Even for complex diseases with high rates of monozygotic twin concordance, disease-associated alleles remain elusive. One explanation is that multiple common genetic variants with weak effects cause these diseases and identification of any single allele requires large cohorts. Conversely, if the allelic spectrum of complex disease is heterogeneous, strong effects of rare variants might be offset by their presence in only a small proportion of the patient population. Lupus (SLE) is a systemic autoimmune disease, with significant monozygotic twin concordance, protean clinical manifestations, and production of high-affinity pathogenic autoantibodies. This complex phenotype and results from genome scans point to multiple molecular defects. Contrary to this expectation, our analysis of ENU-mutagenized mice indicates that homozygous mutations frequently cause anti-nuclear antibodies (ANAs), and can account for a full blown lupus phenotype. The best characterized example is the sanroque strain, which develops high-affinity dsDNA autoantibodies and fails to censor self-reactive germinal centre T cells. Mapping the underlying mutation identified not only a novel gene, Roquin, but also a novel pathogenic pathway for SLE. Identification of such rare variants with strong effects is likely to identify pathogenic pathways that underlie pathology in many patients, lead to interacting molecular partners that also cause pathology, and identify the most effective therapeutic targets.