Genomic approaches for improving grain yield in maize using formerly plant variety protected germplasm.

摘要

The rapid increase in size of the global human population and growing consumer preference for diets higher in protein have led to an increased demand for maize (Zea mays L.) grain. This has created concerns about our ability to provide enough maize grain in coming decades. Public sector plant breeders attempt to develop improved germplasm and also better understand molecular mechanisms that will enhance the performance of high yielding commercial hybrids. The public availability of commercially-derived inbreds formerly protected through the Plant Variety Protection Act (ex-PVP) may provide an enhanced germplasm source that will make public research efforts more valuable and adaptable to private sector breeding efforts. Several genome level methods have been proposed to increase the rate of genetic gain in maize. These include mapping quantitative trait loci (QTL) for use in marker-assisted selection (MAS), simultaneously identifying QTL and develop high yielding lines through advanced backcross populations, and applying genomic selection across the genome to account for small-effect QTL. The objectives of this study were to 1) extensively characterize the phenotypic performance of a set of ex-PVP inbreds in a diallel mating scheme and relate these well-characterized inbreds to a larger set of publicly available lines using molecular markers, 2) identify QTL stable across multiple testers for yield, yield components, and agronomic traits in a population derived from two ex-PVP inbreds 3) develop and test the effectiveness of a new QTL mapping method in maize termed reciprocal advanced backcrossing (RAB) which allows for a greater ability to detect and refine the position of QTL while concurrently developing high yielding lines derived from elite ex-PVP inbreds, and 4) compare the accuracies of genomic selection and phenotypic selection in a conventional F2:3 breeding population derived from ex-PVP inbreds.;We identified B73, PHG84, LH123HT, and PHZ51 as inbreds with high general combining ability for yield by crossing 10 ex-PVP inbreds and two public inbreds in a half-diallel mating scheme. Several high yielding crosses between inbred lines with high GCA were identified. We determined the genetic relatedness of uncharacterized ex-PVP lines to those extensively characterized in the diallel analysis, which could be useful to public and private breeders in selecting inbreds to develop new breeding populations. Yield QTL that were stable across multiple testers were identified on chromosomes one and seven using an F2:3 population derived from ex-PVP inbreds LH51 and PHG35. Yield QTL have in these regions have also been identified in other studies and can be incorporated into MAS breeding programs designed to improve yield. Additional yield QTL were identified on chromosomes three and ten using a reciprocal advanced backcross population derived from LH51xPHG35. High yielding lines that can be further enhanced by selecting for all of the yield QTL were developed as part of the reciprocal advanced backcross QTL mapping (RAB-QTL) approach. We demonstrated that RAB-QTL improves the ability to detect QTL over traditional advanced backcross QTL mapping (AB-QTL) by equalizing allele frequencies and allowing for a contrast between homozygous marker classes. Genomic selection and phenotypic selection for yield, moisture, and test weight were compared by estimating allelic effects from the 358 individuals in the LH51xPHG35 F2:3 mapping population testcrossed to LH119 and PHG39. The LH119 testcross was evaluated in 2010 and 2011 and the PHG39 testcross was evaluated in 2011 and 2012. Estimated allelic effects were used to estimate GEBVs of 180 untested lines and 100 tested lines. Following genomic selection, untested lines from the prediction population and tested lines from the training population were evaluated in the field to compare the accuracy of genomic and phenotypic selection. The two methods performed equally well. The reduced cost of genomic selection relative to phenotypic selection couple with the equal performance of both methods suggests that the increased use of genomic selection relative to phenotypic selection will allow breeders to test more lines with greater potential and increase genetic gain without the need for additional spending. The results of this dissertation can be applied towards increasing rate of genetic gain and at the same time help to meet the future grain yield demands of a growing human population.