基于主判别变量法的近红外光谱STR基因分型专家系统

Genotyping-Dendrogram of D16S539 Locus Based on Near-infrared Spectroscopy and Principal Discriminate Variate

基于近红外光谱结合主判别变量算法对短串联重复序列(STR)基因分型专家系统进行研究。选择STR基因座D16S539中6种总核心重复串数不同且出现频率较高的基因型9-9、9-11、9-12、10-10、10-11和11-11作为研究对象。6种基因型样本分别通过两次聚合酶链反应(PCR)平行扩增,以扩增产物的近红外光谱作为判别变量建立判别模型。最后,基于该基因座6种基因型的总核心重复序列数的差异来构建其基因分型专家系统。结果表明:方法适用于不同差异程度的基因型分型,预测准确率高且稳健性好,这为建立其他基因座的分型专家系统提供了可行性。

The genotyping-dendrogram of short tandem repeat （STR） has been established based on near infrared spectroscopy in combination with the principal discriminate variate method. Six genotypes 9 9, 9-11, 9 12, 10-10, 10-11 and 11 11 of D16S539 locus, which had different total number of core-repeat units and the high frequencies of occurrence, were selected as the research objects. The samples of 6 genotypes were obtained by polymerase chain reaction （PCR） in twice. Discriminate models with the near-infrared spectrum of each PCR product as discriminating variable were built. Based on the differences of total number of the core-repeat units, the genotyping-dendrogram was established. It was shown that these models in the dendrogram were appropriate for discriminating the genotypes of all the samples and the method was of good stability and predicting ability, h provided potential to establish expert system for genotyping of other STR loci.