THE EFFECTS OF SELECTIVE GENOTYPING ON ESTIMATES OF PROPORTION OF RECOMBINATION BETWEEN LINKED QUANTITATIVE TRAIT LOCI

Selective genotyping is the marker assay of only the more extreme phen otypes for a quantitative trait and is intended to increase the effici ency of quantitative trait loci (QTL) mapping. We show that selective genotyping can bias estimates of the recombination frequency between l inked QTLs - upwardly when QTLs are in repulsion phase, and downwardly when QTLs are in coupling phase. We examined these biases under simpl e models involving two QTLs segregating in a backcross or F-2 populati on, using both analytical models and computer simulations. We found th at bias is a function of the proportion selected, the magnitude of QTL effects, distance between QTLs and the dominance of QTLs. Selective g enotyping thus may decrease the power of mapping multiple linked QTLs and bias the construction of a marker map. We suggest a large proporti on than previously suggested (50%) or the entire population be genotyp ed if linked QTLs of large effects (explain > 10% phenotypic variance) are evident. New models need to be developed to explicitly incorporat e selection into QTL map construction.