Type 1A diabetes is an autoimmune disease associated with the major histocompatibility complex (MHC), in particular, with the high-risk HLA-DRB1*0301- DQA1*0501-DQB1*0201/HLA-DRB1*04- DQA1*0301-DQB1*0302 (DR3/4-DQ8) genotype. Accumulating evidence suggests that genes other than the HLA-DR and HLA-DQ (HLA-DR/DQ) genes within or linked to the MHC contribute additional risk. My first aim was to quantify the risk associated with these other non-HLA-DR/DQ MHC or MHC-linked genes. Prospective analysis of siblings with the DR3/4-DQ8 genotype who shared both of their MHC haplotypes identical-by-descent with their proband (first affected family member) showed that they had a much greater risk (85% (+/-12%)) for anti-islet autoimmunity by age 15 than the DR3/4-DQ8 siblings who did not share both of their haplotypes with their proband (20% (+/-11%) risk). However, extensive linkage disequilibrium within the MHC confounds efforts to identify MHC or MHC-linked genes associated with diabetes independent of the HLA-DR/DQ genes. My second aim was to examine the conservation of a common, extended, "ancestral" MHC haplotype with the HLA-A*0101, HLA-B*0801, HLA-DRB1*0301, and HLA-DQB1*0201 alleles, the 8.1 haplotype. A high-density panel of MHC SNPs revealed that 8.1 haplotypes from unrelated individuals were essentially identical for a 2.9 million nucleotide region of the classic MHC and that the conservation extended as long as 9 million nucleotides. My third and final aim was to search for diabetes-associated polymorphisms in the MHC region with correction for the extended linkage disequilibrium of MHC haplotypes, such as the 8.1 haplotype. Three studies of families (N=1,006 families) from different populations, genotyping high-density SNP panels, revealed variants in the UBD/MAS1L region at the far telomeric end of the MHC that were significantly associated with diabetes. The association with the UBD/MAS1L region remained even after correction for the confounding effects of linkage disequilibrium with the 8.1 haplotype and other HLA-DR/DQ haplotypes. The specific identification of the functional variants, such as in the UBD/MAS1L region, that account for the increased risk associated with haplotype sharing in siblings will likely further improve our ability to predict genetic risk for type 1A diabetes in children with and without a family history of diabetes. The form and content of this abstract are approved. I recommend its publication.