基于伪氨基酸和支持向量机的蛋白质亚细胞定位预测

用电子—离子伪势能(E IIP)对蛋白质序列数字化,经离散傅立叶变换(DFT)后,取5个最高幅值对应的频率和20种氨基酸在序列中所占的百分比组成伪氨基酸。用支持向量机(SVM)方法得到分类的模型,并用几个标准的测试方法测试模型的性能。自身一致性测试和Jackkn ife测试均取得高的预测准确率,独立数据集测试的准确率超过80%。和之前报道的方法相比,本方法具有较高的预测准确率。

Protein subcellular location prediction based on pseudo amino acid composition and support vector machines

Select electron-ion interaction potential(EIIP) to represent digital protein.After the digital proteins had been processed with DFT,the highest 5 amplitudes were selected,then they were applied to construct pseudo amino acid composition.The proteins subcellular locations were predicted using support vector machines(SVM) and pseudo amino acid composition.High success rates thus obtained by both self-consistency test and jackknife test.The overall predictive accuracy about 80% had been achieved in independent dataset test.The result demonstrated that this new method is practical.