Polycystic kidney disease (ADPKD) is a condition with an autosomal dominant mode of inheritance and adult onset. Two forms of the disease, ADPKD1 and ADPKD2, caused by mutations in PKD1 and PKD2, respectively, are very similar, except that ADPKD1 patients run a more severe course. At the cellular level, ADPKD1 was first shown to be recessive, since somatic second hits are perhaps necessary for cyst formation. The near identical phenotype had suggested that ADPKD1 and ADPKD2 might have a similar pathogenesis and that the two gene products, poly- cystins 1 and 2, are part of a common developmental pathway. Work in transgenic mice showed that somatic loss of Pkd2 expression is necessary for renal cyst formation, and recently we showed that somatic mutations inactivating the inherited healthy allele were present in 9 of 23 cysts from a human ADPKD2 kidney, supporting a two-hit loss-of-function model for ADPKD2 cystogenesis. Here, we provide the first direct genetic evidence that polycystins 1 and 2 do interact, perhaps as part of a larger complex. In cystic DNA from a kidney of an ADPKD1 patient, we showed somatic mutations not only in the PKD1 gene of certain cysts, but also in the PKD2 gene of others, generating a trans -heterozygous state with mutations in both genes. One mutation in PKD1 is of germinal nature and the mutation in the PKD2 gene is of somatic nature. The implications of such a situation are enormous, not only for ADPKD, but also for many other conditions with phenotypic heterogeneity and age-dependent penetrance.