The genetic basis for rheumatoid arthritis (RA) is likely to be extremely complex. Even the role of MHC genes remains to be fully defined, and may involve interactive genetic effects. The difficulty of precisely defining the clinical phenotype, as well as underlying genetic heterogeneity, complicates the problem. In addition, stochastic genetic or physiologic events may contribute to the low penetrance of susceptibility genes. This situation parallels developing paradigms for other autoimmune disorders, in which many different genes each appear to contribute a small amount to overall risk for disease, and where severity and specific phenotypic subtypes are subject to genetic effects. The completion of the human genome project, along with advances in informatics, will be required to reach a deeper understanding of RA. It is likely that this will involve an iterative and interactive process between several different scientific disciplines.