Irina Tikhonova
Molecular Modelling Reader Queen's University Belfast
Dr. Irina Tikhonova is a Reader in Molecular Modelling at Queen’s University Belfast and an expert in computational chemistry and computer-aided drug discovery for G protein-coupled receptors (GPCRs). Following postdoctoral experience at the National Institutes of Health (Bethesda), University of California San Diego, and Institute of Molecular Medicine (Toulouse), she has established herself as a pioneer in developing innovative computational approaches to identify allosteric binding sites and elucidate GPCR pharmacology. Her research focuses on developing integrated computational and experimental platforms for targeted discovery of biased allosteric GPCR modulators as more precise medicines.
Seminars
- Harnessing cryo-EM and X-ray crystallography to reveal receptor structures and guide simulation-ready models
- Animating GPCRs through molecular dynamics in lipid bilayers to decode ligand binding, activation, and biased signalling
- Accelerating drug discovery via structure-based virtual screening to identify novel therapeutics and mutation-specific strategies
Despite GPCRs accounting for a significant proportion of all approved drugs, it is estimated that 60–85% of potentially therapeutic GPCRs remain untargeted, leaving substantial untapped opportunity within this receptor class. Historically, limited structural data has posed a major challenge to advancing GPCR-targeted drug discovery. However, recent breakthroughs in AI-driven modelling and machine learning are beginning to transform this landscape. From predicting receptor structures with greater precision to enhancing target identification and streamlining drug design, these technologies are rapidly accelerating discovery and enabling access to previously unreachable targets. This workshop will explore how the latest advancements in AI and ML are driving innovation in GPCR drug discovery, bridging the gap from target identification to clinical translation.
Key Takeaways:
- Utilise computer-based drug design to overcome structural gaps by supporting structure prediction, binding-site identification and functional conformation modelling, particularly for orphan and lipid-sensing receptors
- Enhance virtual screening and hit prioritisation by leveraging AI for molecular docking, chemical space exploration, and machine learning-driven scoring, ranking, and selection of tractable compounds
- Predict functional states and downstream signalling using AI models to gain deeper insights into receptor behaviour and distinguish between agonist and antagonist responses
