The key functions are:
   metropolis.2   MCMC across DAGs
   model.averaging   model averaging for individual edges
   averaged.net   summary of model average network
   pheno.net.str.qtl   QTL architecture per node as list
   plot.network.3   plot using Rgraphviz

Here are some hints into the code:
score.model:
      ## Drop these and references to them if you don't have QTLs
      scan <- scanone(cross, pheno.col=i, addcov=aux.addcov,
        intcov=intcovM.dat, method="hk")
      ss <- summary(scan,thr=thr[[i]])
      markers <- row.names(ss)

myformula:
## rebuild this as you need.
## Code is built around QTLs (Qnames)
## addcov can be factor or numeric

