The simultaneous analysis of multilocus genotypes provides more information for linkage than a series of pairwise analyses. In a scan of the human genome, 10 to 15 markers are typed on each chromosome, and one wants to extract the full information provided by these joint genotypes. As we just saw in section 1.2, the Lander-Green Hidden Markov model described in section 5.3 of week 3 is best suited for multilocus computations. This algorithm has been implemented in the programs MAPMAKER/SIBS for sib-pair analysis and GENEHUNTER for analysis of general pedigrees of moderate size. (Kruglyak and Lander [2], Kruglyak et al. [1]).