AI prediction of virus-host protein interactions
Supervisors
Prof David L. Robertson, University of Glasgow
Dr Ben Collins, Queens University Belfast
Dr Ke Yuan, University of Glasgow
Summary
Viruses remain a major threat to human and animal health, causing chronic, opportunistic and recurrent diseases. Their replication depends on intricate interactions between viral and host proteins that control how viruses enter cells, replicate, and evade our immune response. This PhD project will use deep-learning AI methods to predict these molecular interactions with focus on protein-protein interactions. The student will develop innovative models inspired by natural language processing to extend recent advances in protein language models, already applied successfully for protein structure prediction, to predict biomolecular interactions. Working with curated datasets of protein-protein interactions, genome and structural data, the project will focus on RNA viruses of importance to animal and human health. In an analogy with the use of language models in NLP, AI models can be used both for understanding the relationship among words (amino acids in a protein sequence) and sentences (protein structures). Predictions will be validated using proteomics using affinity purification and proximity labelling of viral proteins combined with mass spectrometry. Experimental validation will focus on the well-established influenza A virus model in a variety of host systems. The successful candidate will gain skills in computational biology, deep-learning, virology and proteomics at the forefront of AI-driven biology.