Linkage-based QTL mapping has become a standard method for examining the genetic architecture underlying phenotypic variation. This approach is now routinely employed in species for which genetic maps are available, including both crop species and wild species. QTL mapping is essentially a statistical method for localizing chromosomal regions where genetic variation can be associated with measurable phenotypic variation. Two individuals showing phenotypic differences for a trait are crossed. Their progeny are then either selfed, backcrossed to one parent, or crossed with each other to generate a large mapping population, in which the parental alleles are shuffled into different genetic backgrounds. Associations between phenotypes and genetic marker variation are identified, and the specific genomic locations of the genetic markers can then be inferred by comparison with the genetic map.
For addressing questions on the genetics of adaptive divergence, one of the most successful applications of QTL mapping has involved crop species and traits selected upon during domestication. Foundational studies using the cereal crop maize and its wild relatives elucidated the genetic bases of a variety of domestication traits, including plant architecture and inflorescence developmental traits (Dorweiler et al. 1993; Doebley et al. 1997). Many crops share a suite of 'domestication traits'. Among cereal crops, for example, there has been parallel selection for increased seed size and loss of seed disarticulation (shattering) upon ripening. As QTL studies have spread beyond maize to include other cereal crop species, it has become possible to assess the degree to which parallel selection pressures have acted on orthologous QTLs in related species. A substantial degree of QTL shar ing has been observed in cereal crops such as maize, rice and sorghum (Paterson et al. 1995). A similar pattern of QTL sharing is also seen in crops of the Solanaceae, including tomato, eggplant, pepper, and potato (Doganlar et al. 2002). Thus, it appears that independent domestication events may have involved artificial selection on many of the same QTLs.
Among wild species, QTL mapping in the model organism Arabidopsis thaliana has provided important inferences into the genetics of adaptation. QTLs have been identified for a variety of life history traits that, if underlying natural phenotypic variation, would be expected to have key ecological and evolutionary relevance. Examples include chemical defense (Kliebenstein et al. 2002), timing of reproduction (e.g., El-Assal et al. 2002; Ungerer et al. 2002), inflorescence and floral morphology (Juenger et al. 2000); and fruit and seed characteristics (Alonso-Blanco et al. 1999). Unlike crop species, where a relatively limited number of QTLs are involved in domestication traits, multiple QTLs (often 10 or more) are typically found for these traits in A. thaliana, with complex epistatic interactions among them. One recent study (Weinig et al. 2002) specifically sought to test the degree to which the flowering time QTLs that were identified under controlled laboratory conditions corresponded to those detected in a natural ecological setting. Interestingly, a substantial proportion of the QTLs with major effects in the lab was undetectable in the field, and vice versa; in addition, QTLs in the field differed between seasonal environments. These results suggest a level of complexity in the genetic interactions underlying natural adaptive divergence that would remain undetected with standard QTL mapping approaches under controlled growth conditions.
QTL mapping in wild plant species other than A. thaliana have also provided significant insights into the evolutionary genetics of adaptation and adaptive divergence. One of the most fruitful areas of research has focused on the evolution of traits underlying reproductive isolation, and the connection between intraspecific adaptive divergence and the process of speciation. Studies involving two species in the genus Mimulus have proved particularly insightful (Brad-shaw et al. 1995, 1998; Schemske et al. 1999). Mimulus lewisii and M. cardinalis are interfertile and sympatric, but are reproductively isolated in nature due to differences in floral traits and resulting pollinator preferences (the former species is pollinated by bumblebees, the latter by hummingbirds). A relatively few QTLs of major effect can account for major phenotypic differences in floral traits between these species. This finding suggests that the barrier to reproduction between these species may be attributable to a small number of 'speciation' QTLs. Studies in other species, including Louisiana irises (reviewed by Arnold 2000) and Helianthus species (e.g., Kim and Rieseberg 1999) have focused on natural hybridization and the genes underlying ecological differentiation between hybridizing species. This approach can be used in directly assessing the relationship between interspecific introgression and the movement of QTLs for environment-specific adaptations.
While an important tool for examining the genetics of adaptation, QTL approaches also suffer from a number of drawbacks. First, QTL analysis requires that a genetic map be available for a species, which limits its application out side of crops and model taxa. In addition, multiple generations of crossing are required to generate mapping populations after the initial cross, and large populations of progeny must be maintained. These factors hinder the use of this approach in physically large species and in perennial plants. QTL mapping also suffers from limited resolution and power for identifying genes underlying phenotypic variation. The chromosomal region identified as a QTL is typically on the order of 5-10 cM, even for species with well resolved genetic maps. Even in the small genome of A. thaliana, this often translates into a physical region of 1-2 Mb. Thus, it is no small task to pinpoint a specific candidate locus (or possibly loci) within a QTL, although fine-mapping studies with large recombinant populations are possible (see for example Yano et al. 2000) and may permit the isolation of adaptive genes.
A final drawback of QTL mapping is particularly problematic for studies aimed at examining the genetics of adaptation in natural species. Only those QTLs that segregate between the two individuals used the initial cross may be identified. This leaves the vast majority of naturally occurring genetic variation in a species undetected. Moreover, since the QTLs identified from a particular cross may vary in their expression in different genetic backgrounds and different environmental conditions (see Arabidopsis example above), the actual ecological relevance of QTLs identified from a particular cross remains tenuous at best.
Was this article helpful?