Related Experiment Video
Updated: Jan 16, 2026
06:50
Author Spotlight: A Computational Approach to Decipher Amino Acid Preferences in Multispecific Protein-Protein Interactions
Published on: January 26, 2024
2.5K
HEPAD: enhancing hemolytic peptide prediction with adaptive feature engineering and diverse sequence descriptors
Sih-Han Chen1, Jen-Chieh Yu1, Yi-Hsiang Lin2
1Graduate Institute of Genomics and Bioinformatics, National Chung Hsing University, Taichung, Taiwan.
BMC Bioinformatics
|October 1, 2025
Summary
HEPAD is a new machine learning tool that accurately identifies potentially hemolytic peptides, which are problematic for drug development. This computational model aids in peptide drug discovery by predicting and flagging these peptides early.
Area of Science:
- Computational biology and bioinformatics
- Drug discovery and development
- Machine learning in medicine
Background:
- Peptides are valuable therapeutic agents due to their selectivity and low cost.
- Hemolytic peptides pose a challenge in peptide drug development by damaging red blood cells.
- Computational identification of hemolytic peptides is crucial for advancing peptide therapeutics.
Purpose of the Study:
- To develop a machine learning predictor, HEPAD, for identifying hemolytic peptides.
- To enhance the accuracy and efficiency of hemolytic peptide prediction in drug discovery.
Main Methods:
- Utilized diverse sequence descriptors for feature encoding of peptides.
- Implemented an adaptive feature engineering method to create customized feature subsets.
- Employed five distinct machine learning algorithms for rigorous validation.
Main Results:
- HEPAD achieved high accuracy, with Matthew's correlation coefficients (MCCs) of 0.973, 0.643, and 0.609 on independent datasets.
- Demonstrated significant improvements over existing methods, ranging from 1.9% to 13.3% in MCC.
- Data visualization confirmed the effectiveness of customized feature subsets in distinguishing hemolytic peptides.
Conclusions:
- HEPAD provides an efficient method for identifying potential hemolytic peptides.
- The tool accelerates experimental processes in peptide drug discovery.
- Source code, datasets, and models are publicly available for further research.
