Add option to do two-dimensional scan with interaction covariates (ie, interact with founder allele probs)

[fboehm]

We can probably assume that both dimensions of the scan would have these interaction covariates. By looking at rqtl/qtl2 code, especially that for scan1(), this can be addressed.