On August 21, the State Key Laboratory of Genome and Multi-omics Technologies, led by BGI-Research, collaborated with multiple institutions to unveil a groundbreaking single-cell omics technology named Stereo-cell in Science. This technology represents a transformative advancement in cellular analysis, promising to revolutionize precision medicine, rare disease research, and world’s fundamental understanding of biological systems.

Based on this technology platform, the Laboratory collaborates with 18 institutions to launch the “10 Billion Cells Alliance” (10BC). The alliance, which brings together global scientists to advance the decoding of the fundamental principles of life, will play a critical role in fostering open global collaboration and accelerating the translation of cutting-edge research into tangible public health benefits.

Wang Jian, Chairman and Co-founder of BGI Group, stated, "New technological breakthroughs will lead to new scientific discoveries. We hope to use this technology to serve more people, not only in China but also to make a greater impact internationally."

The study “Stereo-cell: Spatial enhanced-resolution single-cell sequencing with high-density DNA nanoball-patterned arrays” was published in Science.

Over the past decade, single-cell sequencing technologies have dramatically advanced global understanding of cellular heterogeneity and biological complexity, enabling analysis of genomes, epigenomes, and transcriptomes at single-cell resolution. However, existing single-cell approaches still face significant challenges, including throughput limitations, capture uniformity issues, cell size compatibility constraints, and technical scalability barriers.

Published in Science, Stereo-cell represents a breakthrough that overcomes these long-standing limitations. Stereo-cell uses a chip densely tiled with DNA nanoballs (DNBs) as "landing pads" for individual cells, serving as an array of nanoscale capture elements. These elements, spaced at high density across the chip surface, enable precise spatial positioning of individual cells.

Stereo-cell on high-density DNB arrays enables spatially resolved, multimodal single-cell profiling across scales; this integrative design supports rare-cell discovery, microenvironment analysis, and large-structure studies.

Cells are deposited onto a poly-L-lysine–coated surface that enhances electrostatic interactions to enable cell attachment, where RNA is captured by embedded oligo-dT probes, imaging identifies cell positions, and sequencing reads transcripts. This direct capture approach eliminates the need for droplet-based encapsulation while enabling precise spatial positioning of individual cells and maintaining their morphological integrity. Furthermore, optional workflows enable multiplex immunofluorescence and protein profiling, while on-chip culture allows time-resolved measurements of cellular dynamics.

Up: Droplet-free, imaging-guided in situ capture and deep-learning segmentation on a poly-L-lysine DNB array power accurate single-cell calling; this lowers doublets and strengthens multimodal readouts.
Down: Scalable chips and spatial UMI maps with thousands of segmentations demonstrate high-throughput, unbiased capture; this scale enables detection of rare populations and robust atlas construction.

The study demonstrates strong performance across multiple datasets and cell types. On a 6 by 6 cm chip, the team captured 445,467 peripheral blood cells in a single experiment and detected rare hematopoietic stem and progenitor cells (HSPCs) at approximately 0.05% of the population, achieving what researchers describe as "finding needles in a haystack" at unprecedented scale.

This capability could prove crucial for early disease detection, as many diseases begin with changes in rare cell populations that traditional methods might miss. Imaging-guided filtering reduced doublets from 4.38% to 1.29% in mixed human-mouse cell tests, demonstrating improved accuracy over traditional droplet-based methods.

In benchmarks using human peripheral blood mononuclear cells (PBMCs), Stereo-cell achieved cell-type proportions closer to flow-cytometry measurements than public datasets from mainstream droplet platforms, with comparable gene detection metrics. This means the technology provides more accurate representations of what actually exists in patients' blood samples.

For large cells, the oocyte dataset included 719 cells with an average of 8,972 genes per cell, enabling high-throughput studies of fertility and reproductive health that were previously limited to analyzing just a few cells at a time.

"In a single experiment, millions of cells can be captured, along with their morphological, transcriptional, and protein characteristics, enabling deeper analysis of cellular pathological states," said Liu Chuanyu, co-first author of the paper and researcher at the BGI-Research. "Undoubtedly, Stereo-cell is a milestone in the progression from single-cell omics to clinical cell omics, with the potential to play a significant role in disease mechanism research and clinical translation."

