关于全基因组关联研究的自动化元分析初探  

Exploring Automated Meta Analyses of Genome-Wide Association Studies

作　　者：冀燃 李冬果[1] 张大保[1] JI Ran;LI Dong-guo;ZHANG Da-bao(School of Biomedical Engineering, Capital Medical University, Beijing 100069, China)

机构地区：[1]首都医科大学生物医学工程学院,北京100069

出　　处：《中国医疗设备》2017年第5期1-5,27,共6页China Medical Devices

基　　金：科技部"973"项目(2014CB744604);北京市教委科技计划面上项目(KM201010025004;KM201410025013);北京市脑重大疾病研究院基金项目(BIBDPXM2014_014226_000016)

摘　　要：随着自然语言分析、文本挖掘等技术高速发展,元分析中数据提取工作逐渐从人工手动提取向计算机自动提取转变。本文以基因关联研究(Genome-Wide Association Study,GWAS)研究为例,通过预先对纳入研究中感兴趣的数据元素进行定位并明确注意事项,来制定元分析的数据自动提取策略方案,使计算机通过搜索少量文献即可快速准确地提取完整的研究数据。以阿尔兹海默疾病的GWAS研究的元分析为例,将纳入的研究按照上文提出的方法进行搜索并提取数据。结果显示,本研究有效缩短了搜索、提取数据的时间,同时提取数据的成功率和准确度可以保持在90%以上。本文为GWAS研究自动提取数据提供了一种有效的策略和向导作用,对于元分析向大数据时代发展有着推进作用。With the rapid development of natural language processing and text mining technology,the step of extracting data from literature began changing from manual extraction to automation by computer.In the past cases,researchers searched entire articles sentence by sentence to looking for key words or key sentences.But the thorough searching without focus points wasted much time.In this paper,we took genome-wide association study(GWAS)as the example to develop the strategies of data automatics extraction for meta-analysis through clearing the positions of data elements we cared about in the included studies in advance to help computers extract the complete data quickly and accurately by searching only parts of the literature.At the same time,we used a GWAS study about Alzheimer's disease as a case study to search and extract data from all the included studies according to the strategies that we developed.Results showed that our strategies not only shortened the time of extraction,but also kept the success rate and accuracy more than90%.Our research provided effective strategies and a guide for the research of automatic extraction of GWAS data,which has a promoting effect on the development of meta-analysis to the big data era.

关 键 词：基因关联研究 元分析 数据定位 数据提取 单核苷酸多态性 

分 类 号：Q789[生物学—分子生物学]