marker-assisted selection in wheat - ictsd

Chapter 15 – Marker-assisted selection in forestry species 291Kirst et al. (2004) measured transcriptabundance in 2 608 genes in the differentiatingxylem of 91 E. grandis backcrossprogeny. QTL analysis of lignin-relatedtranscripts (expressed gene QTL [eQTL])showed that their mRNA abundance isregulated by two genetic loci. Coordinateddown-regulation of genes encoding ligninenzymes was observed in fast growing individuals,indicating that the same genomicregions are regulating growth and the lignincontent and composition in the progeny.Comparative mapping has shown that genecontent and gene order are conserved overlong chromosomal regions among relatedspecies. Comparative maps are thereforelikely to play an important role in enablinginformation on gene location andfunction to be transferred between speciesand genera. However, this will depend onorthologous genetic markers being mappedin each species (Krutovskii and Neale,2001). To date, comparative maps have beenpublished for Populus (Cervera et al., 2001),Pinus (Devey et al., 1999; Krutovsky et al.,2004), Quercus and Castanea (Barrenecheet al., 2004).MAS, based on QTL, is most likely to beused for within-family selection in a limitednumber of elite families that can be propagatedclonally for deployment in large-scaleindustrial plantations. It is most suitablefor traits that are expensive to measure orcan only be detected after plants have beensubjected to a particular stress or pathogen,and that have poor juvenile-mature correlations.Limitations of the approach includethe low resolution of the marker-trait associations,the low proportion of phenotypicvariation explained by QTL (generallyless than 10 percent), and the low successrate in validating QTL in different geneticbackgrounds and environments (Sewell andNeale, 2002). Recombination-based methodologieshave been applied to inbred croplines to positionally clone genes of interestin QTL regions (Salvi et al., 2002); however,the use of this technique in forest treesis not practicable due to their outcrossingbreeding system.Population-based associationstudiesLimitations of the QTL approach have ledto a change in research focus towards identifyingvariations in the DNA sequence ofgenes directly controlling phenotypic variation,known by some as GAS. One of themain advantages of association genetics isthe high resolution of marker-trait associations.As natural populations are usedin association studies, recombinations thataccumulate over many generations of thepopulation break any long range associationsbetween marker and trait leaving shortstretches of the genome associated with thetrait. If alleles or SNPs can be found that arestrongly associated with superior phenotypes,they can be used for selection acrossbreeding populations. This methodology isbetter suited to tree breeding programmes,which aim to maintain a broad genetic base,i.e. programmes with a large number offamilies. In contrast, the QTL approach isused for within-family selection. Spuriousassociations may be observed in associationstudies where there is undetectedgenetic structure in the breeding populationthat invalidates standard statisticaltests. Strategies for dealing with populationstratification have been developed to avoidthese problems (Pritchard et al., 2000; Wuand Zeng, 2001).In the first association study publishedin forest trees, Thumma et al. (2005)identified 25 common SNP markers inthe lignin biosynthesis gene CCR fromEucalyptus nitens. Using single-marker