Fig. 3.

Pupal weight variation in female hybrids maps to many additive Z-linked effects in Papilio. (A) One-dimensional QTL scan shows that variation in pupal weight is mainly explained by the Z chromosome. LOD scores peak at the Z chromosome center. Alternating colors represent different chromosomes. (B) Introgressed ancestry fraction on the Z chromosome ($f_Z$) is highly informative of pupal weight (adjusted by brood). (C) Genotype–phenotype regression does not support any 1-QTL architecture for pupal weight. The distribution of regression power is measured by $R_{\text{1-marker}}^2/R_{\text{Z-ancestry}}^2$ (the ratio between $R^2$ using a single marker versus a polygenic model of Z chromosome ancestry). The polygenic model provides a better fit for the observed curves. (D and E) Bayesian QTL model selection (33) based on 15 sparsely spaced markers favors the simultaneous inclusion of multiple markers ($\ge$3) in predicting pupal weight. Markers more favored by models have higher marginal probabilities of being selected. Marginal probabilities equal the partial sum of the posterior probabilities of QTL models.