Over 40 loci for deafness have been genetically mapped but only in very recent years have causative genes been identified. Two of these are connexins and this promted Xiaet al to search for new human connexin genes and look for mutations in families with deafness. From a database, two overlapping ESTs were identified with 83% identity to rat Gjb3 and a homologous fragment amplified from human DNA. This identified human GJB3 (connexin 31) which mapped to 1p32-p35. RT-PCR analysis showed it to be expressed in the inner ear. Of six families with sensorineural deafness linked to 1p32-p35, two were found to have mutations in the connexin 31 gene. One resulted in an amino acid change and the other in a premature stop codon causing absence of part of the C terminus of the protein. Both of these mutations were in regions highly conserved in other connexins. In both families, inheritance was autosomal dominant but males were affected with progressive bilateral high frequency hearing impairment with onset from 20 to 40 years of age, whereas females were either unaffected or much less severely affected. A significant proportion of families with AD hearing loss show linkage to 1p32-35, so Xia et al may have identified a common cause of deafness. Interestingly, in the same issue ofNature Genetics, an independent group describe the identification of the same gene but found mutations in a different region to be responsible for erythrokeratodermia variabilis in which deafness is not a feature. These findings are discussed in the “News and views” section and represent an interesting example of different mutations in the same gene being responsible for two completely different phenotypes.