基因预测中的信噪比计算新模型

Calculation of Signal-to-Noise Ratio in Gene Prediction

以快速傅立叶变换(FFT)求出DNA序列功率谱后,用重复计算实验的办法,对大量DNA序列中核苷酸的分布情况进行统计,得出综合考察变量,然后以核苷酸分布频数和DNA序列长度为自变量,建立Z-curve映射规则得到DNA序列信噪比的新模型,并进行实例分析。新模型不需要对序列进行离散Fourier变换(DFT),而且不要求DNA指示序列长度必须为3的倍数,应用范围更大。

Fast Fourier Transform （FFT） was used to get the power spectrum. By means of repeating computational experiments, the distribution frequency of nucleotides distribution was statistically analyzed in a large number of DNA sequences to get a aggregate variable. Then let the distribution frequency and the length of the DNA sequence be the independent variable,the SNR under Z-curve was calculated to establish the new model, which avoids the constraints of DNA sequence length and Discrete Fourier Transform （DFT） in traditional methods,so it＇s easy to popularize.