Detecting Human Phosphorylated Protein by Using Class Imbalance Learning and Ensemble Classifier

Abstract

Protein phosphorylation plays a critical role by altering the structural conformation of a protein, causing it to become activated, deactivated, or modifying its function. Encouraged by Qiu’s pioneer work, this paper has developed a new ensemble classifier for detecting human protein phosphorylation. In the predictor, a protein sample is formulated by incorporating the stationary wavelet features derived from the numerical series of protein chain and two types of pseudo amino acid composition (PseAAC). The operation engine to run the predictor is an ensemble classifier formed by fusing nine individual random forest engines via a voting system. It is demonstrated with a larger dataset obtained from Uniprot web. The approach may also has notable impact on prediction of the other PTMs, such as ubiquitination, crotonylation, methylation, and succinylation, among many others.