Welcome to iCarPS

    iCarPS is a server for the prediction of carbonylation sites based on omputational method. A novel feature encoding scheme including 9 physicochemical properties was proposed to formulate carbonylated protein and non-carbonylated protein samples, and then F-score was adopted to exclude noise and redundant information. Besides, the incremental feature selection (IFS) curve was used to determine the optimal feature subsets which could produce peak values of AUC. As a result, the total accuracies rate of 87.08%, 87.09%, 86.35%, 85.56% were achieved in predicting the carbonylation sites of K, P, R, and T residues, respectively, based on the 10-fold cross-validation test by using random forest (RF) algorithm. We hope that the webserver will become a useful tool for carbonylation analysis and further experimental researches.