Foundation Artificial Intelligence Models in Animal Biotechnology: From Protein Structure Prediction to Genomic Language Models and Autonomous Laboratory Systems
DOI:
https://doi.org/10.14741/ijab/v.16.1.1Keywords:
Foundation AI models, AlphaFold, Protein language models, Livestock genomics, CRISPR design, Vaccine development, Large language models, Genomic AI, Autonomous laboratory, Synthetic biologyAbstract
Foundation artificial intelligence models — large neural networks pre- trained on vast, diverse biological datasets that can be fine- tuned or prompted for a wide range of downstream tasks — are transforming the pace and scope of discovery and application across all domains of biological science. In the period 2021–2026, the introduction of AlphaFold2, AlphaFold3, ESM-2, the Evo genomic language model, the Nucleotide Transformer, and numerous protein design models (RFdiffusion, Genie2, ProteinMPNN) has provided the life sciences community with a suite of AI tools that compress decades of conventional structural biology and functional genomics work into hours of computation. This review examines the transformative impact of foundation AI models on animal biotechnology, with particular emphasis on applications to livestock species. The major areas covered include protein structure prediction for livestock disease proteins (viral capsids, bacterial surface proteins, host receptors) and its impact on rational vaccine design; genomic and epigenomic language models for prediction of regulatory sequence function, variant effects, and gene expression in livestock genomes; AI- accelerated CRISPR guide RNA design and off- target prediction; foundation model- powered drug and vaccine target discovery workflows; AI integration in laboratory automation (liquid handling robots, imaging AI, automated phenotyping); and large language model (LLM) applications in scientific literature mining, protocol generation, and hypothesis- driven research planning. The review critically evaluates the current capabilities and limitations of foundation AI models in livestock biotechnology contexts, where training data for livestock species remain substantially less abundant than for human biomedical applications, addressing strategies for data augmentation, cross- species transfer learning, and livestock- specific model fine- tuning. Emerging regulatory and biosafety frameworks for AI- designed biological entities (AI- designed vaccines, gene circuits, engineered proteins) are reviewed, along with the research infrastructure and skills development investments needed to realise the full potential of AI- driven animal biotechnology.
