HemoFinder is a newly introduced tool in dbAMP 3.0, designed to predict the hemolytic activity and half-life of antimicrobial peptides (AMPs). This tool assists users in identifying AMPs with better drug-like potential, supporting peptide-based drug discovery.
HemoFinder follows a structured machine learning pipeline to analyze AMPs.
HemoFinder integrates five widely used peptide descriptors to capture the sequence and physicochemical characteristics of AMPs:
- AAC: Amino Acid Composition
- DPC: Dipeptide Composition
- PAAC: Pseudo Amino Acid Composition
- CKSAAGP: Composition of k-spaced Amino Acid Group Pairs
- PHYC: Physicochemical Properties
These features are fed into a set of machine learning base learners, including:
- XGBoost
- Decision Tree
- K-Nearest Neighbors (KNN)
- Random Forest
The final classifier is built using a soft voting strategy, enhancing prediction robustness by leveraging the strengths of each base learner. This ensemble approach allows HemoFinder to accurately identify peptides with high hemolytic potential.
We have developed a user-friendly website interface for HemoFinder, making it easy for researchers and developers to access the tool without requiring any programming knowledge.
🔗 Try it now: https://awi.cuhk.edu.cn/~dbAMP/HemoFinder.php
For more information or technical support, please visit the dbAMP Homepage or contact the development team.