Related Experiment Video
Updated: Aug 20, 2025
08:37
The Application of Open Searching-based Approaches for the Identification of Acinetobacter baumannii O-linked Glycopeptides
Published on: November 2, 2021
2.3K
LABAMPsGCN: A framework for identifying lactic acid bacteria antimicrobial peptides based on graph convolutional
Tong-Jie Sun1, He-Long Bu1, Xin Yan1
1College of Computer and Information Engineering, Inner Mongolia Agricultural University, Hohhot, China.
Frontiers in Genetics
|November 21, 2022
Summary
Researchers developed a graph convolutional neural network (GCN) to efficiently predict lactic acid bacteria antimicrobial peptides (LABAMPs). This new model accelerates the identification of these vital compounds for food and agriculture applications.
Area of Science:
- Bioinformatics
- Computational Biology
- Food Science
Background:
- Lactic acid bacteria antimicrobial peptides (LABAMPs) are crucial polypeptides with applications in food production and agriculture.
- Current experimental screening methods for LABAMPs are time-consuming and labor-intensive.
- There is a significant need for computational models to expedite the identification of LABAMPs.
Purpose of the Study:
- To design and implement a novel graph convolutional neural network (GCN) framework for the accurate prediction of LABAMPs.
- To overcome the limitations of traditional experimental screening methods by developing an efficient computational approach.
Main Methods:
- A heterogeneous graph was constructed using amino acids, tripeptides, and their interrelationships.
- A graph convolutional neural network (GCN) was employed to learn embedded words and sequence weights within the graph structure.
- The GCN model was trained and validated using supervised learning with input sequence labels.
Main Results:
- 10-fold cross-validation on two training datasets yielded high accuracies of 0.9163 and 0.9379, outperforming other machine learning and GNN algorithms.
- Independent testing on a separate dataset achieved accuracies of 0.9130 and 0.9291.
- The GCN model demonstrated superior performance, exceeding the best online webserver methods by 1.08% and 1.57%.
Conclusions:
- The developed graph convolutional neural network framework provides an effective and efficient method for identifying lactic acid bacteria antimicrobial peptides (LABAMPs).
- This computational approach significantly reduces the time and effort required for LABAMP screening, paving the way for broader applications in food and agriculture.
