27 September 2024 in Science
The molecular and cellular organization of the primate cerebellum remains poorly characterized. We obtained single-cell spatial transcriptomic atlases of macaque, marmoset, and mouse cerebella and identified primate-specific cell subtypes, including Purkinje cells and molecular-layer interneurons, that show different expression of the glutamate ionotropic receptor Delta type subunit 2 ( GRID2 ) gene. Distinct gene expression profiles were found in anterior, posterior, and vestibular regions in all species, whereas region-selective gene expression was predominantly observed in the granular layer of primates and in the Purkinje layer of mice. Gene expression gradients in the cerebellar cortex matched well with functional connectivity gradients revealed with awake functional magnetic resonance imaging, with more lobule-specific differences between primates and mice than between two primate species. These comprehensive atlases and comparative analyses provide the basis for understanding cerebellar evolution and function.
21 August 2025 in Science
Single-cell sequencing technologies have advanced our understanding of cellular heterogeneity and biological complexity. However, existing methods face limitations in throughput, capture uniformity, cell size flexibility, and technical extensibility. We present Stereo-cell, a spatial enhanced-resolution single-cell sequencing platform based on high-density DNA nanoball (DNB)–patterned arrays, which enables scalable and unbiased cell capture at a wide input range and supports high-fidelity transcriptome profiling. Stereo-cell further allows integration with imaging-based modalities and multiomics strategies, including immunofluorescence and epitope profiling. This platform is also compatible with profiling extracellular vesicles, microstructures, and large cells, whereas its spatial resolution facilitates in situ analysis of cell-cell interactions, cellular microenvironments, and subcellular transcript localization. Together, Stereo-cell provides a flexible framework for expanding single-cell research applications.
12 September 2026 in SSRN Electronic Journal
Transcriptional programmes are interpreted through the tissue neighbourhoods in which they occur, yet most transcriptomic foundation models learn from dissociated proffles or add spatial information after representation learning. Here we introduce OmniCell, a tissue-contextual foundation model pretrained on 67 million dissociated and spatially resolved proffles. OmniCell encodes each cell or spot as a sequence of gene tokens that combine gene identity, expression magnitude and observed tissue coordinates, allowing spatially aware attention and a symmetric cell–gene objective to learn coupled cell and contextual gene representations. Across cellular, molecular and tissue scales, OmniCell preserved transferable cell-state structure across species, batches and rare populations, recovered marker genes and tissuealigned gene programmes, and improved spatial identity mapping, deconvolution and anatomical domain reconstruction. In human liver cancer Stereo-seq data, OmniCell resolved a tumour-margin transition zone and showed that gene-neighbourhood relationships shift between tumour core, transition zone and adjacent non-malignant tissue, consistent with niche-dependent gene function. In silico spatial perturbations further nominated region-speciffc regulatory hypotheses in developing mouse brain and marmoset cortex. OmniCell therefore provides a framework for modelling gene function as a property of cells embedded in intact tissue, rather than as a global expression pattern alone.