FINE MAPPING OF THE DIABETES-SUSCEPTIBILITY LOCUS, IDDM4, ON CHROMOSOME 11Q13

Genomewide linkage studies of type 1 diabetes (or insulin-dependent di abetes mellitus [IDDM]) indicate that several unlinked susceptibility loci can explain the clustering of the disease in families. One such l ocus has been mapped to chromosome 11q13 (IDDM4). In the present repor t we have analyzed 707 affected sib pairs, obtaining a peak multipoint maximum LOD score (MLS) of 2.7 (lambda(s) = 1.09) with linkage (MLS g reater than or equal to 0.7) extending over a 15-cM region. The proble m is, therefore, to fine map the locus to permit structural analysis o f positional candidate genes. In a two-stage approach, we first scanne d the 15-cM linked region for increased or decreased transmission, fro m heterozygous parents to affected siblings in 340 families, of the th ree most common alleles of each of 12 microsatellite loci. One of the 36 alleles showed decreased transmission (50% expected, 45.1% observed [P = .02, corrected P = .72]) ar marker D11S1917. Analysis of an addi tional 1,702 families provided further support for negative transmissi on (48%) of D11S1917 allele 3 to affected offspring and positive trans mission (55%) to unaffected siblings (test of heterogeneity P = 3 x 10 (-4), corrected P = .01]). A second polymorphic marker, H0570polyA, wa s isolated from a cosmid clone containing D11S1917, and genotyping of 2,042 families revealed strong linkage disequilibrium between the two markers (15 kb apart), with a specific haplotype, D11S131703-H0570pol yA02, showing decreased transmission (46.4%) to affected offspring an d increased transmission (56.6%) to unaffected siblings (test of heter ogeneity P = 1.5 x 10(-6), corrected P = 4.3 x 10-4). These results no t only provide sufficient justification for analysis of the gene conte nt of the D11S1917 region for positional candidates but also show that , in the mapping of genes for common multifactorial diseases, analysis of both affected and unaffected siblings is of value and that both pr edisposing and nonpredisposing alleles should be anticipated.