Method for and lessons from mapping QTL in livestock
Recently research on mapping QTL in livestock species has moved from linkage analysis using microsatellites, to genome scans using linkage disequilibrium (LD) between QTL and markers, based on medium density SNPs (eg 10,000 SNPs). In livestock species the pattern of LD is typically very different to that found in humans. In cattle, for instance, the effective population size has decreased steadily since domestication and this has caused LD to exist at Mb scale but to be less than expected at kb scale. This is an ideal pattern for LD mapping of QTL and the reverse of that found in humans.
All QTL mapping methods use markers to estimate the probability that individuals carry QTL alleles that are identical by descent (IBD)at a particular position in the genome. It is advantageous to combine both linkage and LD when estimating this probability that individuals are IBD. These probabilities can then be used in a linear model to detect and map the QTL. The likelihood of false positives can be reduced by including in the linear model environmental effects and relationships among the individuals.
In planning experiments it is useful to know the size of effects to be expected. We have estimated that the typical quantitative trait is controlled by 50-200 genes. The effects of these genes are not all equal but follow approximately an exponential distribution. Thus, the average gene has a gene substitution effect of approximately 0.05 standard deviations. Genes of much larger effect occur but rarely and the typical trait has no genes with effect greater than 0.7 standard deviations. However, experiments grossly overestimate the effect of significant QTL, often by a factor of 2 or more. As a result of this, experiments to validate QTL often have insufficient power because the effect is much smaller than expected. Methods are available to make an unbiased estimate of the size of the effect of a QTL and these should be used in planning validation experiments.