阿尔茨海默病基因一疾病关联的知识挖掘  

Knowledge Mining of Alzheimer's Disease Gene-Disease Associations

作　　者：王雪[1,2] 武俊伟[3] 陈观群[4] 李燕琼 马路[1] Wang Xue;Wu Junwei;Chen Guanqun;Li Yanqiong;Ma Lu(Medical Humanities School,Capital Medical University,Beijing 100069;Department of Library,Xuanwu Hospital,Capital Medical University,Beijing 100053;Medical Information Section,Chinese PLA General Hospital,Beijing 100853;Department of Neurology,Xuanwu Hospital,Capital Medical University,Beiing 100053)

机构地区：[1]首都医科大学医学人文学院,北京100069 [2]首都医科大学宣武医院图书馆,北京100053 [3]中国人民解放军总医院医学信息室,北京100853 [4]首都医科大学宣武医院神经内科,北京100053

出　　处：《图书情报工作》2020年第13期120-132,共13页Library and Information Service

基　　金：首都医科大学宣武医院院级管理课题“基于科技影响力排行的医院重点学科影响力分析”(项目编号:XWGL-2019003);首都医科大学宣武医院院级教学课题“基于元素养理论的医学生信息素养教学路径研究”(项目編号:2019XWJXGG-10)研究成果之一。

摘　　要：[目的/意义]对阿尔茨海跌病(AD)进行基因-疾病关联挖掘,以捕捉潜力研究方向。[方法/过程]基于LBD理论构建开放式知识发现架构,结合MeSH词表、DisGeNET等医学术语组学数据对PubMed中AD文献进行知识挖掘,采用关联规则与算法排序等方法对部分基因重合的强关联主题共现疾病和优先候选基因进行筛选,结合时间切片和其他LBD工具对比加以验证。[结果/结论]对88334篇AD文献进行基因-疾病识别,并与2120种AD基因进行匹配;以XYZ分析视角对识别出的992种主题共现疾病及11899种候选基因进行关联排序;精炼10种强关联疾病与25种优选候选基因,结合文献报道加以论述。通过LBD挖掘目标疾病-共现疾病-基因之间潜在关联,可快速捕捉潜力研究方向,缩小基因测序范围,为新研究假设的生成提供重要指导依据。[Purpose/significance]To explore the gene disease association of Alzheimer's disease(AD)in or-der to capture the potential research directions.[Method/process]An open knowledge discovery framework was constructed based on LBD theory.Combined with MeSH thesaurus,DisGeNET and other medical terms and group data,knowledge mining was carried out in AD literatures in PubMed.Association rules and algorithm sorting were used to screen strongly associated MeSH terms co-occurrence diseases and priority candidate genes for partial gene co-incidence,results of time slicing and comparison with other LBD tools were used to verify them.[Result conclu-sion]88334 AD literatures were identified and matched with2120 AD genes,11899 candidate genes and 992 co-morbidity genes were identified according to XYZ analysis,10 strongly associated co-occurrence diseases and 25 pre-ferred candidate genes were refined and discussed in combination with literature reports.Mining the potential associa-tions between target disease,co-occurrence diseases and genes by LBD can quickly capture the potential research di-rections,narrow the scopes of gene sequencing,and provide important guidance for the generations of new research hypotheses.

关 键 词：知识发现 基因组学 阿尔茨海跌病 实体识别 数据挖掘 排序算法 时间分析 

分 类 号：G250[文化科学—图书馆学] R745[医药卫生—神经病学与精神病学]