In the field of life sciences, single-cell sequencing technology serves as a "precision scalpel" that allows scientists to analyze the molecular characteristics of individual cells, providing unique insights into life's complexity. However, existing single-cell sequencing technologies often force scientists to make trade-offs: either missing certain cells or failing to capture cell morphology. This limitation results in a "blind men touching an elephant" scenario when dealing with complex research environments, where existing technologies cannot see the complete picture of cellular environments.

On August 22nd Beijing time, the National Key Laboratory of Genomic Multi-dimensional Analysis Technology, led by BGI Life Sciences Research Institute in collaboration with multiple institutions, published the cellomics technology Stereo-cell in Science. This technology achieves breakthroughs in multi-modal integration, in-situ dynamic capture, extreme sample compatibility, and million-level throughput, breaking through traditional single-cell sequencing limitations.

The achievement is expected to help single-cell sequencing move beyond the "planar analysis" mode that typically captures only single molecular-level information, advancing toward a "three-dimensional insight" capability that integrates multi-modal information. This will provide strong support for large-scale research in cellular pathology, development and aging, immunity and disease, and genetic evolution of animals and plants.

**Capturing and Identifying Millions of Cells in One Experiment**

Currently, one of the mainstream technologies in single-cell research is high-throughput single-cell sequencing based on droplet microfluidics. However, traditional single-cell sequencing technologies have numerous limitations, including low cell recovery rates, inability to directly observe cell morphology and shape, difficulty in analyzing large-sized cells with incomplete information capture, and significant cell data variations between experimental batches.

In this research, the Stereo-cell technology proposed by researchers is based on high-density DNA nanoball (DNB) array chips. Without relying on specific equipment for cell compartmentalization and encapsulation, it can directly perform in-situ cell capture and transcriptome sequencing, achieving unbiased capture from hundreds to millions of cells while simultaneously analyzing cellular transcription, protein signals, and morphological information.

"The chip consists of spheres with diameters of only 220 nanometers arranged densely at 500-nanometer intervals, acting like nanoscale 'catchers' that capture cells through electrostatic adsorption, avoiding cell loss or deformation caused by physical limitations in traditional methods," explained Liu Chang, co-first author and associate researcher at BGI Life Sciences Research Institute.

By combining microscopic imaging and spatial positioning, Stereo-cell can perform "CT scanning" and "GPS positioning" on every captured cell, achieving precise identification of each cell and avoiding misjudgments caused by overlapping cells or cluttered backgrounds.

Research shows that single-cell gene expression profiles generated by this technology are highly consistent with data obtained from traditional platforms, with cell type proportions more accurately reflecting real situations. In this study, researchers used Stereo-cell technology to achieve ultra-wide throughput ranges from hundreds to millions of cells on different chip sizes (0.5×0.5 cm and 6×6 cm), capable of identifying rare cell subpopulations from large-scale samples in a single experiment. Even rare cells comprising only 0.05% can be precisely found, achieving "finding a needle in a haystack." Additionally, Stereo-cell chips can reach sizes of up to 13×13 cm, providing possibilities for future larger cell throughput research.

**"Live Broadcasting" Cellular Dynamic Changes**

Whether it's possible to obtain multi-dimensional cellular information in one go - such as cell morphological characteristics, molecular types, and functional states - has long been a technical challenge in single-cell research. Researchers explain that Stereo-cell integrates fluorescent staining and antibody labeling technologies, simultaneously capturing cell morphology, transcripts, and cell surface proteins, identifying both cell types and their functions, equivalent to taking a "multi-modal three-dimensional photo" of cells.

"In the past, single-cell sequencing required many validation experiments, such as immunofluorescence staining or flow cytometry. Now we integrate these technologies together. One experiment can capture millions of cells and obtain morphological, transcriptional, and protein characteristics, enabling deeper analysis of cellular pathological states," said Liu Chuanyu, co-first author and researcher at BGI Life Sciences Research Institute. The Stereo-cell technology will significantly advance single-cell omics toward clinical cellomics, with potential for tremendous impact in disease mechanism research and clinical translation.

Furthermore, Stereo-cell supports direct cell culture on chips, enabling in-situ dynamic sequencing that captures changes in gene transcriptional activity while preserving spatial position and temporal change information. In research culturing fibroblasts on chips, this technology not only captured gene expression changes during cell migration and fibrosis processes but also analyzed the spatiotemporal distribution patterns of extracellular vesicles and interaction signals between cells in real physical contact. This opens new perspectives for understanding intercellular communication and greatly expands the boundaries of single-cell sequencing research, including applications in large-scale drug screening and small molecule perturbation experiments.

Researchers note that Stereo-cell can also analyze "extreme samples" that traditional technologies struggle to handle. For example, in skeletal muscle fiber research, Stereo-cell can precisely distinguish gene expression in different functional regions, revealing spatial heterogeneity of muscle fibers. In studies of large-volume egg cells, Stereo-cell can capture 719 oocytes in high throughput in situ, mapping gene expression changes, chromatin morphology dynamics, and RNA subcellular spatial distribution from growth to maturation.

**Supporting Breakthroughs in Life Science Fundamental Theory**

At a results presentation held on the 21st, based on the Stereo-cell technology platform, the "10 Billion Cell Coalition (10BC)" was officially established, aiming to map cellular atlases, construct virtual cells, deeply decode fundamental life laws, and promote life sciences to achieve industrial innovation from biological data reserves to intelligent technology-driven approaches.

"Currently in life sciences, we haven't formed fundamental regulatory summaries like those in physics, chemistry, and materials science - we're still in descriptive and observational stages," said Xu Xun, director of the National Key Laboratory of Genomic Multi-dimensional Analysis Technology and chief researcher at BGI Group. With the development of artificial intelligence (AI) technology, integrating effective single-cell data to construct virtual cell models for breakthrough discoveries in fundamental life science theories has gradually become a consensus in the scientific community.

In March this year, the Chan Zuckerberg Initiative (CZI) announced the launch of the "Billion Cell Project," aiming to generate billion-level single-cell data to accelerate AI biological model development. However, the technical foundation of this dataset still relies mainly on previous-generation single-cell technologies, lacking morphological and multi-tissue information.

Xu Xun explained that Stereo-cell is not only a technology platform but also a new-generation life data engine. The 10 Billion Cell Coalition launched based on this platform will construct "three major cellular universe databases" including life atlases, disease atlases, and perturbation response atlases. He called for: "We welcome global research teams to co-build and share, jointly promoting the development of cellular AI large models and virtual cell systems, achieving systematic leaps from data to diagnosis and treatment."

"At the current stage of life science development, we need to integrate more information and data. If single-cell analysis data can cover large cells and intercellular interactions, even down to subcellular structural levels, integrating these billion or trillion-level cellular multi-dimensional information and conducting systematic analysis with AI models will greatly deepen our understanding of disease mechanisms and promote scientific innovation in major fields, especially having profound significance for the prevention and control of complex and chronic diseases," said Cao Yanan, director of the Yangtze River Delta Health Research Institute at Shanghai Ruijin Hospital.

Related paper information: https://www.science.org/doi/10.1126/science.adr0475