유전체 데이터를 위한 효율적인 데이터 관리자의 구현 및 성능분석

Abstract

Recently, many studies have been done on managing and analyzing genome data because researches of the medical field have a lot of interests. Genome data analysis process has consist of two main parts, One is pre-processing which does mapping and tagging to genome data and another is post-processing which does generating mutations of genome data and analyzing mutated data. The pre-processing jab maps and tags to the combinations of whole genome sequences which are the parts of three billions of sequences in a genome data, so that the job takes too much time. However, it generates same data results if the input data is same. The post-processing job makes mutated genome data which is effected by the order of used mutation algorithms, the kinds of mutation algorithms and the number of iterations for reprocessing jobs. In this paper, we design an efficient genome data manager by considering the characteristics of genome data analysis process. First, we design a data sharing table to share genome data efficiently. Secondly, we design a provenance index tree to handle the provenances of genome data processing. Thirdly, based on the designed methods, we design algorithms for managing genome data. At last, we implement our algorithm and show its performance.