Amino acid composition and classification of acidic and alkaline enzymes / 生物工程学报

It is of theoretical and practical significance to understand the mechanism of enzyme adaptation to acidic and alkaline environments and classification of them based on sequence information. In present work, the amino acid composition of 105 acidic and 111 alkaline enzyme sequences was systematically analyzed. Acidic enzymes contained significantly more Trp, Tyr, Thr and Ser, whereas less Glu, Lys, Met and Arg. On the other hand, alkaline enzymes have slightly more Trp, Ala and Cys, whereas less Lys, Arg and Glu. The amount of Ala, Glu, Leu, Asn, Arg, Ser and Thr in acidic and alkaline enzymes varied largely. Hence, a weighted amino acid composition method was developed for the discrimination of acidic and alkaline enzymes. Using the back-check and the 5-fold cross validation methods, the overall accuracy could reach 86.1% and 83.3%, respectively. A new method to classify acidic and alkaline enzymes based on their sequences was established.

Discriminating acidic, neutral and alkaline enzymes with secondary structure amino acid composition / 生物工程学报

In this work, we systematically analyzed the secondary structure amino acid compositions of acidic and alkaline enzymes and compared them with neutral ones. We found that the propensity of the individual residues to participate in secondary structures and the consistently higher composition of neutral and tiny residues might be the general stability mechanisms for their adaptation to pH extremes. Based on this, we presented a secondary structure amino acid composition method for extracting useful features from sequence. The overall prediction accuracy evaluated by the 10-fold cross-validation reached 80.3%. Comparing our method with other feature extraction methods, the improvement of the overall prediction accuracy ranged from 9.4% to 18.7%. The random forests algorithm also outperformed other machine learning techniques with an improvement ranging from 2.7% to 21.8%.

Prediction of lipases types by different scale pseudo-amino acid composition / 生物工程学报

Lipases are widely used enzymes in biotechnology. Although they catalyze the same reaction, their sequences vary. Therefore, it is highly desired to develop a fast and reliable method to identify the types of lipases according to their sequences, or even just to confirm whether they are lipases or not. By proposing two scales based pseudo amino acid composition approaches to extract the features of the sequences, a powerful predictor based on k-nearest neighbor was introduced to address the problems. The overall success rates thus obtained by the 10-fold cross-validation test were shown as below: for predicting lipases and nonlipase, the success rates were 92.8%, 91.4% and 91.3%, respectively. For lipase types, the success rates were 92.3%, 90.3% and 89.7%, respectively. Among them, the Z scales based pseudo amino acid composition was the best, T scales was the second. They outperformed significantly than 6 other frequently used sequence feature extraction methods. The high success rates yielded for such a stringent dataset indicate predicting the types of lipases is feasible and the different scales pseudo amino acid composition might be a useful tool for extracting the features of protein sequences, or at lease can play a complementary role to many of the other existing approaches.