Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/44702
Title: | 以15個STR基因進行台灣地區漢人手足關係鑑定 Determination of sibship by analysis of 15 short tandem repeat loci in Taiwan Han population |
Authors: | Yen-Yang Lin 林彥仰 |
Advisor: | 李俊億 |
Keyword: | 親子鑑定,手足關係,半手足關係,累積似然比,虛擬族群, parentage analyses,fullsib relationship,halfsib relationship,cumulative likelihood ratio,virtual family, |
Publication Year : | 2010 |
Degree: | 碩士 |
Abstract: | 親屬關係鑑定在現今人身鑑定中占有非常重要的地位,除了親子關係之外,手足和半手足關係算是其中相當重要的一項,在父母存在的狀況下,可以借助親子鑑定的結果,使手足和半手足關係鑑定變的比較容易,若是父母不在的話,便只能靠單純的手足或半手足關係鑑定,本研究係利用台灣地區3794位漢人STR基因頻率作為標準數據,以符合隨機婚配與族群基因頻率建立虛擬族群,包含三代完整的家族個體,運用Excel函數共建立3500個虛擬家族,作為具有親屬關係的基礎,再隨機生成三人和這些家族中成員比對,作為不具親屬關係的基礎,在手足、半手足之有親屬關係與無親屬關係配對的STR基因型作為樣本,使用FORDDAS計算其累積似然比,可得出有親屬關係與無親屬關係的累積似然比分布狀況,畫出曲線圖可以看出此檢定法能不能將具有親屬關係和不具親屬關係的族群區分開來,進一步分析可得接受器操作特性曲線(ROC curve),變動截點時敏感度(sensitivity)和特異度(specificity)也會跟著變化,實際分析10500組虛擬家族中二人組手足關係STR基因,結果顯示此二人組手足關係分析法在特異度達0.9999時正確度約為0.905,可信度尚可,而半手足關係同樣分析10500組虛擬家族二人組半手足關係STR基因,結果顯示分析特異度若要達0.9999,正確度只有約0.56,顯示此二人組半手足分析可信度不足,建立三人組分析方法之後,實際分析10000組虛擬家族中三人手足STR基因,結果顯示在相關組中,累積似然比全部不為0,而非相關組有98.91%累積似然比為0,即排除率為98.91%,若判斷標準訂在累積似然比為0,則特異度為0.9891,敏感度為1,正確度為0.99455,另外分析10000組虛擬家族中三人半手足STR基因,結果顯示在相關組中,累積似然比全部不為0,而非相關組有26.24%累積似然比為0,即排除率為26.24%,此結果顯示三人組手足和半手足關係鑑定法,可以直接排除部分非相關組配對。 The kinship determination has important role in human identify testing now. Apart from parentage analyses, fullsib and halfsib analyses are very important. When parent is alive, we can use parentage analyses to improve test result. In many forensic parentage cases, however, only children undergo genetic tests. If parent had died, we can only use sibship analyses. We use 3794 Taiwanese’s STR frequency to create 3500 virtual family by computer program. These virtual family include three generations and fit random mating rule. We use FORDDAS to calculate likelihood ratio of fullsib and halfsib relationship in virtual family. In non-relationship analyses, we create 3 random man each virtual family. We can draw relationship and non-relationship’s cumulative likelihood ratio distribution graphs. Then we can calculate sensitivity and specificity by excel software. In two full-sibship analyses, we used 10500 pairs virtual data to calculate combined full-sibship index (CFSI). When specificity was 0.9999, accuracy was 0.905. In two half-sibship analyses, we used 10500 pairs virtual data to calculate combined half-sibship index (CHSI). When specificity was 0.9999, accuracy was 0.56. It was a very low accuracy. When we had created three fullsib and halfsib analyses, we calculated cumulative likelihood ratio by 10000 groups virtual data. In full-sibship, 100% relative groups’ cumulative likelihood ratio diden’t equaled 0 but 98.91% non-relative groups’ cumulative likelihood ratio equaled 0. The exclusion rate of three full-sibship analyses was 98.91%. If cut point set on 0, sensitivity was 1 and specificity was 0.9891. The accuracy was 0.99455. On other hand, the exclusion rate of three half-sibship analyses was 26.24%. The three sibship analyses can direct exclude partial non-sibship cases. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/44702 |
Fulltext Rights: | 有償授權 |
Appears in Collections: | 法醫學科所 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-99-1.pdf Restricted Access | 847.12 kB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.