Prediction of Lysine-Malonylation Sites via Sequential and Physicochemical Features

Abstract

Lysine Malonylation is Post Translational Modification responsible for Type2 diabetes, Cancer etc. It is a challenging problem as the data from kmal studies are highly imbalanced. In this work we propose Hybrid sampling a combination of RUS and SMOTE at certain ratios in combination with mutual information feature selection, Balanced Random Forest to solve this problem.