AbstractWe report a combined segregation and linkage analysis of a Danish sample of 216 insulin‐dependent diabetes mellitus (IDDM) nuclear families: of these 216, twenty multiplex families were haplotyped regarding HLA‐DR and ‐B markers. The analysis was conducted using the computer program COMBIN, which includes a modifier to absorb family resemblance that is additional to the effect of the major locus that is assumed linked to a marker locus, eg, within the HLA region. The initial analysis could clearly reject a dominant major locus but could not discriminate between other models with or without modifier. However, after adding supplementary information on population associations between HLA and IDDM together with the identity‐by‐descent (IBD) distribution to the analysis, a final model was identified. This invokes an additive major locus, linked to HLA with recombination not significantly different from 0, a disease gene frequency of 0.217, plus a dominant modifier. From this model it can be predicted that about 0.15 of the general population is at 100 risk of IDDM, about 5 is at intermediate risk (approximately 10), while the remaining population has a risk of 0. The model predicts recurrence risks compatible with empirically estimated values. Particularly strong, positive haplotype associations were found for theDR3,B8, DR3,B18, and DR4,B15haplotypes, but detailed analyses showed that neither these particular haplotypes nor theDR3andDR4haplotypes in general could entirely explain the HLA‐associated susceptibility. TheDR2haplotypes showed a strong negative association. The results are discusssed in the light of available data on the epidemiology of IDDM in order to provide a framework for further epidemiologi
展开▼
