本文对实验证实的741条大肠杆菌Sigma70启动子的序列进行预测研究。首先，基于RNA聚合酶与DNA的相互作用，利用位置打分函数对序列中的保守位点进行了衡量；然后，根据启动子的序列特征，利用离散性指标对序列中不同的碱基信息含量进行测量；最后，利用多元非线性判别分析实现了对大肠杆菌启动子的预测。10折叠交叉检验结果显示，总体预测精度达到85%以上。与其它算法比较结果显示，我们开发的这一算法能够更好地预测大肠杆菌启动子。

According to the known knowledge of 741 experimentally confirmed Sigma70 promoters, the promoters are predicted. At first, based on the interaction between RNAp and DNA elements, the positioncorrelationscorefunction (PCSF) algorithm is used to measure the conservative sits in promoter sequences. Subsequently, according to the characteristics of promoters, a diversity index is applied to measure the information content in different regions. Finally, the modified Mahalanobis Discriminant is proposed to perform prediction. The overall accuracies of 10fold crossvalidation of 85%+ are achieved. By comparing with other methods, it is shown that the proposed method can recognize the  Escherichia coli promoters with high accuracy.