The "edge strand" hypothesis: Prediction and test of a mutational "hot-spot" on the transthyretin molecule associated with FAP amyloidogenesis