PreAIP: Computational Prediction of Anti-inflammatory Peptides by Integrating Multiple Complementary Features

Numerous inflammatory diseases and autoimmune disorders by therapeutic peptides have received substantial consideration; however, the exploration of anti-inflammatory peptides via biological experiments is often a time consuming and expensive task. The development of novel in silico predictors is desired to classify potential anti-inflammatory peptides prior to in vitro investigation. Herein, an accurate predictor, called PreAIP (Predictor of Anti-Inflammatory Peptides) was developed by integrating multiple complementary features. We systematically investigated different types of features including primary sequence, evolutionary and structural information through a random forest classifier. The final PreAIP model achieved an AUC value of 0.833 in the training dataset via 10-fold cross-validation test, which was better than that of existing models. Moreover, we assessed the performance of the PreAIP with an AUC value of 0.840 on a test dataset to demonstrate that the proposed method outperformed the two existing methods. These results indicated that the PreAIP is an accurate predictor for identifying anti-inflammatory peptides and contributes to the development of anti-inflammatory peptides therapeutics and biomedical research. The curated datasets and the PreAIP are freely available at http://kurata14.bio.kyutech.ac.jp/PreAIP/.