Fig. 4.

Power of QTL detection 1- β ^∗ of joint composite interval mapping (JCIM) model 1 (black line) and 2 (red line) with cofactor selection method 1 at different α ^∗ levels in a scenario with heritability h ²= 0.8, 50 backcross nested association mapping (BC-NAM) subpopulations and (a) 0.5 of explained ratio by QTL × genetic background interactions to the total genetic variance, where 1 QTL interacted with 5 background markers; (b) 0.75 of explained ratio by QTL × genetic background interactions to the total genetic variance, where each of 25 QTLs interacted with 10 background markers. Vertical lines at each point indicate the standard errors