文章摘要
沃红梅,易洪刚,潘红星,唐少文,赵杨,陈峰.核函数logistic回归模型在全基因组关联研究中的应用[J].中华流行病学杂志,2013,34(6):633-636
核函数logistic回归模型在全基因组关联研究中的应用
Application of gene-based logistic kernel-machine regression model on studies related to the genome-wide association
收稿日期:2012-10-31  出版日期:2014-09-18
DOI:
中文关键词: 核函数;Logistic回归;全基因组关联研究
英文关键词: Kernel function;Logistic regression;Genome-wide association study
基金项目:国家自然科学基金;江苏省高校自然科学研究重大项目;高等学校博士学科点专项科研基金;江苏高校优势学科建设工程项目
作者单位E-mail
沃红梅 Department of Epidemiology and Biostatistics,School of Public Health, Nanjing Medical University,Nanjing 211166,China
 
fengchen@njmu.edu.cn 
易洪刚 Department of Epidemiology and Biostatistics,School of Public Health, Nanjing Medical University,Nanjing 211166,China
 
 
潘红星 Jiangsu Provincial Center for Disease Control and Prevention  
唐少文 Department of Epidemiology and Biostatistics,School of Public Health, Nanjing Medical University,Nanjing 211166,China
 
 
赵杨 Department of Epidemiology and Biostatistics,School of Public Health, Nanjing Medical University,Nanjing 211166,China
 
 
陈峰 Department of Epidemiology and Biostatistics,School of Public Health, Nanjing Medical University,Nanjing 211166,China
 
 
摘要点击次数: 1325
全文下载次数: 3195
中文摘要:
      [导读]探讨基于基因水平的核函数logistic回归模型及其在全基因组关联研究中的应用.以全基因组关联研究模拟数据为例,介绍核函数logistic回归模型在基因水平检测遗传变异与复杂性疾病之间关联的分析策略.模拟结果表明,在所有已知基因检验结果中致病位点所在基因假设检验的P值最小.结果提示基于基因水平的核函数logistic回归模型能够充分提取和综合基因中多个遗传突变位点信息,降低统计学检验的自由度,同时还能够控制多种协变量因素和交互作用,在检测致病基因与疾病关联时具有一定的效能.
英文摘要:
      [Introduction] To explore the gene-based logistic kemel-machine regression model and its application in genome-wide association study (GWAS).Using the simulated genome-wide singlenucleotide polymorphism (SNPs) genotypes data,we proposed a practical statistical analysis strategynamed ‘ the logistic kernel-machine regression model',based on the gene levels to assess the association between genetic variations and complex diseases.The results from simulation showed that the P value of genes in related diseases was the smallest among all the genes.The results of simulation indicated that not only it could borrow information from different SNPs that were grouped in genes and reducing the degree of freedom through hypothesis testing,but could also incorporate the covariate effects and the complex SNPs interactions.The gene-based logistic kernel-machine regression model seemed to have certain statistical power for testing the association between genetic genes and diseases in GWA.
查看全文   Html全文     查看/发表评论  下载PDF阅读器
关闭