| 摘要 |
PURPOSE: A domain prediction device, a method, and a computer-readable recording medium for recording a program for performing thereof in a computer are provided to predict a domain, based on physico-chemical properties of an amino acid sequence and a secondary structure of a protein, which is difficult to predict with a method for comparing only similarities between sequences. CONSTITUTION: A domain prediction device, a method, and a computer-readable recording medium for recording a program for performing thereof in a computer comprise the following steps: a database (100) stores sequential information of a domain, physico-chemical property information of an amino acid sequence, and secondary structure information; a sequence feature detecting part (200) extracts a sequence feature about the sequential information by using the sequential information, the physico-chemical property information, and the secondary structure information; a machine learning part (300) machine-learns a sequence feature; a sequence analysis part (400) extracts a candidate sequence from inputted protein sequential information; and the sequence analysis part predicts a domain corresponding to the candidate sequence by using a sequence feature of the domain which is machine-learned from the machine learning part. [Reference numerals] (100) Database; (110) Domain DB; (120) Physicochemical feature DB; (130) Secondary structure DB; (200) Sequence feature detecting part; (210) Physicochemical feature generating unit; (220) Secondary structure generating unit; (230) Sequence feature combination unit; (300) Machine learning part; (310) Separating unit; (320) Training unit; (330) Importance determining unit; (400) Sequence analysis part; (410) Input unit; (420) Candidate sequence extracting unit; (430) Domain prediction unit; (440) Output unit |
