Method and apparatus for discovering sequence motif using probabilistic model

摘要

PURPOSE: A method for excavating a ranking motif with a probabilistic model and a device thereof are provided to infer the DNA combination mechanism of a transcription factor by automatically finding the number and location of the ranking motif combined through the recognition of the transcription factor from protein-DNA combination data. CONSTITUTION: A ranking processing unit(10) classifies DNA(Deoxyribonucleic Acid) rankings combined with a transcription factor into sets according to a type of the transcription factor and obtains sub rankings from the sets. A model generating unit(20) generates an infinite mixture model modeling the probability distribution of concealed variables about the location and the type of a motif according to a base pattern of the sub rankings on condition that a total number of motifs is not limited. A model learning unit(30) learns the infinite mixture model to maximize the ex-post facto probability distribution of the infinite mixture model. [Reference numerals] (1) Ranking motif excavating device; (10) Ranking processing unit; (2) Ranking DB; (20) Model generating unit; (30) Model learning unit; (40) Output unit