| 题名 |
運用15個STR基因進行叔姪關係鑑定之研究 |
| 并列篇名 |
The Study of Uncle-Niece Relationship Determination by 15 STR Loci |
| DOI |
10.6134/tjfm.2010.0202.05 |
| 作者 |
屈保慶(Pao-Ching Chu)；李俊億(James Chun-I Lee) |
| 关键词 |
DNA鑑定 ； STR基因 ； 親子鑑定 ； 叔姪鑑定 ； 似然比 ； 虛擬族群 ； Forensic DNA analysis ； STR locus ； Parentage testing ； Uncle-niece relationship testing ； Likelihood ratio ； Simulated population |
| 期刊名称 |
台灣法醫學誌 |
| 卷期/出版年月 |
2卷2期（2010 / 12 / 01） |
| 页次 |
24 - 33 |
| 内容语文 |
繁體中文 |
| 中文摘要 |
本研究依據台灣地區3,794位漢人的15個STR基因頻率，利用Excel建構出隨機婚配之模擬三代家庭資料庫，隨機選取各12,000組「叔姪二人關係配對」與「隨機二人關係配對」的STR基因型虛擬樣本，計算出叔姪關係累積似然比或叔姪指數之分布，並以ROC曲線做兩樣本群的統計分析，將僅代表基因相似性強度的叔姪指數得以轉成叔姪關係機率模式，二人叔姪指數達到179.80以上，叔姪關係鑑定結果之特異度大於99.99%，可強烈支持二人具叔姪關係。計算使用15個STR基因的142組「叔姪二人關係」與200組「隨機二人關係」的真實檢體樣本叔姪指數，分析叔姪指數分布，發現「虛擬樣本」與「真實檢體樣本」兩樣本群分布之平均數與標準差均無顯著差異，虛擬樣本群得以代表真實樣本群。自虛擬樣本群比較使用不同的STR基因數目之鑑別叔姪關係的效果，使用越多STR基因鑑定數目，鑑別叔姪關係的能力越強，但因為離散程度的增加，「叔姪二人關係」與「隨機二人關係」兩樣本群交集的區域無法明顯的分開，而使鑑別效果增加有限。本研究推演出「二叔一系爭姪」所有可能的基因型配對組合，再使用家庭組合配對演算法推導出「二叔一系爭姪」的所有基因型配對組合公式，以分析三人叔姪關係及計算三人叔姪指數。自模擬家庭資料庫選出各12,000組「二叔一姪關係」與「二叔一隨機之人關係」的虛擬樣本，在「二叔一隨機之人關係」的虛擬樣本中，共有26.02%的配對樣本不符合遺傳法則的「二叔一姪」基因型組合關係而排除；而餘下的73.98%，則計算三人叔姪指數作進一步的鑑別。因此在系爭叔姪（伯姪、姑姪、舅甥、姨甥）二人關係下，再增加一位叔（伯、姑）或一位舅（姨）的STR基因型，可同時應用遺傳法則排除法與三人叔姪指數進行鑑別，增加對系爭姪之親屬關係的鑑別力。 |
| 英文摘要 |
The simulated three-generation-family database was generated by Excel software according to the allele frequencies of 15 STR loci of 3,794 Taiwanese population and random mating method in this study. There were 12,000 virtual STR profiles of uncle-niece pairs and random pairs respectively picked out from the simulated family database. The distributions of uncle-niece index (accumulated likelihood ratios of uncle-niece relationship) were calculated and then were analyzed by ROC curve, so the uncle-niece indices which only stood for the strength of genetic similarity could be transformed into the probability mode about uncle-niece relationship. If the uncle-niece index of 15 STR loci between two people was higher than 179.80, the specificity of uncle-niece relationship determination would be higher than 99.99%, and it strongly supported that uncle-niece relationship was true. Real human samples of 142 uncle-niece pairs and 200 random pairs were taken into analysis. There was no statistically significant difference in variance and in mean between simulated data and real human data, and it was assumed that the simulated population could represent the real population. The identification effect corresponding to the STR loci numbers was estimated from the simulated data, and the result revealed that it was more powerful to determine uncle-niece relationship while using more numbers of STR loci. However, the degrees of dissemination were also augmented, and therefore may restrict the effect in relationship determination. In order to calculate uncle-niece indices among two uncles and one alleged niece, all possible genotype constituents about the triples were listed to deduce the formulae of likelihood ratio. There were 12,000 triples of uncle-uncle-niece and uncle-uncle-random respectively taken out from the simulated family database for testing. The exclusion rate to analyze two uncles and one alleged niece by genetic inconsistency principle was 26.02%, and the remained 73.98% could be estimated by uncle-niece index formulae of triples which were already deduced. While adding one more uncle or aunt, the combination use of genetic inconsistency principle and uncle-niece indices could increase the power of discrimination in uncle-niece identification. |
| 主题分类 |
醫藥衛生 >
基礎醫學 社會科學 > 法律學 |
| 参考文献 |
|
