Sequence Based Prediction of Antioxidant Proteins Using a Classifier Selection Strategy

Abstract

Antioxidant proteins perform significant functions in maintaining oxidation/antioxidation balance and have potential therapies for some diseases. Accurate identification of antioxidant proteins could contribute to revealing physiological processes of oxidation/antioxidation balance and developing novel antioxidation-based drugs. In this study, an ensemble method is presented to predict antioxidant proteins with hybrid features, incorporating SSI (Secondary Structure Information), PSSM (Position Specific Scoring Matrix), RSA (Relative Solvent Accessibility), and CTD (Composition, Transition, Distribution). The prediction results of the ensemble predictor are determined by an average of prediction results of multiple base classifiers. Based on a classifier selection strategy, we obtain an optimal ensemble classifier composed of RF (Random Forest), SMO (Sequential Minimal Optimization), NNA (Nearest Neighbor Algorithm), and J48 with an accuracy of 0.925. A Relief combined with IFS (Incremental Feature Selection) method is adopted to obtain optimal features from hybrid features. With the optimal features, the ensemble method achieves improved performance with a sensitivity of 0.95, a specificity of 0.93, an accuracy of 0.94, and an MCC (Matthew’s Correlation Coefficient) of 0.880, far better than the existing method. To evaluate the prediction performance objectively, the proposed method is compared with existing methods on the same independent testing dataset. Encouragingly, our method performs better than previous studies. In addition, our method achieves more balanced performance with a sensitivity of 0.878 and a specificity of 0.860. These results suggest that the proposed ensemble method can be a potential candidate for antioxidant protein prediction. For public access, we develop a user-friendly web server for antioxidant protein identification that is freely accessible at http://antioxidant.weka.cc.

Tumstatin fragment selectively inhibits neutrophil infiltration in experimental asthma exacerbation

“…This may in part be explained by antioxidative capabilities of these class of peptides . CP17 is predicted to be an antioxidant protein (confidence 0.945) with SP17 showing almost identical capabilities (confidence 0.935; Table ) . This indicates a certain capacity for these peptides to act as ROS scavengers.…”

“…[43][44][45] CP17 is predicted to be an antioxidant protein (confidence 0.945) with SP17 showing almost identical capabilities (confidence 0.935; Table 2). 46 This indicates a certain capacity for these peptides to act as ROS scavengers. Secondly, and highly supportive of our initial hypothesis, the CP17-specific reduction in the ROS production rate indicates that this function is conferred in the sequence of CP17 (acting via a yet unknown receptor).…”

Structure and function of a novel antioxidant peptide from the skin of tropical frogs

“…They combine low-molecular-weight, high activity, easy absorption, and much reduced negative side effects because of fast tissue and blood clearance [69]. The bioprospection of this class of molecules is of great interest due to its wide applications in different industries, including cosmetics [70] or food industry to control lipid oxidation that produce undesirable offflavors and potentially toxic reactions products [71]. Moreover, they might also be used as leading templates for designing novel molecules for preventing neurodegeneration.…”

Identification of Antioxidant Proteins With Deep Learning From Sequence Information

“…To evaluate the predictive performance of the final model, we compared our proposed model IDAod with AodPred (Feng et al, 2016 ), other frequently used classifiers like logistic, decision tree, random forest implemented in WEKA (Frank et al, 2004 ), and the ensemble model proposed by Zhang et al (Zhang et al, 2016 ). In order to facilitate the comparison of experimental results, we reproduced AodPred in 10-fold cross validation.…”

“…Besides, its feature selection method ANOVA may be oversimplified and the model was trained on single g-gap feature. Afterward an ensemble model was adopted based on secondary structure information, mutation probability and solvent accessibility (Zhang et al, 2016 ). The model achieved accuracy of 86.3% and sensitivity of 87.8% on independent testing dataset.…”