Journal: Science Advances

Article Title: BrainSTEM: A single-cell multiresolution fetal brain atlas reveals transcriptomic fidelity of human midbrain cultures

doi: 10.1126/sciadv.adu7944

Figure Lengend Snippet: ( A ) Schematic depicting the computational workflow for constructing the fetal whole-brain single-cell atlas and fetal midbrain subatlas, followed by their downstream applications. This includes projecting in vitro–derived datasets onto these reference atlases to evaluate their transcriptional fidelity. Created in BioRender. Toh, H. (2025); https://BioRender.com/pl47oha . ( B ) Dotplot of gene expression of known marker genes across different annotated cell types in the fetal whole-brain atlas. Uniform Manifold Approximation and Projection (UMAP) plot of fetal whole-brain atlas showing ( C ) the annotated cell types, ( D ) the single-cell study, ( E ) the embryonic age in postconception week, and ( F ) the gene expression of known marker genes for selected cell types. ( G ) Schematic depicting the determination of region specificity by comparing the region identity of each cell’s k -nearest neighbors. Created in BioRender. Toh, H. (2025); https://BioRender.com/pl47oha . ( H ) Heatmap showing the average proportion of k -nearest neighbors mapping to the corresponding brain region (midbrain, forebrain, or hindbrain) across different cell types. UMAP plot of dopaminergic (DA) neurons showing ( I ) the brain region identity, the midbrain region gene expression signature, the forebrain region gene expression signature, and the hindbrain region gene expression signature. ( J ) Heatmap of gene expression of top 10 marker genes, ordered by log2 fold change, for each brain region in DA neurons.

Article Snippet: Library preparation was conducted using the Parse Biosciences Evercode Whole Transcriptome (WT) kit (v2.0.1).

Techniques: In Vitro, Derivative Assay, Gene Expression, Marker

Journal: Science Advances

Article Title: BrainSTEM: A single-cell multiresolution fetal brain atlas reveals transcriptomic fidelity of human midbrain cultures

doi: 10.1126/sciadv.adu7944

Figure Lengend Snippet: ( A ) Schematic depicting the BrainSTEM framework’s two-tier mapping process of query datasets onto our fetal whole-brain atlas and midbrain subatlas. Created in BioRender. Toh, H. (2025); https://BioRender.com/pl47oha . Proportion plots of in vivo brain tissue single-cell studies across different developmental stages showing ( B ) proportion of cells confidently assigned, ( C ) region.lineage proportions for confidently assigned cells, and ( D ) cell-type proportions for confidently assigned cells. ( E ) Cell-type proportions of synthetic dataset consisting of 5:1 hindbrain-to-forebrain cells projected either by BrainSTEM or by DM. Proportion plots of in vitro–derived single-cell datasets across different time points and conditions showing ( F ) proportion of cells confidently assigned, ( G ) region.lineage proportions for confidently assigned cells, ( H ) cell-type proportions for confidently assigned cells, and ( I ) brain region proportions for confidently assigned DA neurons. ( J ) Ridge plots showing the distribution of dopaminergic neuron gene expression signature (DA N score) and line graphs showing the proportion of DA N in the midbrain-associated cells across different time points for selected in vitro midbrain differentiation time series studies. Spearman’s correlation coefficient for the DA N score and the proportion of DA N in each study presented. Two-sided Wilcoxon rank sum test applied to assess the pairwise difference between the in-house dataset and the other three datasets, with significance indicated by asterisk [ P value for the in-house dataset versus Fiorenzano et al. : 0.000, the in-house dataset versus Xu et al. : 0.078, and the in-house dataset versus You et al. : 0.779]. ( K ) Ridge plots showing the distribution of serotonergic neuron gene expression signature (Ser N score) in the hindbrain-associated cells in the same setting.

Article Snippet: Library preparation was conducted using the Parse Biosciences Evercode Whole Transcriptome (WT) kit (v2.0.1).

Techniques: In Vivo, In Vitro, Derivative Assay, Gene Expression