Left: Dual omics capture strategy of Stereo-cell-CITE and study design.
Right: Spatial visualization of the distribution of captured RNA and protein on an S1 chip with an input of 10,000 human PBMCs.

The authors describe multimodal advantages that provide insights beyond RNA analysis alone, creating what researchers call "multidimensional cellular profiles." Using multiplex immunofluorescence and Stereo-cell-CITE workflows, protein markers including CD3, CD45RA, CD112-positive T cells, and CD103-positive tissue-resident signatures matched transcriptomic clusters in PBMCs, while stimulation experiments revealed regulatory networks in natural killer cells. This work advances the world’s understanding of how immune cells coordinate responses to threats.

With cells remaining in place on the chip, the team cultured fibroblasts directly on arrays and recorded time-resolved changes, capturing elevated migration and fibrosis pathway activity, providing new insights into wound healing and tissue scarring processes. In multinucleated skeletal muscle fibers, Stereo-cell defined spatial regions that localized gene modules at key junctions and distinguished fiber-type markers, potentially informing treatments for muscle diseases and age-related muscle loss.

Stereo-cell enables in situ sequencing for cultured cells.

The technology can also provide insights into cell–cell interactions, microenvironments and subcellular localization under the study's experimental conditions. In oocytes, the authors mapped maturation trajectories and identified RNA localization patterns consistent with single-molecule RNA FISH, tracking the correlation between different subcellular gene modules’ region-specific distributions within large cells. This knowledge could advance fertility treatments and reproductive medicine.

Subcellular gene modules in oocytes (e.g., OOEP aggregation, SLC45A3 dispersion) reveal organized RNA landscapes; this resolution enables high-throughput mapping of intracellular regulation in large cells.

Experts have positioned Stereo-cell as a groundbreaking advancement in single-cell analysis, transitioning from traditional "flat analysis" to comprehensive "three-dimensional insights" that support large-scale studies in cellular pathology, development, immune research, and genetics. They emphasized four key advances: multimodal integration of spatial, RNA, and protein signals; in situ, time-resolved readouts; compatibility with extreme sample types; and scalability from hundreds to nearly a million cells per chip. This breakthrough paves the way for transformative applications, including the digitization of clinical pathology and large-scale drug screening.

Professor Wang Xiangdong, Chief Officer of Scientists at Zhongshan Hospital affiliated to Fudan University, and Director of Shanghai Institute of Clinical Bioinformatics and Fudan University Center of Clinical Bioinformatics, stated, "From a clinical perspective, Stereo-cell technology has pioneered a new pathway in clinical molecular medicine, which will help us provide better services for patients." Currently, Professor Wang is collaborating with experts from six hospitals, including Zhongshan Hospital of Fudan University, Shanghai Tongji Hospital, and Henan Provincial People's Hospital, to form the Stereo-cell clinical team. They are conducting projects on the clinical translation of single-cell technology based on cutting-edge Stereo-cell technology, aiming to provide patients with multidimensional and multi-faceted clinical diagnostics and treatments.

Professor Ruan Yijun from the Life Sciences Institute at Zhejiang University stated: "Stereo-cell is a groundbreaking technology. It has completely expanded our imagination, enabling us to explore the functions of every human cell during the life process, the changes that occur, and the conditions under which diseases begin to develop. In the future, it will have unlimited application prospects in the field of clinical medicine. I look forward to everyone working together to advance single-cell technology from the billion-level to the trillion-level, allowing us to truly and comprehensively decode the fate of every cell in the human body."

 "Stereo-cell is not just a technology platform, but a new generation of biological data engine," said Xu Xun, co-corresponding author of the paper, director of the State Key Laboratory of Genome and Multi-omics Technologies, and chief scientist at BGI Group. "Based on this platform, the 10 Billion Cells Alliance (10BC) was launched to construct the 'three major cellular universe databases':  life atlases, disease atlases, and perturbation-response atlases. We welcome global research teams to collaborate and share, jointly promoting the development of cell-scale AI foundation models and virtual cell systems, and achieving a systematic leap from data to diagnosis and treatment."

This study can be accessed here: https://www.science.org/doi/10.1126/science.adr0475