抄録
Three modules of a computer system for predicting membrane protein structure have been developed. First module distinguishes membrane proteins from other type of proteins based upon the information of amino acid sequence alone. In this process, only hydrophobicity of amino acids has been used and the discrimination of 99% was possible, analyzing amino acid sequence by Fourier transform method. The secondary structure is predicted in the second module, in which the hydrophobicity played dominant role. The last module predicts the configuration of transmembrane helices solving a jigsaw puzzle, the pieces of which are rods of helices with polar areas on their surface.
