Two percent of the residents of Bengkala, Bali, have profound, congenital, neurosensory, nonsyndromal deafness due to an autosomal recessive mutation at the DFNB3 locus. We have employed a direct genome-wide disequilibrium search strategy, allele-frequency-dependent homozygosity mapping (AHM), and an analysis of historical recombinants to map DFNB3 and position the locus relative to flanking markers. DFNB3 maps to chromosome 17, closest to D17S261, pRM7-GT and D17S805. In individuals homozygous for DFNB3, historical recombinant genotypes for the flanking markers, D17S122 and D17S783, place DFNB3 in a 5.3 cM interval of the pericentromeric region of chromosome 17 on a refined linkage map of 17p-17q12. Based on conserved synteny, the murine sh2 gene may be the homologue of DFNB3.

The nonsyndromic congenital recessive deafness gene, DFNB3, first identified in Bengkala, Bali, was mapped to a approximately 12-cM interval on chromosome 17. New short tandem repeats (STRs) and additional DNA samples were used to identify recombinants that constrain the DFNB3 interval to less, similar6 cM on 17p11.2. Affected individuals from Bengkala and affected members of a family with hereditary deafness who were from Bila, a village neighboring Bengkala, were homozygous for the same alleles for six adjacent STRs in the DFNB3 region and were heterozygous for other distal markers, thus limiting DFNB3 to an approximately 3-cM interval. Nonsyndromic deafness segregating in two unrelated consanguineous Indian families, M21 and I-1924, were also linked to the DFNB3 region. Haplotype analysis indicates that the DFNB3 mutations in the three pedigrees most likely arose independently and suggests that DFNB3 makes a significant contribution to hereditary deafness worldwide. On the basis of conserved synteny, mouse deafness mutations shaker-2 (sh2) and sh2J are proposed as models of DFNB3. Genetic mapping has refined sh2 to a 0.6-cM interval of chromosome 11. Three homologous genes map within the sh2 and DFNB3 intervals, suggesting that sh2 is the homologue of DFNB3.

Bengkala is an Indonesian village located on the north shore of Bali that has existed for over 700 years. Currently, 2-2% of the 2185 people in this village have profound congenital deafness. In response to the high incidence of deafness, the people of Bengkala have developed a village specific sign language which is used by many of the hearing and deaf people. Deafness in Bengkala is congenital, sensorineural, non-syndromal, and caused by a fully penetrant autosomal recessive mutation at the DFNB3 locus. The frequency of the DFNB3 mutation is estimated to be 9-4% among hearing people who have a 17-2% chance of being heterozygous for DFNB3. (J Med Genet 1995;32:336-343) A medical and genetic analysis of deafness in Bengkala, Bali began in 1990 with the identification of deaf subjects and their relationships. An Bali.1-3 Goitre has been the concern of Indonesian physicians for many years and has been the subject of extensive regional iodide therapy since 1980.2 In addition to goitre, 47 of the residents of Bengkala are congenitally deaf. Fig 1 shows the families of all of the deaf people in Bengkala. It is suspected that the history of deafness in Bengkala spans many generations since virtually the entire population of Bengkala uses a village specific sign language as a means of integrating the deaf and hearing into a community. FAMILY HISTORIES AND ASSESSMENT OF DEAFNESSFamily histories and pedigrees were collected over a three year period and were used as a guide to interviews needed to obtain multiple independent confirmations of the hearing status, the time of onset of deafness, and the progression of hearing impairment for each deaf person.

Genotypes for 53 short tandem repeat (STR) markers distributed at an average of 39 cM intervals throughout the genome were determined for 46 individuals from the village of Bengkala, Bali. This village of approximately 2200 individuals has an oral and written tradition suggesting genetic bottlenecks. The allele frequency distributions in Bengkala were compared with distributions obtained by typing individuals in the CEPH data base using a Kolmogorov-Smirnov two sample test. Twenty-eight of the 53 markers showed differences (P < 0.05) in distribution between the two populations. Allele frequencies of tetranucleotide STRs were much more similar between the two populations than were those of dinucleotide STRs (P < 0.043). Population heterogeneity in Bengkala was indicated by an excess of observed homozygosity, deviations from Hardy-Weinberg equilibrium at seven loci, and significant allelic associations between physically unlinked loci. In addition to providing information pertinent to the issue of genetic diversity of STRs in the human population, these analyses serve as a resource to map a gene causing non-syndromal autosomal recessive deafness in Bengkala, and to corroborate the anthropological study of the history and social structure of the village.

A total of 71 pairs of like-sexed dizygotic twins were studied, comparing within-pair differences for plasma total, free and esterified cholesterol, and triglyceride with the number of HLA haplotypes the twins had in common. If associations are present between HLA and the blood lipids studied, the twins with no haplotypes in common would be expected to have the largest within-pair mean square, those with two in common the smallest, and those with one in common an intermediate value. No significant differences were found comparing within-pair mean squares for the variables studied.