摘要
启动子是调节基因表达的重要元件,对其的研究对于阐明基因转录调控机制具有重要意义。作者依据RNA聚合酶Ⅱ启动子序列特性选取高效的特征提取方法,构建了基于粒子群优化的支持向量机(particle swarm optimization-support vector machine,PSO-SVM)新方法,用以识别真核生物基因RNA聚合酶Ⅱ启动子。结合5-折交叉检验方法,得到启动子-外显子、启动子-内含子和启动子-基因间序列的分类准确率分别为97.1%、96.7%和98.8%,其马修斯相关系数分别为0.962、0.934和0.976。结果说明,对比其它启动子识别方法,PSO-SVM方法更能有效地识别真核生物基因启动子。
Promoters play important roles in gene expression. And the research in promoters has great significance to illuminate the regulatory mechanisms of gene transcription. According to the nucleotide distribution features of gene promoter regions, a support vector machine algorithm based on swarm intelligence optimization was constructed to identify the RNA Pol-II promoter of gene with effectively feature extracting method. With 5-fold cross validation, the accuracies of the new method in classifications of promoter-exon, promoter-intron and promoter-intergenic were 97.1%, 96.7% and 98.8% respectively, and the Mathews correlation coefficients were 0.962, 0.934 and 0.976. The results indicate that, support vector machine based swarm intelligence optimization algorithm can identify the promoter sequence of gene in eukaryotes more effectively than other prediction methods.
出处
《生物物理学报》
CAS
CSCD
北大核心
2015年第2期143-153,共11页
Acta Biophysica Sinica
基金
国家自然科学基金项目(11401311)
江苏省自然科学基金项目(BK20141358)~~
关键词
位置关联权重矩阵
粒子群优化
支持向量机
随机森林
极限学习机
启动子识别
Position correlation weight matrix
Particle swarm optimization
Support vector machine
Randomforest
Extreme learning machine
Promoter prediction
