DynML-Net: a porcine enteric virus identification network based on protein language models and a dynamic heterogeneous multi-branch architecture
- 1School of Artificial Intelligence, Anhui Agricultural University
- 2College of Veterinary Medicine, Anhui Agricultural University
Abstract
Background Porcine Enteric Viruses (PEVs) pose a serious threat to the global pig industry, leading to substantial economic losses. However, traditional clinical diagnostic methods are limited by high costs, long turnaround times, and difficulties in rapidly addressing challenges such as co-infections with multiple viruses and the emergence of different viral species. To overcome the limitations of existing approaches in recognizing viruses under complex clinical conditions, poor adaptability to variant strains, and the insufficient extraction of key features, this study proposes a deep learning network integrating the ESM-2 protein language model and a multi-branch heterogeneous architecture, named DynML-Net. Results The model adaptively integrates decision outputs representing local architectural patterns, global dependencies, and long-sequence evolutionary features through a dynamic gating mechanism, and enhances classification performance via a deep mutual learning (DML) strategy. Experimental results demonstrate that DynML-Net achieves accuracies of 94.52% and 89.16% for binary and multiclass classification tasks, respectively, outperforming current mainstream models, while exhibiting robust performance under conditions of sample imbalance, data reduction, and heterologous viral datasets scenarios. Conclusions This study demonstrated the efficacy of DynML-Net in PEVs identification and classification tasks, and based on DynML-Net, a Streamlit-based locally deployable system was developed to facilitate a closed-loop workflow from sequence upload to automated identification, providing an accurate and efficient computational solution for the rapid detection of PEVs.
Related
- Rapid identification of porcine bacterial pathogens using an artificial intelligence Raman microbial analyzer for biology
- TriTower-m6Am: A triple-tower heterogeneous deep learning architecture integrating semantic, sequential, and structural information for mRNA N6,2′-O-dimethyladenosine site prediction
- Large Language Model Framework for Structured Biomarker Extraction and Triple-Negative Breast Cancer Identification From Pathology Reports: Development and Evaluation Study.
- Muc2-dependent microbial colonization of the jejunal mucus layer is diet sensitive and confers local resistance to enteric pathogen infection
- Enteric ChAT-Expressing Neurons Mediate the Protective Effect of Lactobacillus plantarum in Inflammatory Bowel Disease.
BibTeX
@article{chen2026dynmlnet,
title = {{DynML-Net: a porcine enteric virus identification network based on protein language models and a dynamic heterogeneous multi-branch architecture}},
author = {Qingwei Chen and Xin Xing and Shumei Li and Ying Shao and Xiangjun Song and Zhao Qi},
journal = {BMC Genomics},
year = {2026},
month = jul,
doi = {10.1186/s12864-026-13227-3},
url = {https://doi.org/10.1186/s12864-026-13227-3},
note = {CC BY},
}
RIS
TY - JOUR
TI - DynML-Net: a porcine enteric virus identification network based on protein language models and a dynamic heterogeneous multi-branch architecture
AU - Qingwei Chen
AU - Xin Xing
AU - Shumei Li
AU - Ying Shao
AU - Xiangjun Song
AU - Zhao Qi
PY - 2026
DA - 2026/07/01
JO - BMC Genomics
DO - 10.1186/s12864-026-13227-3
UR - https://doi.org/10.1186/s12864-026-13227-3
AB - Background Porcine Enteric Viruses (PEVs) pose a serious threat to the global pig industry, leading to substantial economic losses. However, traditional clinical diagnostic methods are limited by high costs, long turnaround times, and difficulties in rapidly addressing challenges such as co-infections with multiple viruses and the emergence of different viral species. To overcome the limitations of existing approaches in recognizing viruses under complex clinical conditions, poor adaptability to variant strains, and the insufficient extraction of key features, this study proposes a deep learning network integrating the ESM-2 protein language model and a multi-branch heterogeneous architecture, named DynML-Net. Results The model adaptively integrates decision outputs representing local architectural patterns, global dependencies, and long-sequence evolutionary features through a dynamic gating mechanism, and enhances classification performance via a deep mutual learning (DML) strategy. Experimental results demonstrate that DynML-Net achieves accuracies of 94.52% and 89.16% for binary and multiclass classification tasks, respectively, outperforming current mainstream models, while exhibiting robust performance under conditions of sample imbalance, data reduction, and heterologous viral datasets scenarios. Conclusions This study demonstrated the efficacy of DynML-Net in PEVs identification and classification tasks, and based on DynML-Net, a Streamlit-based locally deployable system was developed to facilitate a closed-loop workflow from sequence upload to automated identification, providing an accurate and efficient computational solution for the rapid detection of PEVs.
SN - 1471-2164
ER -