Porcine Reproductive and Respiratory Syndrome (PRRS) is an infectious disease caused by the Porcine Reproductive and Respiratory Syndrome Virus (PRRSV), primarily characterized by reproductive disorders in sows and respiratory infections in pigs of all ages. The clinical symptoms include diarrhea, fever, abortion, stillbirth, and, in severe cases, death due to secondary complications. The disease may also cause bruising or bluish-purple discoloration of the extremities and ears, leading to its common name, blue ear disease. PRRS has caused substantial economic losses to the global swine industry. Despite continuous advances in vaccine and drug development, viral drug resistance, genetic variability, and potential adverse effects have limited the effectiveness of current prevention and treatment strategies. Antiviral peptides (AVPs) possess potent antiviral activity and have emerged as promising candidates for peptide-based therapeutics in recent years. In the first study, researchers compared the tissues derived from PRRSV-infected pigs with those from healthy pigs and employed mass spectrometry-based proteomics to identify and screen differentially expressed peptides. Meanwhile, three machine learning models, including Random Forest (RF), Support Vector Machine (SVM), and Graph Neural Network (GNN), were trained and assessed using datasets of experimentally validated AVPs and non-AVPs. Among these models, RF demonstrated the best overall predictive performance, achieving an Area Under the Curve (AUC) of approximately 0.95. The trained models were subsequently applied to predict the antiviral potential of candidate peptides identified through proteomic analysis, thereby identifying potential PRRSV-related AVPs. However, these computational predictions require further experimental validation. To further explore AVP prediction methods, this presentation also introduces a study on the GRUATT-AVP model. The study employed multiple sequence encoding methods to convert peptide sequences into numerical features, followed by feature selection. A Gated Recurrent Unit (GRU) combined with an attention mechanism was then utilized to learn sequence information and to improve AVP classification performance. Taken together, these two studies demonstrate that PRRSV-related proteomic analysis can provide biologically relevant candidate peptides, while sequence-based prediction models offer the potential to evaluate their antiviral activity. Integrating these two strategies with subsequent experimental validation could establish a comprehensive screening workflow, facilitating the discovery of additional antiviral peptides against PRRSV.
- Yousaf, W., A., Haseeb, Y., Shen, H., Li, K., Fan, N., Sun, P., Sun, Y., Sun, H., Yang, W., Yin, H., Zhang, Z., Zhang, J., Zhong, J., Wang and N., Huo (2025) Data driven discovery of antiviral peptides against PRRSV using multiple machine learning models. Front. Vet. Sci. 12:1681083. doi: 10.3389/fvets.2025.1681083
- Aziz, M.T., Rupok, A.S., Mahmud, S.M.H., Goh, K.O.M., Hosen, M.F., Shoombuatong, W. and Nandi, D. 2025. GRUATT-AVP: leveraging a novel attention-based gated recurrent unit to advance the accuracy of antiviral peptide prediction. Sci Rep 15, 42509 (2025). https://doi.org/10.1038/s41598-025 26565-1
|
|
