In this work, the application of statistical coupling analysis in a kind of transmembrane protein, Aquaporin, is studied. The major objective of this work is to discover potential correlated residues in E. coli AqpZ involved in molecular selectivity and permeability. Mutating such residues would introduce glycerol permeability to the protein. To obtain this goal, SCA technique combined with network analysis and 3D structural analysis are employed for critical position identification. MD simulation is also used for verifying the accuracy of prediction. This work also applies SCA technique in AQP evolutionary study as well as butelase 1/ligase-related study. To further understand the details of AQP, MD simulation on characteristics of phospholipid bilayer and vesicle is conducted. This research performs SCA analysis on a treated database and establishes corresponding networks according to SCA scores. A set of potential correlated positions are identified through combination analysis of network and molecule structure. Simulation is conducted to approve the correctness of this co-mutation sites prediction. SCA is also shown to be applicable in phylogenetic categorization and to help to find important positions related to the AQP evolutionary process between higher and lower animals. Meanwhile, SCA technique is also demonstrated to allocate critical residues in butelase 1 for ligation catalysis. The findings of this research show the application of SCA technique in bioinformatic research and imply the significance of correlated residues in proteins besides highly-conserved ones. It will be promising to use SCA to allocate critical residues for diseases in AQP family and the research into many other protein families as well. Doctor of Philosophy (MSE)