PREDICTION DEVICE AND METHOD FOR PREDICTING FUNCTIONALLY AND STRUCTURALLY IMPORTANT SITES BY USING PROBABILISTIC GRAPHICAL MODEL

摘要

Disclosed is an apparatus for predicting an important site of protein by using a probability graph model. The apparatus for predicting an important site of protein by using a probability graph model comprises: an input unit which receives an unknown sequence for a sequence of amino acid or nucleic acid; a database in which information of the sequence of amino acid or nucleic acid has been stored, and a 3D structure of protein has been stored; and a control unit which predicts an important site of the unknown sequence by using the database, wherein the control unit measures evolutionary correlation between different locations and evolutionary dependence of individual acid residues by calculating a Markov Random Field (MRF) model parameter based on at least one of a multi-sequence arrangement generated based on sequences evolutionarily related to the received sequence of amino acid or nucleic acid, an MRF model generated by using a protein structure of the received sequence of amino acid or nucleic acid, and a constraint condition set based on the MRF model, and predicts the important site of the unknown sequence based on the measured evolutionary correlation and the measured evolutionary dependence.