Understanding protein structures through molecular modeling is critical to our long-range goal of developing cures or treatments for cystic fibrosis, secretory diarrhea, schizophrenia and drug addiction. One protein molecule (eg CFTR, TAAR1, mu OR) can go through multiple states (eg open, closed, active, inactive), and we want to have accurate models of these states in order to understand the function of the protein and design drugs for pharmacotherapy or design the protein for gene therapy.
Different molecular models have different levels of accuracy. You must keep doing mock modeling trials to understand how accurate models are generated. In order to understand the levels of accuracy required for our purposes, it is not enough to just generate molecular models. You must also think about what you can predict using the models. A good example for you to think about is locking of a protein molecule into different states. For example, we can lock R104C/E116C CFTR into the open state because of a spontaneous disulfide bond. It can also be locked into the closed state with the M2M molecular linker as described in the CFTR project. Dr Norimatsu believes R104C/E116C CFTR with a spontaneous disulfide bond spends most of the time in a state we call the semi-open state. Our goal is to design a CFTR mutant that spends most of its time in the true open state when activated (phosphorylated).