Related Experiment Video
Updated: Apr 17, 2026
Mapping Bacterial Functional Networks and Pathways in Escherichia Coli using Synthetic Genetic Arrays
Published on: November 12, 2012
Prediction of minimum inhibitory concentration of antibacterial peptides using geometric graph networks and dynamic
Yannan Bin1, Jinyu Li2, Ruifen Cao2
1The Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, School of Life Sciences and Medical Engineering, Anhui University, Jiulong Road 111, Hefei, Anhui 230601, China.
Abstract:
Antibiotic resistance has reduced the effectiveness of traditional antibiotics for public health needs. Antibacterial peptides (ABPs) hold clinical value due to their inherent antimicrobial activity and resistance to degradation. Existing deep learning approaches for predicting the minimum inhibitory concentration (MIC) of ABPs typically rely on a single modality or a single scale, limiting their ability to capture ABP complexity. Moreover, these methods cover only a narrow range of bacterial species. To address these challenges, we propose the geometric graph network (GGN)-ABPMIC model, which integrates multi-scale structural features at the atomic and residue scales with peptide sequence features. These features are aggregated through a GGN enhanced with geometric vector perceptrons. To enrich the feature representation, we also incorporate descriptors that quantify the physicochemical properties of ABPs and fuse them with the GGN-processed features. The model predicts the MIC values for ABPs across datasets spanning 10 bacterial species. For training and robustness, we introduce a multi-stage dynamic-weight hybrid loss that combines mean-squared error (MSE), the Huber loss, and a contrastive learning loss. Across the 10 species, GGN-ABPMIC achieves a mean MSE of 0.221, a mean R$^2$ of 0.564, and a mean Pearson correlation coefficient of 0.749, outperforming existing approaches. Additional validate via case studies on supplementary Escherichia coli ABP sequences indicates that the MIC predictions of GGN-ABPMIC are closer to the true values, demonstrating strong predictive performance and generalization for ABP activity. The data and codes for GGN-ABPMIC are available at https://github.com/LiJinYu1231/GGN-ABPMIC/tree/master.
Insights
This study introduces a new deep learning model, GGN-ABPMIC, to predict antibacterial peptide (ABP) activity against antibiotic resistance. The model accurately forecasts minimum inhibitory concentrations (MICs) across diverse bacterial species, offering a promising tool for developing new antimicrobials.
Area of Science:
- Biochemistry
- Computational Biology
- Drug Discovery
Background:
- Antibiotic resistance poses a significant global health threat, diminishing the efficacy of conventional antibiotics.
- Antibacterial peptides (ABPs) present a viable alternative due to their potent antimicrobial properties and resistance to degradation.
- Current deep learning models for predicting ABP efficacy are limited by single-modality/scale approaches and narrow bacterial coverage.
Purpose of the Study:
- To develop an advanced deep learning model for accurate prediction of minimum inhibitory concentrations (MICs) of antibacterial peptides (ABPs).
- To address limitations of existing models by integrating multi-scale structural and physicochemical features of ABPs.
- To enhance the prediction scope across a wider range of bacterial species.
Main Methods:
- Proposed the geometric graph network (GGN)-ABPMIC model, integrating atomic, residue, and sequence-level features.
- Employed a GGN enhanced with geometric vector perceptrons for feature aggregation.
- Incorporated physicochemical descriptors and fused them with GGN features, utilizing a hybrid loss function for training.
Main Results:
- Achieved a mean MSE of 0.221, mean R$^2$ of 0.564, and mean Pearson correlation coefficient of 0.749 across 10 bacterial species.
- Demonstrated superior performance compared to existing methods in predicting ABP MIC values.
- Case studies confirmed the model's strong predictive accuracy and generalization capabilities for ABP activity.
Conclusions:
- The GGN-ABPMIC model effectively predicts ABP activity across multiple bacterial species, outperforming current methods.
- The integration of multi-scale structural and physicochemical features enhances predictive accuracy.
- This approach offers a valuable tool for accelerating the discovery and development of novel antibacterial peptides to combat antibiotic resistance.
Related Concept Videos
01:27Pharmacodynamic Models: Overview
02:26Protein Networks
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
01:15Pharmacodynamic Models: Linear Concentration–Effect Model
01:15Determination of Multiple Dosing Parameters: Steady-State, Minimum and Maximum Concentrations
02:04Protein-protein Interfaces
01:20Estimation of k and VD of Aminoglycosides