蛋白质二级结构的协同训练预测方法

Protein secondary structure co-training prediction method

针对蛋白质二级结构机器学习预测方法,忽略氨基酸疏水性特征以及氨基酸之间的长程作用和准确率不高的现状,进行了比较实验分析。采用氨基酸对应的疏水能值替换蛋白质中相应的氨基酸,得到疏水能值的序列。实验结果表明,用长的疏水能值序列训练BP网络,对长程作用起主导的E结构的预测效果好。由于Pro-file编码特征和疏水能值特征是独立的冗余视图,基于协同训练思想,提出Co-training算法。该算法的主要步骤是在Profile特征空间训练SVM分类器,在疏水性特征空间训练BP神经网络分类器,协同对氨基酸二级结构进行预测;SVM分类器和BP分类器有分歧的样本,基于主动选择思想,分析分类器以及特征空间的特点,定义质疑样例和可信样例,给予两个分类器不同的优先级进行仲裁。实验表明,Co-training方法有较高的准确性,对短程起主导的E结构和长程起主导的H结构预测准确率都有所提高。

Machine learning based protein secondary structure prediction methods suffered low prediction accuracy because they ignored the amino acid hydrophobic property and the interaction between far away amino acids.In order to solve this problem,comparative experiments had been done.A sequence of hydrophobic value could be build by replacing the amino acid by its hydrophobic value.Experiments show that the BP neural network using long amino hydrophobic value sequence works well in prediction of E structure which is controlled mainly by long amino acid-amino acid interaction.Because both the Profile space and the hydrophobic energy value space were sufficient and redundant views,this paper proposed a Co-training algorithm.In the proposed algorithm,there were two classifiers.One was SVM classifier trained in Profile space,and the other was BP neural network classifier trained in hydrophobic value space,and they predicted one amino acid secondary structure independently.If these two classifiers had different prediction results with one amino acid,an arbitration rule proposed was employed to make the final decision which was based on an active selecting strategy.Suspected sample and creditable sample were defined according to the characteristics of the classifiers and spaces to arbitrate the controversial prediction results.The experimental results show that the proposed algorithm has higher prediction accuracy both in E structure which controlled mainly by long interaction and H structure which controlled mainly by short interaction than existing algorithms.