Browsing by Person "Steinhoff, Jana"

Now showing 1 - 1 of 1

Quantitative trait loci (QTL) mapping in multi-line crosses of European maize
(2012) Steinhoff, Jana; Reif, Jochen Christoph
Multiple-line crosses (MC) have been proposed as promising mapping resource for quantitative trait loci (QTL) detection for agronomic important traits. In contrast to mapping populations derived from a single biparental population, MC can increase the statistical power of QTL detection, the accuracy of QTL location and of QTL effect estimates. Additionally, MC-QTL mapping has the advantage of using data routinely collected in plant breeding programs. The objectives of this study were to (i) assess the reliability of the maize genetic consensus map by comparing it to its six single population linkage maps, (ii) exploit the benefits of a combined analysis by applying two MC-QTL mapping models and to compare the results to single-population analyses, and (iii) investigate the genetic architecture of grain yield, grain moisture, adaptation, and flowering time in elite maize. The experiment comprised six populations with 109 to 150 individuals, resulting from crosses of elite maize breeding material. The germplasm was provided by Syngenta Seeds, Bad Salzuflen, Germany. The 788 genotypes were genotyped with 857 SNP markers. After constructing genetic linkage maps of the six single populations, the genotypic information of the single populations was integrated in a consensus map and its reliability was tested for QTL studies. The average distance between adjacent markers was 1.84 cM suggesting that the marker density is not a limiting factor for QTL analyses. Moreover, we observed medium to high heritabilities for all traits. Consequently, the quality of both genotypic and phenotypic data should allow QTL detection with substantial power. We applied two different MC-QTL mapping models on the data assuming fixed allele effects. The disconnected model estimates QTL effects nested within populations, whereas the connected model takes into account the relationship between the populations. Both models outperformed the single population analyses with regard to QTL detection rate, variance explained by the detected QTL, and the size of the confidence intervals. In all analyses, the disconnected model outperformed the connected model in terms of number of QTL and size of confidence intervals. This superiority seems to be caused by the high background dependencies of QTL effects in connected crosses, which was revealed by a modified diallel analysis of the QTL effects. We investigated the genetic architecture of grain yield, grain moisture, adaptation to maturity zones, and flowering time. Our findings suggest that all traits exhibit a complex genetic architecture with an absence of large QTL effects. Some of the studied traits appear to be influenced by epistasis, interactions between loci. In particular, for flowering time, the two-dimensional scan for epistatic interactions suggested the presence of digenic epistasis. The absence of QTL with large effects suggests that marker-assisted selection is not an appropriate tool to breed for adapted maize hybrids with improved grain yield. Consequently, more suitable approaches for complex traits such as genomic selection should be applied. The joint analyses across populations resulted in higher QTL detection power and resolution compared to single population analyses. Thus, for traits with a less complex genetic architecture, MC-QTL mapping is a powerful tool for the identification of robust diagnostic markers.