| dc.description.abstract | Maize has a significant and increasingly crucial role in global food security systems, but the threat of gray leaf spot (GLS) disease, caused by Cercospora zeina, persists and continues to impact its yields on a global scale.Forty-eight inbred lines of maize from the International Maize and Wheat Improvement Center (CIMMYT)were sorted into 6 sets and factorial crosses of each set generated according the North Carolina mating design II. The lines alongside their crosses evaluated across 12 environments to characterize partial disease resistance to artificial GLS epidemics,temporal progress of the disease, genetic association of resistance with 9 phenological traits, and combining abilities alongside heterotic effects of the inbred lines for resistance. Eight measures of resistancethat includedstandardized area under disease progress curve (SAUDPC), and weighted mean absolute rate of disease increase (ρ) were examined. SAUDPCs were the most efficient infection measure, buttheir variability were best explained (R2 = 93.9%) by disease ratings between the VT (full tassel emergence preceding pollen shed) and R4 (dough stage) stages of development. Individual disease ratings at R4 were almost as effective as SAUDPCs in characterizing genotype reactions.The Levenberg-Marquardt algorithm, along with integration of data eyeballing and gridding to determine starting parameter values, were the best methods for estimating the nonlinear regression model that accurately fit the gray leaf spot (GLS) epidemics.The generalized version of the Richards model outperformed all other growth curve models examined except in some instances where its curve shape parameter matched the fixed curve shape models that best fit any particular epidemic. Goodness of fit of the monomolecular, Gompertz, logistic and exponential models varied with the resistance of entries and favorability of the trial environments to GLS development. Genotypic and phenotypic correlations of SAUDPC were strongest with the Stay-green characteristic (SGR) (r = −0.87). The magnitude and direction of coheritability estimates mirrored trends in genotypic and phenotypic correlations. The covariation of GLS resistance with agronomic traits was mainly due to the direct effects of days to anthesis (DTA) and days to silking (DTS), and the indirect effects of stay-green capacity (SGR) and silking-maturity interval (SMI). General combining ability (GCA) and specific combing ability (SCA) effects were significant. However, the GCA effects were more important than SCA effects signifying the preponderance of additive gene action. Although this was confirmed by significant correlations between GCA effects and per seresistance of inbred parents, the correlations were not strong enough to be of predictive value. Estimates of potence ratio suggested that non-additive effects were due to multiple loci and alleles with recessive, dominance, over-dominance, and various types of epistasis. These findings indicate that comparative analysis of GLS epidemics should be based on the generalized Richards model and breeding for resistance should prioritizereciprocal recurrent selection, pedigree selection, restricted index selection, multiple population improvement, and multistage selection depending on the genetic background of the sources of resistance. | en_US |
