Once QTL have been identified, the program can search for interactions. For each pair of QTL, the information criterion is calculated with an interaction effect. The minimum over all pairs is compared to the information criterion of a model without the epistatic effect. If the difference is greater than the threshold, the effect is retained and the process is repeated. In each step, the new effect is tested within the backgroud of all the QTL main effects as well as previously identified interactions.
For crosses with two marker genotypic classes, only additive by additive effects are tested. For those with three classes, additive by dominance, dominance by additive and dominance by dominance effects are also tested.
By default, MImapqtl does a forward stepwise search for epistatic terms. If you want to try a backward elimination approach, use a B instead of an E in the appropriate position. Also, be aware that if there are too many epistatic terms, the request for a backward elimination approach will be ignored in favor of a forward search. MImapqtl only allows for parameters. If you want to relax this constraint, then use a U instead of the B. MImapqtl will then allow for up to parameters.