Files in this item
|(no description provided)|
|Title:||Application of molecular markers to early generation testing in maize breeding|
|Author(s):||Eathington, Samuel Ray|
|Doctoral Committee Chair(s):||Dudley, John W.|
|Department / Program:||Crop Sciences|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
Agriculture, Plant Culture
|Abstract:||Early generation testing is used to develop improved maize (Zea mays L.) inbreds. Even though the expected genetic correlations between testcross performance in different generations is high, non-genetic effects reduce the effectiveness of early-generation testing. This experiment was conducted to determine the usefulness of molecular marker information to enhance the predictability of early generation testing.
From the maize population by inbred cross, BS11(FR)C7 x FRMO17, 190 random families represented in the $\rm F\sb2:S\sb1$ and $\rm F\sb2:S\sb4$ generations were genotyped at 157 marker loci. Families were testcrossed to a B73 type inbred and evaluated in 8 environments in the Midwest Corn Belt.
The phenotypic correlations between a family's testcross performance in the two generations ranged from 0.36 for grain yield and percent root lodging to 0.65 for grain moisture. Genetic correlations for all traits except test weight were similar to the expected genetic correlation. The correlations between marker effects estimated from the $\rm F\sb2:S\sb1$ and $\rm F\sb2:S\sb4$ testcrosses ranged from 0.43 for percent root lodging to 0.77 for percent stalk lodging.
For grain yield, grain moisture, and percent stalk lodging, significant genotype x environment interaction was detected. At the molecular level, significant marker x environment interaction was detected for all three traits. Marker loci with consistent expression of the marker associated effect across environments and marker loci with inconsistent expression of the marker associated effect across environments were detected.
Five models were evaluated for their ability to predict $\rm F\sb2:S\sb4$ testcross performance. The models combining $\rm F\sb2:S\sb1$ phenotypic information with marker information performed the best. A model consisting of only phenotypic information was in general better than the models that utilized only marker information. For grain yield, the phenotypic model required that 15% of the $\rm F\sb2:S\sb1$ families be selected to retain the top 4 $\rm F\sb2:S\sb4$ families. However, a model combining phentotypic data and marker information required only 6% of the $\rm F\sb2:S\sb1$ families to be selected to identify the top 4 $\rm F\sb2:S\sb4$ families.
|Rights Information:||Copyright 1995 Eathington, Samuel Ray|
|Date Available in IDEALS:||2011-05-07|
|Identifier in Online Catalog:||AAI9624340|