Date of Award
Doctor of Philosophy
Kendall R. Lamkey
In this study, MAS was applied to select the superior individuals from a single-cross hybrid population in maize. The 133 single-crosses, developed by randomly crossing between the recombinant inbred lines of the B73xDe811 (F6:7) and Mo17xH99 (F6:8) population, were evaluated in 2001 and 2002. Eight economically important traits were recorded on a plot mean basis. The first study involved the comparison of the predictive ability of four predictive models for assessing genetic values of individual single-crosses for each trait. Four predictive models including principal component regression (PCR) and stepwise regression (STR) with the individual and integrated marker data from parental populations (ind-PCR, int-PCR, ind-STR, and int-STR), were compared. The result showed that the int-PCR model was the best model in predicting genetic values of individuals with minimizing the integrated means square error (IMSE) and maximizing coefficient of the correlation between the observed and the predicted values (r[macron]op) for tested and untested individuals. The int-PCR model was then used for estimating genetic values (marker scores) of individual single-crosses in the second study, where the objective of the second study was to compare the efficiency of MAS to phenotypic selection (PS) when selection was performed for each trait. Four selection schemes were used: (1) selection based on phenotype only (PS); (2) selection based on marker scores only (MS); (3) selection based on the combination of marker score and phenotype via selection indices (MPS); and (4) tandem marker scores and phenotypic selection (TMPS). Each was applied in selecting 20 individuals with the best performance for each trait of interest. The results showed that all selection schemes were successful in selecting the superior individuals according to their selection differentials and responses to selection. MS was inferior to PS in most cases. While TMPS and MPS appeared to be as effective as or better than PS in selecting for superior individuals according to their responses to selection, their percentages of selected lines in common with PS, and their theoretical efficiency in genetic improvement relative to PS, particularly for the traits with low heritability, i.e., root and stalk lodging.
Digital Repository @ Iowa State University, http://lib.dr.iastate.edu
Kongsamai, Buppa, "Marker-assisted selection and prediction of hybrid performance in maize" (2004). Retrospective Theses and Dissertations. 793.