基于信噪比和频谱特征改进的基因预测方法研究

Research of Prediction Method Based on the Characteristics of Gene's SNR and Spectrum

指出了正确识别定位DNA序列中外显子区间是当前生物信息学最基础的问题之一.通过介绍DNA序列的3-周期性、信噪比等概念,引入了固定长度滑动窗口和移动信噪比曲线识别两种基因识别算法,在综合两种算法的复杂度和准确性优势的基础上,提出了基于信噪比特征的基因识别改进算法,并通过实验验证了该算法的有效性与准确性.

Identifying and positioning exons intervals of the DNA sequence correctly is one of the most basic issues in current bioinformatics. In this paper, through describing the 3-periodicity of DNA sequences, SNR and other concepts, we introduce two gene identification algorithms which are called fixed-length sliding windows and mobile SNR curve. After synthesizing the advantage about complexity and accuracy of the two algorithms, we propose an improved identification algorithm based on SNR and spectrum. Finally, the experiments are used to verify the validity and accuracy of the algorithm.