مقالات پذیرفته شده در نهمین کنگره بین المللی زیست پزشکی
Artificial Intelligence in Veterinary Virology: Genomic Surveillance, Pathogenesis Modeling, and Predictive Epidemiology
Artificial Intelligence in Veterinary Virology: Genomic Surveillance, Pathogenesis Modeling, and Predictive Epidemiology
Seyedeh Fatemeh Angoshtan,1,*
1. Department of clinical Sciences, Faculty of Veterinary Medicine, Semnan University, Semnan, Iran
Introduction: The rise in complexity related to viral diseases of livestock is creating a high-tech need to understand, monitor, and manage transmission. These pathogens exhibit very high levels of heterogeneous mutation in the population, changes or selection of tissue tropism, and complicated dynamics of transmission that often include wildlife reservoirs and/or vector species. While traditional methods used in virology are beneficial, they are not robust enough to effectively manage the range of genomic and ecological data involved in this field of research. As a result, artificial intelligence (AI) has become an accepted and powerful integration tool that can enhance mechanistic understanding, predict epidemic patterns, and optimize molecular surveillance in veterinary virology. This review will highlight how artificial intelligence is changing veterinary virology through genomic surveillance, host-pathogen modeling, and outbreak prediction of viral diseases in livestock.
Methods: We searched PubMed, Scopus, and Web of Science databases for studies published from 2018 to 2025. The keywords used to filter the search are as follows: “Viruses”, “Artificial Intelligence”, “Machine Learning”, “Host-Pathogen Interactions”, “Metagenomics”, “Zoonoses”, “Vector-Borne Diseases”, and “Disease Outbreaks”. Boolean operators were used to combine studies on both livestock viral pathogens and AI-informed geospatial modeling and ecological surveillance.
Results: Genomic surveillance has benefited immensely through the use of artificial intelligence to accurately predict recombination and mutation hotspots and track antigenic drift in fast-evolving RNA viruses. Deep learning architecture designed from viral sequence databases has allowed for the identification of new viral lineages from a metagenomics perspective, in particular, for vector species such as Culicoides and mosquitoes. In the area of pathogenesis modeling, machine learning based merging of transcriptomics and proteomics data has shown unique host response patterns suitable for differentiating neuroinvasive EHV-1 strains from non-neuroinvasive strains, and identified immunotolerance pathways during chronic BVDV infection. Predictive epidemiology studies have recently harnessed climate-related neural networks to forecast vector-borne outbreaks with high spatial and temporal resolution for early warning systems in endemic regions.
Conclusion: Despite the advancements in AI, issues still exist regarding data standardization, model interpretability, and the application of these models at the field level. To maximize the use of AI, we require solutions to concerns related to poor annotated datasets, limited transdisciplinary collaboration, and ethical issues in relation to algorithmic bias. The AI application in veterinary virology represents a paradigm shift in diagnostics, a movement from reactive diagnosis to proactive. Future efforts should shift towards a hybrid approach using AI, high-resolution omics data, ecological surveillance, and experimental virology to develop systems to detect, monitor, and manage viral threats. The combination of computational intelligence with veterinary medicine represents a real paradigm shift in the potential to minimize the burden from viral disease in livestock, maximizing the global One Health effort toward improving health resilience.