酵母基因组中核小体偏好序列的识别

Prediction of nucleosome preferred sequences in yeast genome

应用多样性增量结合二次判别分析(Increment of Diversity with Quadratic Discriminant analysis,IDQD)方法,对酵母基因组中的核小体强/弱偏好序列进行了识别。10交叉检验的预测成功率超过了97%,受试者操作特性(receiver operating characteristic,ROC)曲线下面积达到了0.99,预测成功率高于现有SVM算法。最后利用构建好的分类器对酵母基因组中三类包含TATA盒基因的起始密码子ATG上游400nt下游100nt区域进行了分析。结果表明,IDQD算法有能力应用于基因组中核小体序列的识别。

The method of Increment of Diversity combined with Quadratic Discriminant analysis （IDQD） was used to predict the nueleosome strong/weak preference- sequences in yeast genome. The results of 10- fold cross- validation test gave an accuracy of 97% and the area under ROC （auROC） curves of 0.99. The accuracy is superior to that of the ettrrently published SVM method. Finally the classifier was applied in the analysis of regions around ATG for three types of TATA - containing genes in yeast genome. The above results show that the IDQD method is capable of recognizing nueleosome positioning sequences in yeast genome.