Using support vector machine to predict β‐ and γ‐turns in proteins

Abstract

By using the composite vector with increment of diversity, position conservation scoring function, and predictive secondary structures to express the information of sequence, a support vector machine (SVM) algorithm for predicting beta- and gamma-turns in the proteins is proposed. The 426 and 320 nonhomologous protein chains described by Guruprasad and Rajkumar (Guruprasad and Rajkumar J. Biosci 2000, 25,143) are used for training and testing the predictive model of the beta- and gamma-turns, respectively. The overall prediction accuracy and the Matthews correlation coefficient in 7-fold cross-validation are 79.8% and 0.47, respectively, for the beta-turns. The overall prediction accuracy in 5-fold cross-validation is 61.0% for the gamma-turns. These results are significantly higher than the other algorithms in the prediction of beta- and gamma-turns using the same datasets. In addition, the 547 and 823 nonhomologous protein chains described by Fuchs and Alix (Fuchs and Alix Proteins: Struct Funct Bioinform 2005, 59, 828) are used for training and testing the predictive model of the beta- and gamma-turns, and better results are obtained. This algorithm may be helpful to improve the performance of protein turns' prediction. To ensure the ability of the SVM method to correctly classify beta-turn and non-beta-turn (gamma-turn and non-gamma-turn), the receiver operating characteristic threshold independent measure curves are provided.

Using factor analysis scales of generalized amino acid information for prediction and characteristic analysis of β-turns in proteins based on a support vector machine model

“…The ratio of -turns to non--turns in globular proteins is about 1:3, so MCC is the most reliable parameter to evaluate the prediction performance. Table 3 shows that MCC of the present study is larger than that of other prediction methods except Hu et al's SVM [27] and BTNpred method [29]. Besides, Q predicted of this study is the largest one in those prediction methods listed, indicating that the model developed can well deal with the over-prediction problem.…”

“…Kirschner and Frishman developed prediction of -turns and -turn types by a novel bidirectional Elman-type recurrent neural network with multiple output layers (MOLEBRNN), and obtained a MCC of 0.45 [26]. A new SVM-based predictor developed by Hu and Li [27] combined the increment of diversity, position conservation scoring function, and predicted secondary structures to compute the inputs for prediction of -turns and -turns. More recently, two reports [28,29] have achieved a prediction of -turns at over 80% accuracy.…”

“…Currently, a number of outstanding -turn structure prediction methods [4,[18][19][20][21][22][23][24][25][26][27][28][29] have been developed. In the Chou-Fasman method [4], a set of probabilities was assigned to each residue and the conformational parameters and positional frequencies were determined by calculating the relative frequency of a given type of secondary structure.…”

Identifying the Sulfate Ion Binding Residues in Proteins

“…The SVM algorithm has been widely used for prediction of protein structure and function [1,2,20]. SVM has been compiled into software packages.…”

Predicting Βeta-Turns and Βeta-Turn Types Using a Novel Over-Sampling Approach

“…Though predicted secondary structures of protein were effective in predicting β-turns and their types [7] [9] [12] [14] [23], the way of classifying a secondary structure of protein into three states of backbone conformation as α-helix, β-sheet, and coil leads to the circumstance that 50% total number residues are assigned as coils while they are believed to belong to a large set of distinct local structures [25] [26]. Therefore, the structural alphabets (SAs), that are sets of specific prototypes approximating the local protein structure, were developed to overcome this drawback [25].…”