基于3-周期性的基因预测  

An approach to gene prediction based on 3-periodicity

作　　者：胡连果 李强[1] 张瑾[1] 张良云[1] 杨涛[1] 骈聪 

机构地区：[1]南京农业大学理学院,南京210095

出　　处：《安徽农业大学学报》2014年第3期351-357,共7页Journal of Anhui Agricultural University

基　　金：国家自然科学基金(11226070)资助

摘　　要：基因预测是DNA序列分析的一项重要任务,真核生物基因外显子序列的识别是生物信息学中的难点,本研究提出一种基于3-周期性小鼠基因的预测方法。基于碱基幅角的功率谱,使小鼠外显子序列的频谱图具有更显著的3-周期性,且内含子的频谱图更加平稳,从而增大了序列中外显子与内含子的信噪比的区分度,使小鼠基因序列的预测获得更好的效果,在平均值法阈值R2下预测准确率达94.53%;在分布图法阈值R1下,准确率达94.50%,敏感性达99.45%。Gene prediction is a significant task, in which the recognization of exon of eukaryotic gene is a great challenge. In this study, a gene prediction method which is based on 3-periodicity in gene sequences was performed. A more prominent peak value in exon and more steady values in intron were shaped in the power spectrum curves based on arguments of the four bases. The distinction degree of SNR of exons and introns was enlarged and a better effect in short gene prediction was presented. In the condition of threshold R2, the accuracy reached to 94.53%; under the threshold R1, the accuracy reached to 94.50%, and the sensibility reached to 99.45%.

关 键 词：功率谱 3-周期性 幅角 信噪比 阈值 

分 类 号：Q811.4[生物学—生物工程]