Interrelationship among Sampling Scheme, Sample Size and QTL Gene Action

抄録

Insufficient sample size is a primary factor contributing to the conflicting results about the effects of genetic markers on quantitative traits. It is of importance to know the minimum sample size required for DNA genotyping before the start of an experiment. A general formula has been developed to serve as guideline for determining the necessary sample size for trait-based analysis. A list of interrelated factors affecting the sample size for DNA genotyping experiment has been identified. Significance level: the smaller the significance level, the greater is the sample size needed. Sampling scheme: multi-sampling requires a greater sample size than tri-sampling, which in turn needs a greater sample size than two-tail sampling. Differences in marker frequency between sampling groups: the greater the difference in marker frequency between groups, the smaller the sample size is needed. Mode of gene action: non-additive gene action requires a smaller sample size compared to additive gene action. A smaller sample size is needed for complete dominance or over-dominance than for incomplete dominance. Conservative level: genome-wide significance reduces the frequency of false positives, but requires a much greater sample size than does point-wise significance. Adjusting point-wise significance to genome-wide significance would increase the required sample size drastically. If the sole purpose of the experiment is to detect marker-QTL association, it is practical to conduct two-tail sampling. If the experiment is designed to detect both the linked QTL gene action and the marker-QTL association, tri-sampling scheme could be applied. The application of multi-sampling is costly and unjustified because there is little, if any, extra genetic information to gain. Genetic markers provide indirect information about linked QTL for marker-assisted selection, which depends solely on linkage disequilibrium. With further perfecting of molecular genetic techniques, direct identification of QTL may become an eventual reality for “QTL-assisted selection” (combining QTL information and phenotype) or direct QTL selection, which will become an important milestone in the history of animal production.