A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features

Abstract

The thermostability of proteins is a key factor considered during enzyme engineering, and finding a method that can identify thermophilic and non-thermophilic proteins will be helpful for enzyme design. In this study, we established a novel method combining mixed features and machine learning to achieve this recognition task. In this method, an amino acid reduction scheme was adopted to recode the amino acid sequence. Then, the physicochemical characteristics, auto-cross covariance (ACC), and reduced dipeptides were calculated and integrated to form a mixed feature set, which was processed using correlation analysis, feature selection, and principal component analysis (PCA) to remove redundant information. Finally, four machine learning methods and a dataset containing 500 random observations out of 915 thermophilic proteins and 500 random samples out of 793 non-thermophilic proteins were used to train and predict the data. The experimental results showed that 98.2% of thermophilic and non-thermophilic proteins were correctly identified using 10-fold cross-validation. Moreover, our analysis of the final reserved features and removed features yielded information about the crucial, unimportant and insensitive elements, it also provided essential information for enzyme design.

Keywords: machine learning methods; mixed features; non-thermophilic protein; reduced amino acids; thermophilic protein.

FIGURE 1

The whole framework of the proposed method in this manuscript.

FIGURE 2

The figure of model performance. (A) The first two dimensions of the result of compression characteristics of the TSNE method; (B) the figure of the ultra-classification surface of SVM method; (C) The accuracy values of four different models; (D) the comparison results with other methods.

FIGURE 3

The comparison results of experiments. (A) The receiver operation characteristic (ROC) curve of three methods; (B) the results of experiments over the database (Fan et al., 2016).

FIGURE 4

The critical and removed features in the proposed method: (A) the most important features; (B) the deleted amino acid frequency features; (C) the deleted reduced-depiptides (I); (D) the deleted reduced-depiptides (II). The symbol “*” means any one of the 20 amino acids, it may be “A”, “C”, “P”, or others. Besides, “**” has the same meaning; it represents a two-letter combination of 20 amino acids, “AA”, “DC”, “VP”, for example.