| 引用本文: | 张帅燕,刘毅慧.自动编码器方法的蛋白质二级结构预测[J].生物信息学,2018,16(1):36-42. |
| ZHANG Shuaiyan,LIU Yihui.The prediction of protein secondary structure using auto encoder[J].Chinese Journal of Bioinformatics,2018,16(1):36-42. |
|
| 摘要: |
| 蛋白质二级结构预测是进行蛋白质三级结构研究的重要基础,氨基酸的编码方式对二级结构预测有一定的影响。本文应用了一种新的组合编码方式,即将基团编码与位置特异性打分矩阵(PSSM)进行组合的编码方式。本文中提出的基团编码是针对氨基酸的一种新的编码方式,基团编码是根据氨基酸内部组成来进行编码的,由42位属性组成。本文选取位置特异性打分矩阵(PSSM)中的Blosum62进化矩阵和新的基团编码进行组合,形成新的编码方式。然后对CB513和25pdb两组数据分别进行实验。本文中将采用贝叶斯分类器与自动编码器两种方法来对这种新的编码方式进行实验,然后比较这两种方法得到的两组数据的结果。可以很明显的发现采用自动编码器的实验结果要比使用贝叶斯分类器的结果要高出1.65%。在本文的实验中,可以提取特征的自动编码器的预测准确率更好。 |
| 关键词: 蛋白质二级结构预测 基团编码 PSSM 贝叶斯分类器 自动编码器 |
| DOI：10.3969/j.issn.1672-5565.201705007 |
| 分类号:TP391 |
| 文献标识码:A |
| 基金项目:国家自然科学基金项目(61375013)；山东省自然科学基金项目(ZR2013FM020). |
|
| The prediction of protein secondary structure using auto encoder |
|
ZHANG Shuaiyan, LIU Yihui
|
|
(School of Information, Qilu University of Technology, Jinan 250353,China)
|
| Abstract: |
| The secondary structure prediction is the basis of tertiary structure of protein, and the encoding method has influence on the prediction of secondary structure. A new encoding method composed of radical group encoding and position-specific scoring matrix(PSSM) is proposed. The radical group encoding contains 42 features, which is generated according to amino acids composition. A new encoding method was generated by combining the radical group encoding and the evolution matrix Blosum62. The Bayes classifier and auto encoder are used to predict the secondary structure for CB513 and 25pdb datasets. According to the comparison of the accuracy, the accuracy of auto encoder is higher 1.65% than the accuracy of Bayes classifier. In the experiment, the auto encoder extracting features can achieve higher accuracy. |
| Key words: Protein secondary structure prediction Radical group encoding PSSM Bayes classifier Auto-encoder |
