All of the type I collagen in connective tissue is the product of one structural gene for the pro alpha 1(I) chain and another for the pro alpha 2(I) chain of type I procollagen. An intriguing question therefore is how the expression of the two genes differs in mineralizing and non-mineralizing tissues. One approach that our laboratory has pursued to answer this and related questions is to develop a new system whereby one can examine the self-assembly of collagen fibrils de novo by controlled enzymic cleavage of procollagen to collagen under physiological conditions. The system has made it possible for the first time to define thermodynamic parameters for the self-assembly process. We are now using the system to define the normal kinetics for fibril formation. The results should make it possible to study the effects of other components of extracellular matrix on fibril assembly, including the effects of bone-specific components that initiate mineralization. A second approach has been to define mutations in type I procollagen genes that cause increased brittleness of bone. Over a dozen mutations in type I procollagen genes have been found in probands with osteogenesis imperfecta. One of the surprises has been that at least 25% of the probands with lethal variants of osteogenesis imperfecta have mutations in type I procollagen genes. Another surprise has been the observation that a number of the mutations are tissue specific in terms of their phenotypic manifestations even though the same abnormal pro alpha chains are being synthesized in a variety of tissues.