We previously reported that nonobese diabetic (NOD) congenic mice (NOD.c3c4 mice) developed an autoimmune biliary disease (ABD) with similarities to human primary biliary cholangitis (PBC), including anti-mitochondrial antibodies and organ-specific biliary lymphocytic infiltrates. We narrowed the possible contributory regions in a novel NOD.Abd3 congenic mouse to a B10 congenic region on chromosome 1 (“Abd3”) and a mutated Pkhd1 gene (Pkhd1del36−67) upstream from Abd3, and we showed via backcrossing studies that the NOD genetic background was necessary for disease. Here, we show that NOD.Abd3 mice develop anti-PDC-E2 autoantibodies at high levels, and that placing the chromosome 1 interval onto a scid background eliminates disease, demonstrating the critical role of the adaptive immune system in pathogenesis. While the NOD genetic background is essential for disease, it was still unclear which of the two regions in the Abd3 locus were necessary and sufficient for disease. Here, using a classic recombinant breeding approach, we prove that the mutated Pkhd1del36−67 alone, on the NOD background, causes ABD. Further characterization of the mutant sequence demonstrated that the Pkhd1 gene is disrupted by an ETnII-beta retrotransposon inserted in intron 35 in an anti-sense orientation. Homozygous Pkhd1 mutations significantly affect viability, with the offspring skewed away from a Mendelian distribution towards NOD Pkhd1 homozygous or heterozygous genotypes. Cell-specific abnormalities, on a susceptible genetic background, can therefore induce an organ-specific autoimmunity directed to the affected cells. Future work will aim to characterize how mutant Pkhd1 can cause such an autoimmune response.