基于序列信息理论预测线虫基因选择性剪切位点

基因的选择性剪切使得在DNA上一段相同的序列翻译成多个不同的蛋白质序列.选择性剪切的出现把剪切位点分为选择性供体位点、组成性供体位点、选择性受体位点和组成性受体位点.基于EBI中的线虫基因选择性剪切位点数据库,选取不同位点的单碱基频率和序列片段的三联体频数作为参数,利用位置权重矩阵和离散增量结合支持向量机,对选择性剪切位点进行了理论预测.对选择性供体位点和选择性受体位点的预测成功率分别为63.78%和72.63%,特异性分别为68.02%和83.96%.

Prediction of Alternative Splice Sites of C. elegans Genome Based on Sequence Features

Alternative splicing can make a same DNA sequence to product more than one protein sequences. Alternative splicing can divide splice sites into alternative donor sites, constitutive donor sites, alternative acceptor sites and constitutive acceptor sites. Based on C. elegans alternative splicing database in EBI, by selecting mononucleotide frequencies of different sites and 3-mer frequencies of sequence fractions as the informational parameters, alternative splice sites are predicted by the SVM method combined with position weight matrix and increment of diversity. The accuracies of prediction are 63. 78% and 72. 63% for alternative donor sites and alternative acceptor sites, respectively. And their specificities are 68. 02% and 83.96%, respectively.