Related Experiment Video
Updated: Jul 2, 2025
13:49
Semi-automated Biopanning of Bacterial Display Libraries for Peptide Affinity Reagent Discovery and Analysis of Resulting Isolates
Published on: December 6, 2017
11.4K
AntiBP3: A Method for Predicting Antibacterial Peptides against Gram-Positive/Negative/Variable Bacteria.
Nisha Bajiya1, Shubham Choudhury1, Anjali Dhall1
1Department of Computational Biology, Indraprastha Institute of Information Technology, Okhla Phase 3, New Delhi 110020, India.
Antibiotics (Basel, Switzerland)
|February 23, 2024
Summary
This study introduces a novel alignment-free method for predicting antibacterial peptides (ABPs) effective against gram-positive, gram-negative, and gram-variable bacteria, outperforming existing approaches.
Area of Science:
- Computational Biology
- Bioinformatics
- Drug Discovery
Background:
- Existing antibacterial peptide (ABP) prediction methods often target specific bacterial types (gram-positive or gram-negative).
- There is a need for a universal prediction method applicable to a broader range of bacteria, including gram-variable types.
Purpose of the Study:
- To develop a robust and accurate method for predicting antibacterial peptides (ABPs) effective against gram-positive, gram-negative, and gram-variable bacteria.
- To overcome the limitations of sensitivity and specificity in existing ABP prediction approaches.
Main Methods:
- Initial development of alignment-based (BLAST) and motif-based approaches for ABP prediction, noting their limitations.
- Implementation of alignment-free machine/deep learning models utilizing diverse peptide features like composition, terminal residue binary profiles, and fastText word embeddings.
- Rigorous model evaluation using five-fold cross-validation and an independent test dataset.
Main Results:
- Machine learning models achieved high performance, with the amino acid binary profile model yielding maximum AUCs of 0.93 (gram-positive), 0.98 (gram-negative), and 0.94 (gram-variable) on the independent dataset.
- The developed method demonstrated superior performance compared to existing approaches when evaluated on the independent dataset.
- The models exhibited improved sensitivity and precision for predicting ABPs across different bacterial types.
Conclusions:
- The novel alignment-free machine learning approach provides a significant advancement in predicting antibacterial peptides (ABPs) for diverse bacterial species.
- The developed prediction tool, available as a web server, standalone package, and pip package, facilitates the discovery of novel peptide-based therapeutics.
- This method offers a promising solution for developing broad-spectrum antibacterial agents.
