模拟数量变化对全同胞似然比的影响

The influence of likelihood value of sib when the number of simulation pairs varies

目的本文讨论模拟人群数量和STR数量对全同胞似然比的影响,为法医遗传学鉴定中似然比临界值的制定提供参考。方法采用Familias3软件,分别用15个,19个,29个,39个STR模拟100对到100万对全同胞,观察似然比(LR值)的分布和临界值改变时相应的阳性率、阴性率、假阳性率和假阴性率的变化规律。结果当模拟人口数超过105时,99%,95%,5%和1%的LR值保持在相同的数量级,当模拟群体数量低于104时,波动范围较大。当似然率(LR)为1时各体系的灵敏度分别是:15个STR系统为98.01%-99.0%;19个STR系统为99.0%-99.3%;29个STR系统为99%-99.85%;39个STR系统为99.93%-100%。当似然率(LR)为1时各体系假阳性率为:15个STR系统为1.7225,19个STR系统为0.7370,29个STR系统为0.1311,39个STR系统为0.0593。当STR基因座个数由15个增加到39个时,LR值的参数,包括中位数,平均值,最大值,最小值,99%,95%,5%,1%和STDEV也在相应增加。结论模拟人口规模是影响LR分布的一个重要因素;在一定的人口数下,识别系统中的灵敏度和假阳性率是互相制约,增加检测系统中STR数量,可以获得更好和更高的LR值。

Objective The aim of this paper is to discuss the cutoff value variation of LR when the size of population and number of STRs change. Methods We employed Familias 3 to simulate 100 to one million pairs of full-sibs using 15,19, 29 or 39 STRs and observed the LR distribution and the following parameters: positive, negative, false-positive and false negative results when the cutoff value changes. Results When the number of simulated population is over 105, the LR values of 99%, 95%, 5% and 1% keep the same order of magnitude, and the fluctuation range is larger when the number of simulated groups is less than 104. When LR=1, the sensitivity was 98.01%-99.0% for 15 STRs, 99.0%-99.3% for 19 STRs,99%-99.85% for 29 STRs and 99.93%-100% for 39 STRs. When LR=1, the false positive rates was 1.7225 for 15 STRs,0.7370 for 19 STRs, 0.1311 for 29 STRs and 0.0593 for 39 STRs. When the number of STRs expanded from 15 to 39, the LR value increased in the parameters, including Median, Mean, Max, 99%, 95%, 5%, 1% and STDEV. Conclusion The size of the simulated population is an important factor affecting the distribution of LR. With a certain number of population, the sensitivity and false positive rate in the identification system are mutually restricted. By increasing the number of STR in the detection system, a better and higher LR value can be obtained.