Since the introduction of DNA probes as diagnostic tags, closely linked markers have been detected for 18 of the 22 most common Mendelian defects which account for 75% of the morbidity and mortality in this group. Still, as tools for diagnosis and prevention, linked probes are not completely reliable because of occasional crossovers; moreover, this diagnostic approach is not possible for sporadic cases. Fortunately, some of these limitations can be overcome by using a new strategy, termed 'reverse genetics', which aims at the detection of the primary gene defect at the DNA level and should provide the clue to the elucidation of the corresponding biochemical defect which is mostly not understood. These studies should also improve the prospects for therapy, either conventional or through replacement of the defective gene. Furthermore, new techniques have been developed in the mouse that allow the detection of specific gene defects in preimplantation embryos. If combined with in vitro fertilization, similar techniques may provide an alternative to prenatal diagnosis in situations where therapeutic abortion is not considered acceptable.