The goal of this work is to develop a powerful and easily interpretable method for identifying and analyzing therapeutic peptide, especially in tumor homing and hemolytic peptides. The proposed method is also utilized to design highly efficacious tumor homing peptides with less hemolytic activity to facilitate better drug delivery system for cancer. Consequently, the informative features are used for providing good understanding of insight into the characteristics of the tumor homing and hemolytic peptides. In this work, the proposed prediction model aims to maximize both the simplicity and interpretability of used features and prediction method. Finally, a user-friendly web-server is established to help experimental scientists who want to use the prediction method without the interest to follow the mathematical and statistical details.
