Genetic linkage analysis infers recombination parameters between genetic loci from the observed genotypes and phenotypes of related individuals. The pedigree forms a discrete Bayesian network over founder alleles, inheritance, and recombination indicators, with inference computing the data likelihood as a function of the recombination fraction.
The canonical exact-inference implementation is SUPERLINK by Fishelson and Geiger, which represents general pedigrees as Bayesian networks and combines variable elimination with conditioning.