Duke-NUS Medical School researchers, working with the University of Sydney, have developed BrainSTEM—a two-tier, single-cell atlas of the developing human brain that profiles nearly 680,000 cells. Published online in Science Advances on October 31, 2025, the resource focuses on midbrain dopaminergic neurons, flags off‑target cell types in lab-grown models, and will be released openly for the research community.
Scientists at Duke-NUS Medical School and collaborators have created BrainSTEM (Brain Single-cell Two tiEr Mapping), a comprehensive single-cell atlas of the fetal human brain designed to benchmark and improve models for Parkinson’s disease.
- Scope and focus: The atlas profiles nearly 680,000 cells from the developing human brain and includes a higher-resolution midbrain sub-atlas that pinpoints dopaminergic neurons—cells impaired in Parkinson’s disease.
- Why it matters: Parkinson’s disease is Singapore’s second most common neurodegenerative condition, affecting about three in every 1,000 people aged 50 and above. More faithful midbrain models could strengthen research and support future cell-based therapies.
The study, published in Science Advances (Vol. 11, Issue 44; DOI: 10.1126/sciadv.adu7944) on October 31, 2025, reports that leading laboratory differentiation protocols can yield unwanted cells from other brain regions, underscoring the need to refine both experimental methods and data-analysis pipelines.
Lead authors and senior investigators highlighted the tool’s practical value. “Our data-driven blueprint helps scientists produce high-yield midbrain dopaminergic neurons that faithfully reflect human biology. Grafts of this quality are pivotal to increasing cell therapy efficacy and minimizing side effects, paving the way to offer alternative therapies to people living with Parkinson’s disease,” said Dr. Hilary Toh, an MD–PhD candidate at Duke‑NUS.
“By mapping the brain at single-cell resolution, BrainSTEM gives us the precision to distinguish even subtle off-target cell populations. This rich cellular detail provides a critical foundation for AI-driven models that will transform how we group patients and design targeted therapies for neurodegenerative diseases,” said Dr. John Ouyang, a senior author from Duke‑NUS’ Centre for Computational Biology. Assistant Professor Alfred Sun added that the rigorous, data-driven approach “will speed the development of reliable cell therapies for Parkinson’s disease,” while Professor Patrick Tan, Senior Vice‑Dean for Research at Duke‑NUS, called the work a new benchmark for capturing cellular detail in complex systems.
The team will release the atlases as open resources alongside an out‑of‑the‑box BrainSTEM package, enabling researchers to apply the multi‑tier mapping approach to any brain cell type. The project involved collaborators at the University of Sydney and received support including the USyd–NUS Ignition Grant and the Duke‑NUS Parkinson’s Research Fund.
