Molecular Markers Applied to Maize Breeding

244 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1993.Three experiments were conducted to develop methods to practically apply restriction fragment length polymorphisms (RFLPs) to maize (Zea mays L.) breeding. Applications included those requiring linkage between marker loci and loci governing traits of interest and those not requiring linkage.In the first experiment, cluster analysis based on RFLPs resulting from 46 probe-enzyme combinations was used to reveal associations among 148 U.S. inbreds and assign inbreds to heterotic groups. Relationships were estimated using Gower's Coefficient of Similarity and inbreds were grouped by average linkage. Inbreds were ordered into two major groups generally coinciding with distinctions between breeding groups derived from 'Lancaster Sure Crop' open pollinated variety and from Iowa Stiff Stalk Synthetic. Within these two groups, 11 subgroups were formed, each containing an elite inbred (WF9, MO17, C103, PA91, OH43, B14, B73, N28, B37, or OH07).In the second experiment, the classification described above was evaluated to determine if it could represent the true associations among the inbreds. The estimates of relationship were substantiated by pedigree information, using the Hubert $\Gamma$ statistic. Congruence between the phenogram and the proximity matrix indicated that the phenogram depicted estimated relationships accurately. The grouping generally agreed with those obtained in three other cluster analyses utilizing different methods for computing proximities. Thus, the classification appeared to reasonably represent the true associations among the 148 maize inbreds.In the third experiment, RFLP and testcross performance data for yield and harvest moisture for an F$\sb2$:S$\sb4$ population resulting from two related lines were used to explore genotype x environment interaction at the chromosomal level and to evaluate its impact on marker-assisted selection. Differences in the additive effects of chromosomal segments linked to markers across five diverse locations suggested that distinct loci may be operative or the same loci to differing degrees. Indices that assign scores based on genotypes for markers significantly influencing testcross performance in one set of environments were developed to employ MAS for each trait and for both traits simultaneously in three other environments. Generally, significant additive effects were reasonably stable across a diverse range of environments, facilitating gains using MAS beyond those realized using phenotypic selection.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD