Mapping As You Go: An Effective Approach for Marker-Assisted Selection of Complex Traits

Many factors have contributed to the inability to successfully employ a marker-based selection scheme for The advent of high throughput molecular technologies has led to complex traits. One major difficulty has been the effecan expectation that breeding programs will use marker–trait associations to conduct marker-assisted selection (MAS) for traits. Many tive detection, estimation, and utility of QTL and their efchallenges exist with this molecular breeding approach for so-called fects. This is especially the case for traits governed by complex traits. A major restriction to date has been the limited ability context-dependent gene effects (i.e., interaction with other to detect and quantify marker–trait relationships, especially for traits genes and/or environment). Historically, most evaluations influenced by the effects of gene-by-gene and gene-by-environment of mapping and MAS strategies have largely ignored the interactions. A further complication has been that estimates of quantipossibility of context dependency and have instead astative trait loci (QTL) effects are biased by the necessity of working sumed that QTL act independently (i.e., no interaction with a limited set of genotypes in a limited set of environments, with other genes and/or environment). Adoption of this and hence the applications of these estimates are not as effective as assumption has led some to suggest that MAS has little expected when used more broadly within a breeding program. The if any power over traditional phenotypic selection (e.g., approach considered in this paper, referred to as the Mapping As You Go (MAYG) approach, continually revises estimates of QTL Bernardo, 2001). Increasing evidence suggests, however, allele effects by remapping new elite germplasm generated over cycles that context-dependent factors are important in the deof selection, thus ensuring that QTL estimates remain relevant to the termination of trait phenotypes (e.g., Cooper and Hamcurrent set of germplasm in the breeding program. Mapping As You mer, 1996; Schlichting and Pigliucci, 1998; Clark, 2000; Go is a mapping-MAS strategy that explicitly recognizes that alleles Wolf et al., 2000; Holland, 2001; Mackay, 2001; McMulof QTL for complex traits can have different values as the current len et al., 2001; Tuberosa et al., 2002; de Visser et al., breeding material changes with time. Simulation was used to investi2003). The presence of these factors has significant imgate the effectiveness of the MAYG approach applied to complex traits. plications for the way mapping and MAS should be The results indicated that greater levels of response were achieved and conducted for the improvement of complex traits in a these responses were less variable when estimates were revised frebreeding program during the short, medium, and long quently compared with situations where estimates were revised infrequently or not at all. term. Analysis methods have been developed to accommodate for the effects of context dependency (e.g., Crossa et al., 1999; Jannink and Jansen, 2001; Nelson et al., R advances in molecular genetics have led to 2001; Boer et al., 2002; van Eeuwijk et al., 2002). For an enthusiasm for use of MAS to improve the perexample, in the case of epistasis, Holland (2001) outformance of traits in plant breeding (Koornneef and lined an approach that was based on the identification of Stam, 2001; Peleman and Rouppe van der Voort, 2003). preferred allele configurations across interacting genes. The key components to the implementation of this apSimilar approaches have been suggested by others (e.g., proach are (i) the creation of a dense genetic map of Jansen et al., 2003; Kuhnlein et al., 2003). Applications molecular markers, (ii) the detection of QTL based on of this approach will be a challenge, even for well-studstatistical associations between marker and phenotypic ied gene networks (Peccoud et al., 2004). Other advariability, (iii) the definition of a set of desirable marker vances in methodology include the use of multiple line alleles based on the results of the QTL analysis, and crosses among related individuals (Jannink et al., 2001; (iv) the use and/or extrapolation of this information to Yi and Xu, 2001; Bink et al., 2002) and/or haplotype the current set of breeding germplasm to enable markerinformation to increase the power to accurately estimate based selection decisions to be made. To date, this apQTL and their effects (Meuwissen and Goddard, 2000; proach has been effective for relatively simple traits that Jansen et al., 2003). In all cases, the analysis methods are controlled by a small number of genes (e.g., disease assume that the mapping studies can be conducted with resistance; Flint-Garcia et al., 2003) but less effective sufficient power to adequately account for all, or at least for more complex traits controlled by many genes that the important, context dependencies that may exist. are under the influence of epistasis (gene-by-gene interRegardless of what assumptions are made, a common action) and gene-by-environment interaction effects outcome of all QTL analysis methods is the estimation (e.g., grain yield; Openshaw and Frascaroli, 1997; Melof QTL allele effects, whether at an individual gene level chinger et al., 1998; Utz et al., 2000). or across multiple interacting gene complexes (Jansen, 1996). A target combination of marker alleles is defined Pioneer Hi-Bred International, 7250 NW 62nd Ave., P.O. Box 552, from these estimates, forming the basis of selection in Johnston, IA 50131-0552. Received 31 Oct. 2003. *Corresponding author ([email protected]). Abbreviations: MAS, marker-assisted selection; MAYG, Mapping As Published in Crop Sci. 44:1560–1571 (2004).  Crop Science Society of America You Go; MET, multienvironment trial; MSO, Mapping Start Only; QTL, quantitative trait loci. 677 S. Segoe Rd., Madison, WI 53711 USA