China Maps Rice’s Full-Lifecycle Cell Atlas in Cell

A large team of Chinese researchers has constructed the first three-dimensional spatiotemporal cell atlas spanning the entire life cycle of rice, from germinating seed to flowering and grain fill. The work, published online in the journal Cell on the evening of October 6, 2026, integrates genome sequencing, single-cell sequencing and spatial transcriptomics to create a single-cell resolution map of how a rice plant is built — a resource its authors say will support more precise crop improvement.

The study was led by Yazhouwan National Laboratory, working with BGI Life Science Research Institute (also known as BGI Research), Huazhong Agricultural University, Southern University of Science and Technology (SUSTech), Wuhan University, and the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences — a coordinated effort involving a total of 12 research institutions.

What the Atlas Contains

The team used the japonica rice variety “Zhonghua 11” as its material, combining BGI’s self-developed spatial omics technology Stereo-seq, single-cell sequencing and high-throughput gene sequencing. According to Xinhua, the map covers 10 organ or tissue types and 61 developmental stages, drawing on more than 850,000 high-quality cell nuclei and over 340,000 spatial data bins, and annotating 119 cell types and 133 cell subtypes.

An English-language summary from Pandaily gives the precise figures: the datasets profile about 851,725 nuclei and 347,640 spatial bins. An inferred developmental trajectory links cell states from the early embryo to mature organs, allowing researchers to watch how different cells coordinate to “assemble” a complete rice plant.

Put simply, as Xinhua described it, a single grain of rice works internally like a vast and precise super-factory, in which tens of thousands of cells — some responsible for growing roots, others for forming grains — must divide and cooperate at the right time and place. Until now, researchers could not clearly see this internal operation. The new atlas breaks through that blind spot by tying the molecular features of each cell to its physical position and developmental stage.

The Endosperm Finding: Taste and Nutrition

The research connects directly to the dinner table. Rice grain is largely made up of the endosperm, which determines the taste and nutritional quality of cooked rice. The study found that the seemingly uniform endosperm has a fine internal “spatial division of labor” — a dorsoventral split. Near the dorsal region, genes related to carbohydrate metabolism and starch synthesis are enriched; near the ventral region, genes tied to storage-protein synthesis are favored. Mutations in the related genes can shift starch and protein accumulation patterns, the researchers reported.

The team also highlighted the auxin-responsive regulator OsARF1, which is expressed at different spatial and developmental stages and participates in different molecular regulatory networks, including endosperm nutrient allocation. Because OsARF1 ties into multiple developmental branches with largely distinct predicted targets, it illustrates why editing a single “master” gene can ripple widely through an organism — a long-standing problem in breeding known as pleiotropy, where a gene that improves one trait disrupts others.

The practical implication, Xinhua reported, is that scientists seeking to breed better-tasting or more nutritious rice can now use the atlas to devise more targeted, spatially specific regulation strategies rather than crude edits to master regulators.

New Tools for Global Researchers

Alongside the atlas, the team developed two reusable resources. The first is an interactive online atlas platform, the Rice Spatio Temporal Atlas (RSTA), supporting gene queries, spatial expression display and comparative analysis. The second is RICE scGPT — a rice-adapted single-cell foundation model trained on this study’s data and public rice single-cell datasets, designed to help annotate cell types and integrate datasets from different sources.

The associated RSTA GitHub repository curates and openly shares the project’s resources, including Stereo-seq analysis code, the Rice-scGPT model, cell lineage and seed-development tools. The repository describes RSTA as a multiscale, spatiotemporal integrated research initiative “from seed to seed” and states that its resources are “freely accessible” through a user-friendly web interface. It lists 40,155 annotated rice genes, 788,030 high-quality single nuclei profiled across 47 tissues, spatial gene expression mapped in 43 tissue samples, and marker genes defining 202 cell types — broader cumulative totals than the single paper’s figures.

As of the most recent repository updates, the public web interface was still “waiting for public release,” and no exact release date has been confirmed.

‘Organized Science’ and a 26-Year Arc

The work is being framed in China as a model of what officials call “organized scientific research.” State broadcaster CCTV, reported by Yicai, described the atlas as the first of its kind globally — a three-dimensional spatiotemporal cell atlas covering rice’s entire journey from seed germination to flowering and fruiting, integrating genome, single-cell and spatial transcriptome sequencing. China News Service, republishing the Xinhua report, noted the atlas offers new research resources for understanding rice growth and exploring grain quality improvement.

Commentators have placed the achievement within a longer historical arc. Behind the result lies a research relay spanning 26 years: as early as 2000, the agronomist Yuan Longping signed a cooperation agreement with BGI and others that launched BGI’s rice genome research. The rice indica genome working draft was published in the journal Science in 2002. The new atlas, its backers argue, converts that linear genome sequence into a space-and-time-aware instruction manual for the whole organism.

The study’s corresponding institution is SUSTech, with chair professor Hongwei Guo among the corresponding authors, according to Pandaily.

Why It Matters

Rice is a primary staple food in China and much of the world, making its genetic and developmental understanding economically and food-security critical. Improving taste and nutrition via endosperm composition is a major breeding goal. By mapping where functional genes act across tissues and stages, the atlas helps breeders navigate the pleiotropy problem — targeting context-specific downstream modules instead of bluntly altering master regulators.

The scientific significance lies less in any single finding than in the framework itself: the work turns the rice genome from a linear sequence into a spatial and temporal map, and packages it as open infrastructure — an atlas portal and a foundation model — that international crop researchers can reuse.

What to Watch

Several open items remain. The public release date of the RSTA web interface has not been confirmed. The full list of corresponding authors has not been fully enumerated in the available sources. And the slight differences between the paper’s dataset totals and the repository’s cumulative figures will need reconciliation as the resource continues to expand. If the atlas delivers on its promise as a reusable framework, the more consequential test — measurable improvements in rice taste and nutrition through targeted breeding — will play out over the coming years, not weeks.