A large number of QTL mapping breeding designs and estimation methods exist. Some methods are related to interval mapping but utilize more than just pairs of marker loci (composite interval mapping) or utilize all of the linked markers on individual chromosomes (multipoint mapping). Another approach is to test for associations between marker genotypes and phenotypic means in only those individuals that show extreme phenotypic differences in an F2 population such as the individuals at the upper and lower tails of the phenotypic distribution. In such bulked segregant analyses the marker loci near QTLs are expected to be in gametic disequilibrium because of linkage. Therefore, the individuals in the lower tail of the phenotypic distribution would have one marker genotype and the individuals in the upper tail of the phenotypic distribution would have another marker genotype if a marker locus were very tightly linked to a QTL. In contrast, a marker locus independent of any QTL would show all genotypes at equal frequencies in the individuals that represented the upper and lower tails of the pheno-typic distribution.

Limitations of QTL mapping studies

QTL mapping results are highly context-sensitive, just like heritability estimates. The number and effects of QTLs identified in one population may not be representative of QTL effects in another population of the same species. Mapping only identifies QTLs that are segregating in the population at the time of mapping. Further, QTL mapping by the F2 design can only detect and estimate the phenotypic effects of two alleles at a QTL. The two QTL alleles detected are those fixed when inbreeding lines to form P1 in Fig. 9.15. In reality, there may be more than two alleles segregating in a population for any QTLs and detecting these requires screening replicate P1 crosses. Alleles at QTLs that are fixed or lost in the original population or become fixed or lost by genetic drift during the formation of inbred lines cannot be identified. Likewise, the estimates of QTL effect sizes are always relative to the other QTL loci segregating in a population. For example, QTL X could explain 10% of the phenotypic difference between marker-class means in one population whereas QTL Y has even bigger effect but happens to be fixed for one allele. If QTL Y is segregating instead, then QTL X will have a smaller perceived effect on the pheno-type in a mapping study. This occurs because when two alleles are segregating at QTL Y, there will be a greater difference in phenotype between P1 individuals than when QTL Y is fixed.

Several types of statistical power limitations impact QTL mapping results. Actual QTLs of small effect cannot be identified easily because it is difficult to show that a small difference between marker-class means is statistically meaningful given inherent environmental variation, experimental measurement error, and corrections required when carrying out a very large number of statistical tests. The effect sizes of QTLs that are identified as statistically meaningful can be substantially inflated since the QTLs with small effects are not identified. This so-called Beavis effect is pronounced if the number of progeny in a mapping study is about 100 and modest with about 500 progeny (Beavis 1994; Xu 2003).

The number of QTLs identified in mapping studies is likely to underestimate the true number of QTLs that cause variation in a trait (Otto & Jones 2000). Under conditions similar to many actual QTL mapping studies, no more than about a dozen QTLs will be identified as statistically meaningful (Hyne & Kearsey 1995). In addition, two or more QTLs that are adjacent in the genome may appear as a single QTL. If the effects of two linked QTLs are in the same direction, the perceived single QTL will have an inflated effect. On the other hand, if the two linked

QTLs have contrasting (or antagonistic) effects on the trait, then a single perceived QTL will be detected that has a downwardly biased effect size that is therefore less likely to be detected as statistically meaningful. The number and spacing of genetic markers also influences the perception of QTL numbers and effect sizes, since widely spaced markers (relative to the recombination rate) may miss QTLs or aggregate the effects of multiple QTLs. The term quantitative trait region or QTR is sometimes used to describe an association between a marker locus and a marker-class mean difference since multiple linked QTLs may exist within the genome interval mapped. Further fine-scale mapping with genetic markers spaced at smaller intervals along the chromosome in the specific chromosomal regions around QTR can be used to identify true single QTLs (e.g. Kroymann & Mitchell-Olds 2005).

Biological significance of QTL mapping

QTL mapping is now carried out routinely in model and domesticated species, facilitated by the large numbers of molecular markers and high-throughput genotyping techniques as well as genome sequencing and genetic linkage mapping projects. QTL mapping can also be carried out, albeit usually with more difficulty and less resolution, in some non-domesticated species using crosses within and between species (reviewed by Slate 2005). Table 9.7 shows some examples of the number of QTLs identified for various phenotypes in a range of species. In Table 9.7 the results range from one or a few QTLs with large

Table 9.7 Examples of QTLs identified by mapping with genetic marker loci.



Number of



0 0

Post a comment