连续小波变换在蛋白质结构预测中的应用

Prediction of Protein Structure by Continuous Wavelet Transform

将代码为3pgal蛋白质的氨基酸序列映射为疏水值序列,在合适的尺度下,通过连续小波变换方法可对其非规则二级结构进行预测,准确率为83.3％。从PDBsum数据库中随机抽取100个蛋白质作为测试对象,经本法处理后,1 853个非规则二级结构中有1 424个能被准确预测到,平均预测准确率为76.8％。结果表明:该法可较好地预测蛋白质的非规则二级结构,具有极大的发展前景。

Non-standard secondary structures can be predicted by using continuous wavelet transform (CWT) under the appropriate dilation after the amino acids of 3pgal protein are transformed into sequences of hydrophobic values per residue, the predicted accuracy is about 83.3 % . One hundred proteins from PDBsum database were chosen randomly as the test objects, 1 424 non-standard secondary structures can be predicted by this method, comparing with that of the sum 1 853, the average accuracy is 76.8% . The result exhibits that CWT is an effective tool to predict the non-standard secondary structures of proteins and has a tremendous development foreground.