The affected sib-pair method can be applied to investigate linkage bet
ween a marker locus and a disease. Several statistics have been propos
ed to test if the observed pattern of marker alleles shared identicall
y by descent (ibd) is compatible with the null hypothesis of no linkag
e. Here, we consider different optimality criteria for sib-pair linkag
e tests. While for recessive inherited diseases the mean test is found
to be uniformly (in 8) most powerful, it can also be shown that, irre
spective of the mode of inheritance, the mean test is the locally opti
mal test.