人類基因定位研究中採用多基因座資料較單基因座資料能萃取更多基因座間相關性的訊息,因此近年來多基因座之資料結構逐漸受到重視,進行基因座分群的方法也陸續被提出,其中可變長度馬可夫鏈分群分析尚未用於家庭資料。本研究在四種遺傳模式之下,以可變長度馬可夫鏈分群分析配合傳遞不平衡檢定進行病體-親本三元體之單套型資料分群,目的是探討此分群方法之分群結果於家族資料之敏感度分析,並在此資料結構下比較Browning(2006)以及Browning et al.(2007)所設定的門檻值之優劣。經模擬結果顯示Browning(2006)設定之門檻值在連鎖不平衡係數低時,有較好的分群結果,使用傳遞不平衡檢定時檢定力較高;在連鎖不平衡係數高時,二門檻值之分群效果沒有顯著差異,利用傳遞不平衡檢定時檢定力沒有顯著差異。
Due to the advantage of gaining more power on testing association between a disease and a marker, haplotype-based test methods have been studied by many researchers. One approach in these studies is to use clustering methods to separate haplotypes into several clusters and to adjust for multiple testing by permutation. In this thesis we apply the VLMC (variable length Markov chains) method to case-parent triads data, proposed by Browning(2006) and Browning et al.(2007). We compared the two thresholds proposed by the two studies by simulation method. Our result indicates that the method of Browning (2006) enhances power in low linkage disequilibrium compared with that of Browning et al.(2007). However, in power performance there is no difference between the two methods in high linkage disequilibrium.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。