Eight regions of the genome (PARK1-8) have been implicated in autosomal dominant and autosomal recessive forms of early-onset Parkinson's disease. These forms constitute a few of all cases. However, except for a haplotype in six families (PARK3), no study has successfully mapped a gene or described mutations that contribute to the common late-onset Parkinson's disease. Some have even suggested that a genetic component does not exist. We cross-matched our nationwide genealogy database with a population-based list of Icelandic Parkinson's disease patients to search for families with more than one patient. We performed a genomewide scan on 117 patients and 168 of their unaffected relatives within 51 families using 781 microsatellite markers. Allele-sharing, model-independent analysis of the results showed linkage to a region on chromosome 1p32 with a logarithm of odds score of 3.9 (Z(lr) = 4.2). By increasing the information content with additional microsatellite markers in this region, we found that the logarithm of odds score increased to 4.9 (Z(lr) = 4.8). This result corresponds to an unadjusted p value of 1.0 x 10(-6) and p < 0.005 after adjusting for a genomewide search. We designate this region PARK10. We therefore have successfully mapped, to genomewide significance, a susceptibility gene for late-onset Parkinson's disease using multiple families drawn across a whole population. Identification of the susceptibility gene in this region may pave the way for a better understanding of the disease process, which, in turn, may lead to improved diagnostics and therapeutics.