中国科学技术大学学报
中國科學技術大學學報
중국과학기술대학학보
JOURNAL OF UNIVERSITY OF SCIENCE AND TECHNOLOGY OF CHINA
2010年
9期
892-901
,共10页
条件推断%混杂%基因模型选择%配对%最大稳健有效性%MAX3
條件推斷%混雜%基因模型選擇%配對%最大穩健有效性%MAX3
조건추단%혼잡%기인모형선택%배대%최대은건유효성%MAX3
conditional inference%confounding%genetic model selection%matched pair%maximun efficiency robustness%MAX3
针对配对的病例对照数据,利用偏离哈代-温伯格平衡的信息构造了用于估计基因模型的检验,称为配对的哈代-温伯格不平衡检验(mHWD).基于mHWD估计基因模型并应用相应的配对趋势检验进行关联性分析,称为基于模型选择的趋势检验(GMS).GMS是从数据驱动的角度提出的一种稳健有效的检验,模拟结果显示GMS在现有的关联性检验方法中具有最大的稳健有效性.对类肉样瘤病研究数据的分析进一步表明了基于模型选择的趋势检验的良好性质.
針對配對的病例對照數據,利用偏離哈代-溫伯格平衡的信息構造瞭用于估計基因模型的檢驗,稱為配對的哈代-溫伯格不平衡檢驗(mHWD).基于mHWD估計基因模型併應用相應的配對趨勢檢驗進行關聯性分析,稱為基于模型選擇的趨勢檢驗(GMS).GMS是從數據驅動的角度提齣的一種穩健有效的檢驗,模擬結果顯示GMS在現有的關聯性檢驗方法中具有最大的穩健有效性.對類肉樣瘤病研究數據的分析進一步錶明瞭基于模型選擇的趨勢檢驗的良好性質.
침대배대적병례대조수거,이용편리합대-온백격평형적신식구조료용우고계기인모형적검험,칭위배대적합대-온백격불평형검험(mHWD).기우mHWD고계기인모형병응용상응적배대추세검험진행관련성분석,칭위기우모형선택적추세검험(GMS).GMS시종수거구동적각도제출적일충은건유효적검험,모의결과현시GMS재현유적관련성검험방법중구유최대적은건유효성.대류육양류병연구수거적분석진일보표명료기우모형선택적추세검험적량호성질.
The matched case-control design is widely used in genetic association studies to control potential confounding variables. A commonly used method for analyzing such data is the Mantel-Haenszel (MH) test which can be derived as the score test of a conditional logistic regression model and may not be powerful when the genetic model is mispecified. Trend test incorporating the underlying genetic model is known to be more powerful than the MH test. However, in the practice, the genetic model is usually unknown to the researchers. To circumvent this issue and to retain power of the trend test, MAX-type test that takes the maximum of the typical trend tests was proposed, which was shown to be robust efficient. Here another robust efficient method was proposed by incorporating information of deviation from Hardy-Weinberg equilibrium, which is referred to as the matched Hardy-Weinberg disequilibrium test (mHWD). Our method uses the trend test as the association test but the score of which is determined by the information from mHWD. In this sense, the proposed procedure is a data-driven trend test method. The critical values and p-values of the proposed test can be easily obtained using simple Monte-Carlo methods. Simulation studies show that the proposed test exhibits greater efficiency robustness than the existing tests across a class of scientifically plausible genetic models. The proposed method is illustrated by analyzing a real data set in sarcoidosis study.