基于统计组合与CpG含量分类的基因预测算法

A gene-prediction algorithm based on the statistical combination and the classification in terms of CpG content

基因预测是研究基因功能、基因表达、基因之间的关联关系以及如何控制基因转录等工作的基础。现有的基因预测方法在预测内部编码外显子方面能达到较高的精度,但在预测5′端外显子方面却存在着不足。本文针对5′端外显子,构建了一种基于统计组合与CpG含量分类的基因预测算法:将基因区域数据根据CpG含量分成2个相对独立的集合,采用统计组合的方法将多种基因预测方法综合在一起进行基因预测研究。实验结果表明该算法提高了基因预测的精度,为进一步研究基因预测提供了一种可能的方案。

Gene-prediction is the foundation of researches on the gene function, the gene expression, the corelationship among genes, and the way of controlling the gene transcription. At present, gene-prediction methods have achieved superior precision on predicting internal coding exons, however, they are inefficient on predicting 5＇-exons. Focusing on predicting 5＇ -exons. A gene-prediction algorithm based on the statistical combination and the classification in terms of CpG content is shown in this paper, in which the data in the gene areas are divided into two relatively independent groups in terms of the CpG content, and then a statistical combination of various methods of gene-prediction is applied to the research. In the experiment, the precision of the gene-prediction is improved by using this algorithm and a possible way for the future research is also provided.