基于支持向量机的蛋白质编码区辨识

Prediction of Protein Coding Regions Based on Support Vector Machine

提出一种新的真核蛋白质编码区位置预测算法。首先,用支持向量机辨识编码区密码子第一位碱基;然后用短时傅里叶变换分析支持向量机输出值序列的时频特性,进而精确定位编码区。仿真结果表明,该算法具有较高的编码区辨识精度。

A novel method for predicting the position of coding regions is proposed.Firstly,a support vector machine is used to recognize the first nucleotide of the codon in coding regions.Then,according to the difference of the time frequency characteristics of output values of the support vector machine analyzed by short time Fourier transform,the position of coding regions can be determined.Simulation results show that the method has high identification accuracy for protein coding region prediction.