基因识别及其算法研究  

Noise Suppression Based Algorithm for Gene Recognition

作　　者：古毅伟[1] 王松[2] 张旭[3] 张茹[1] 刘建毅[1] 仝辉[4] 

机构地区：[1]北京邮电大学计算学院,北京100876 [2]中国科学院软件研究所,北京100190 [3]北京邮电大学信息与通信工程学院,北京100876 [4]北京邮电大学理学院,北京100876

出　　处：《数学的实践与认识》2013年第14期66-76,共11页Mathematics in Practice and Theory

摘　　要：针对基因识别问题,基于DNA序列的3周期这一性质,首先给出了DNA序列功率和信噪比的快速算法并讨论了不同物种基因类型的阈值确定方法;在此基础上,建立了基于背景噪声抑制和频谱平滑的SNR频谱预处理模型,经过预处理后的频谱不仅大幅度抑制了背景噪声,同时保留了SNR频谱的模式特征.在编码序列识别上,对经典的EPND预测算法进行了改进,使用改进的EPND算法对经过预处理后频谱进行基因识别,实验结果显示这种基因识别模型具有优异的基因识别性能,比传统直接使用基于滑动窗口DFT的EPND识别算法在敏感度、特异性等评价指标上提高了2%-12%左右.To solve the problem of gene recognition, this paper presents a fast algorithm to compute the power and SNR for given DNA sequences and discusses the determination method for threshold of different species gene type; On the basis, this paper proposes an SNR sequence preprocessing model based on background noise suppression and spectrmn smoothing. After such a preprocessing, the SNR spectrum remains its pattern characteristics with the background noise suppressed. On the recognition of coded sequence, this paper improves the classical EPND predication Mgorithm which works on the preprocessed gene SNR sequence. The experiment shows the new improved algorithm gains an excellent performance, which improves 2%-12% on major evaluation indexes compared to conventional method based on sliding window DFT and EPND recognition Mgorithm.

关 键 词：基因识别 频谱分析 噪声抑制 EPND 滤波 信噪比 

分 类 号：TN911.4[电子电信—通信与信息系统] O242.1[电子电信—信息与通信工程]