Abstract Osteogenesis imperfecta (OI) is a heritable form of bone fragility typically associated with a dominant COL1A1 or COL1A2 mutation. Variable phenotype for OI patients with identical collagen mutations is well established, but phenotype variability is described using the qualitative Sillence classification. Patterning a new OI mouse model on a specific collagen mutation therefore has been hindered by the absence of an appropriate kindred with extensive quantitative phenotype data. We benefited from the large sibships of the Old Order Amish (OOA) to define a wide range of OI phenotypes in 64 individuals with the identical COL1A2 mutation. Stratification of carrier spine (L1–4) areal bone mineral density (aBMD) Z-scores demonstrated that 73% had moderate to severe disease (less than −2), 23% had mild disease (−1 to −2), and 4% were in the unaffected range (greater than −1). A line of knock-in mice was patterned on the OOA mutation. Bone phenotype was evaluated in four F1 lines of knock-in mice that each shared approximately 50% of their genetic background. Consistent with the human pedigree, these mice had reduced body mass, aBMD, and bone strength. Whole-bone fracture susceptibility was influenced by individual genomic factors that were reflected in size, shape, and possibly bone metabolic regulation. The results indicate that the G610C OI (Amish) knock-in mouse is a novel translational model to identify modifying genes that influence phenotype and for testing potential therapies for OI. © 2010 American Society for Bone and Mineral Research