As we have discussed previously (Horn et al. 1988a; Erlich et al. 1989b; Horn et al. 1988b), there are no unique class II sequences associated with IDDM, which suggests that "normal" class II alleles confer susceptibility. Given the estimates of concordance--under 50% of monozygotic twins and approximately 15% (Tattersol, Pyle 1972 and Thomson 1988) for HLA-identical sibs--, it is not surprising that some unaffected individuals contain putative susceptibility alleles. Perhaps some environmental "triggering" agent, such as viral infection (Yoon, this volume), is required for the disease to develop in susceptible individuals. Other non-MHC linked genes which contribute to susceptibility may account for the difference in concordance rates for monozygotic twins and for HLA-identical sibs. In the nonobese diabetic mouse and the BB rat models for IDDM, non-MHC susceptibility loci have been identified and mapped (Hattori et al. 1986; Colle et al. 1981), but in humans the analysis of non-MHC candidate loci (i.e., the T cell receptor) has thus far failed to reveal any other susceptibility loci. In general, the HLA-linked genetic susceptibility to IDDM, as well as to other autoimmune diseases, appears to be associated with specific combinations of class II epitopes (e.g., alleles, haplotypes, or genotypes) rather than with specific individual residues or epitopes. Understanding the role of these predisposing sequences will require structural analysis of the class II molecules as well as in vitro and in vivo functional studies of interactions with putative autoantigens and T cell receptors. In the meantime, DNA typing offers the potential for identifying individuals at high risk for IDDM.