The identification of the cystic fibrosis locus (CF) provides a model for the study of single gene defects where the biochemical lesion is not known. Using families each of which has several affected siblings, it was possible to exclude a number of 'candidate genes' which had previously been proposed as possible sites of the CF mutation. Exclusion mapping of the genome using polymorphic protein and DNA markers showed that CF is on the long arm of human chromosome 7. The most closely linked flanking markers were identified, and human chromosome fragments containing them (and therefore the CF locus) were isolated in rodent cell lines by chromosome-mediated gene transfer. The transgenome was then analysed using cosmid contig mapping, pulse-field gel electrophoresis, HTF island identification and linkage disequilibrium. In this way, a candidate coding sequence has been identified which always segregates with CF.