Interaction testing. (A) Interaction graph is given by a two-way analysis of variance considering the two QTL as two fixed factors with an interaction between them. The "Q-q" QTL effect at the proximal QTL is more important when the animals have the Q allele at the distal QTL (-4g/+15g with the q/Q allele at the distal QTL respectively). (B) The table gives the resulting p-value for both factors and the interaction. (C) QTL analysis using the "interaction model": the blue curve corresponds to the classical model testing the hypothesis « No QTL versus 1 QTL», the grey curve to the model testing «One QTL versus 2 QTL», the green curve corresponds to the interaction model « No QTL versus 1 QTL in interaction with another one fixed at 168 cM » and the red corresponds to the interaction model after adding the six novel markers. For each case, the significance threshold at the 0.05 level is displayed.