitative traits. The study was conducted at Kachwekano research farm in Kabale District at an altitude of 2,223 m above sea level, during sorghum growing season of December 2014 to August 2016. The study on variability, heritability and gene action for yield and yield contributing traits was based on F2 segregating populations raised from crosses between high yielding and medium plant height cultivars of Shokani, Ndatira and Buhuri all of which acted as males to less yielding and tall female Kyatanombe. A single Kyatanombe panicle, after hot emasculation of its top three quarters, was sectioned into four approximately equal parts; one section was covered with kaki bag containing Shokani pollen grains, the second with Ndatira pollen grains, and the third with Buhuri pollen grains while the lower quarter that was never emasculated was covered with an empty kaki bag to allow for self-pollination. The F1 generation was raised during sorghum growing season of December 2015 to August 2016 at KAZARDI farm xv where the plants were allowed to self-pollinate by bagging them before anthesis. The F2 plants were evaluated alongside parents in a randomised complete block design with three replications during sorghum growing season of December 2016 to August 2017 in Bugongi farm of Kachwekano Zonal Agricultural Research and Development Institute at an altitude of 1840 metres above sea level. The results on the farmer preferred traits revealed culinary traits for making good quality local beverages (29.4%), high yield (28.2%), early maturity (10.4%) and medium height (9.1%) as most farmer preferred selection traits. The results of preferred diversity studies revealed moderate Shannon wiener diversity index (H= 2.03) implying that a few traits were preferred by majority of the farmers. There were weak correlations between farmers preferred selection traits and social economic background of interviewed farmers which ranged from 0.02 to 0.21. The results of Analysis of variance revealed a significant difference between male preferred traits from female preferred traits with higher calculated F value (F=5.57) than F critical value (F crit=3.92) and lower calculated P value of 0.02 than the alpha P value (P< 0.05) suggesting that an improvement programme of sorghum should include views of both males and females to ensure adoption by both sexes. The results of genetic diversity studies basing on the coefficient of variation for qualitative traits revealed peduncle shape (CV=52.2%), peduncle exertion (CV= 47.2) and grain colour (CV=46.4) as most varied traits. Exertion length (CV=94.8%), panicle weight (CV=34.5%), grain yield (34. 1), plant height (CV= 21.7) and panicle length (CV=20) as the quantitative traits that showed highest variability. Days to 50% flowering had the least variability (CV=3). The results of qualitative trait diversity revealed low genetic diversity basing on low Shannon wiener trait diversity index values ranging from H= 0 for Leaf midrib colour to H=1.4 panicle shape and compactness. Results of UPGMA cluster analysis generated a dendrogram with two major clusters with cluster one comprising of 27 landraces and cluster 2 comprising 20 landraces. Principal Component Analysis revealed that the first PC accounted for 27.49 % of variation loaded on traits of exertion length, panicle weight, stem girth, width of third leaf from top, grain yield and peduncle length and second PC accounted for 11.44% loaded on grain colour, peduncle shape and plant height. Correlation and path coefficients analysis for grain yield against 12 quantitative traits revealed significant trait associations (p<0.001) for grain yield with panicle weight (r=0.938), stem girth (r= 0.674), panicle width (r=0.619) leaf width (r=0.576) and leaf length (r= 0.429). Results xvi of regression analysis revealed that panicle weight had the highest direct effect on grain yield (1.085) and Plant height (0.209). Results of path coefficient analysis revealed that panicle width (0.713), 100-sed weight (0.445) and panicle length (0.421) had the highest indirect contribution to grain yield through panicle weight among panicle related traits and stem girth (0.814), leaf width (0.74901), leaf length and number of leaves (0.489) exhibiting the highest indirect contributions to yield via panicle weight among the vegetative traits. Stem girth (0.700), leaf width (0.626), panicle width (0.531) and number of leaves (0.515) had the highest overall indirect contribution to yield. Residual effect was low (R=0.11) with high adjusted R square value (0.89). The study on variability, heritability and gene action for yield and yield contributing traits based on F2 segregating populations raised from crosses between cultivars of Shokani, Ndatira and Buhuri all of which acted as males to female Kyatanombe revealed that the three crosses showed different patterns of heritability and gene action. All traits revealed variable heritabilities and genetic advances. Some traits for example grain yield and 100 seed weight through all the three crosses revealed additive gene interaction with high heritability and genetic advance as a percentage of the mean values while others such as number of leaves and days to 50% flowering were under non-additive gene interaction with low values of genetic advance as percentage of the mean. Varying levels of phenotypic variabilities existed among most of the traits. Higher genotypic coefficient of variation values than environmental coefficient of variation values for all the traits except - plant height in the cross Kyatanombe X Shokani, days to 50% flowering in the Kyatanombe X Ndatira, and number of leaves in Kyatanombe X Buhuri indicated that majority of the traits were influenced by the genetic makeup. In some traits such as plant height and stem girth, mean values for the F2 were higher than the parents indicating heterosis. The study on farmer preferred traits identified important traits to consider, genetic diversity studies identified the existing trait diversities upon which suitable parents with end-user traits would be selected and heritability studies revealed suitable trait improvement strategies through selection during sorghum breeding.