Preclinical human neural organoid and computational/AI model development
Not stated
- Funding
- Competition Funded PhD Project (Students Worldwide)
- Application deadline
- 8 December 2026
About the project
About the Project Summary Human neural organoids grown from patient-specific iPSCs have emerged as a promising preclinical model that can reproduce important details of human CNS development, tissue architecture and pathology. However, we need precision tools to explore which aspects of complex physiological and pathophysiological cellular behaviour these models can recapitulate, increasing confidence in their value as discovery and drug-testing systems. Recent advances in large-scale artificial intelligence (AI) models, including single-cell-based foundation models and AI-based perturbation models, offer new opportunities to study the dynamics of disease mechanisms at single-cell resolution. In this project, we will combine single-cell profiles from human neural organoids and postmortem CNS samples to develop and apply large-scale AI frameworks to identify disease-associated cellular states, molecular biomarkers, and potential therapeutic targets. Computational predictions will subsequently guide experimental validation in neural organoid models. Parts of the project is in collaboration with various groups in the new Cambridge MRC Translational Models Hub (Lotfollahi, Teichmann, Lancaster) and may involve collaborators from industry (Replicam). Project Aims We will develop and apply large-scale AI frameworks, based on single-cell and spatial transcriptomic analyses combined with in silico perturbation models, to existing organoid and postmortem tissue datasets to explore healthy tissue and disease-specific representations, with a focus on neurodegenerative conditions. We will scale up neural organoid models and explore the extent to which they consistently recapitulate expected physiological and pathophysiological features, based on the benchmarks. We will experimentally validate selected AI-predicted therapeutic targets and biomarkers, using gene editing, delivery or Trim-away-based perturbations and drug-testing in human neural organoid models.