Human Early Embryogenesis Spatiotemporal Transcriptomic Atlas (HESTA)

I. Database Overview

The Human Early Embryogenesis Spatiotemporal Transcriptomic Atlas (HESTA) provides a spatial and temporal atlas of gene expression in human embryos at early gestation, which is critical for understanding embryo development, organogenesis, and disease origins.

Using Stereo-seq, we obtained the spatiotemporal transcriptome from 77 sagittal sections of 13 human embryos at Carnegie stages 12–23 (CS12–CS23). We established the development trajectories and regulatory profiles of 50 organs, and identified the top organ-specific regulons with the highest scores as potential organ-identity regulators. The atlas refines the key organs and cell types that are vulnerable to virus infection and genetic disorders. We found dynamic changes of imprinting gene expression in specific organs at different stages. Our work, for the first time, revealed the time-lapse and spatial transcriptome dynamics of human embryogenesis after gastrulation.

The database consists of the following main sections:

Spatial Explorer: Spatial visualization of clustering and proof-of-concept 3D reconstruction

Organ Atlas: Organ-centric gene, pathway and regulon activity maps

snRNA: Matching single-nucleus RNA-seq visualization

Analysis: Gene overview, cross-species comparison, and allele visualization

Download: Data and metadata download

DCS Cloud: HESTA data are also available on DCS Cloud for customized online analysis (see Section VII)

Each section below includes a page screenshot and a short example walkthrough. Suggested first path for new users: Spatial clustering → Organs-gene (liver marker) → Gene overview.

II. Spatial Explorer – Spatial Visualization

The Spatial Explorer module provides interactive spatial visualizations of sections, clusters and a proof-of-concept 3D reconstruction.

1. Embryo Overview

   Embryo Overview is the visual gateway to the spatial data. Users can browse samples and sections, view where each section originates in the intact embryo, and inspect summary charts such as cell counts per section and tissue composition for the selected section.

   

   Example: orient yourself before diving into a section

   1. Open Spatial Explorer → Embryo Overview.
   2. Pick a sample (for example CS18 or CS23) to see section placement in the intact embryo.
   3. Review cell counts and tissue composition for candidate sections, then open Spatial clustering with a promising section ID.

2. Spatial Clustering

   Visualizes organ-level spatial clustering annotations for each section.

   Sections are selected via a developmental-stage timeline (CS12–13 to CS23), followed by section chips within the chosen stage.

   Supports interactive operations such as zooming, panning and selecting clusters.

   

   Example: locate an organ cluster on a Stereo-seq section

   Goal: find where liver (or another organ) is annotated at a given developmental stage.

   1. Open Spatial Explorer → Spatial clustering.
   2. Click a stage on the timeline (for example CS18 or CS23).
   3. Click a section chip under that stage to load the clustering map.
   4. In the viewer, zoom/pan and select the organ cluster of interest (for example Liver) to highlight its spatial territory.

3. Substructure Clustering

   Visualizes substructure-level clusters within available sections.

   As with Spatial Clustering, users first pick a developmental stage on the timeline, then choose a section chip to load the corresponding visualization.

   

   Example: inspect finer domains within a section

   1. Open Spatial Explorer → Substructure clustering.
   2. Select a stage on the timeline, then a section chip.
   3. Explore substructure clusters in the viewer to resolve domains within major organs.

4. Proof-of-concept 3D reconstruction

   One proof-of-concept reconstruction generated from eight consecutive Stereo-seq sections of the CS12–13 E2 embryo.

   Source section IDs: CS12-13_E2S1, CS12-13_E2S2, CS12-13_E2S3, CS12-13_E2S4, CS12-13_E2S5, CS12-13_E2S6, CS12-13_E2S7, CS12-13_E2S8.

   Supports rotation, zooming, and sectioning in the 3D view.

   

   Example: open the 3D reconstruction

   1. Open Spatial Explorer → Proof-of-concept 3D reconstruction.
   2. Rotate, zoom, and section the volume built from CS12-13_E2S1–S8 to relate 3D structure back to 2D Stereo-seq slices.

III. Organ Atlas – Organ-centric Multi-omics

The Organ Atlas module provides organ-centric views of gene expression, pathway activity and regulon activity across organs and developmental stages.

1. Organs – Gene

   Organ-specific spatial gene expression maps.

   Each dataset is represented as an m bin50 × n genes matrix, where each value is the expression level of a gene in a bin50 region.

   Users can select organs and genes to visualize spatial expression patterns and compare across stages.

   

   Example: find a gene in the developing liver

   Goal: inspect spatial expression of a liver marker (for example AFP or ALB) within liver tissue.

   1. Open Organ Atlas → Organs-gene.
   2. In the File name dropdown, select Liver.gene.h5ad to load the liver gene dataset.
   3. In the embedded explorer, search or select your gene of interest and view its spatial pattern on liver bin50 spots.
   4. Optional: open Analysis → Gene overview with the same gene to compare expression across stages.

2. Organs – Regulon

   Organ-centric spatial regulatory network visualization.

   To infer spatial regulatory networks for each organ substructure:

   The merged raw count matrix was first subsampled to optimize computation.

   Regulatory networks were inferred using Stereopy with method=KNN.

   Regulon activity was quantified by AUCell enrichment scores, yielding an m bin50 × n regulons matrix describing per-bin50 regulon activity.

   The AUCell matrices were converted to h5ad files for interactive visualization.

   

   Example: explore organ regulons

   1. Open Organ Atlas → Organs-regulon.
   2. Select an organ file (for example Liver.regulon.h5ad).
   3. Choose a regulon of interest and examine its spatial activity on bin50 spots.

3. Organs – Pathway

   Organ-centric spatial pathway activity maps.

   To quantify pathway activity:

   The R package irGSEA was used to perform rank-based gene set enrichment analysis on the merged raw count matrix.

   The UCell algorithm was applied to compute per-bin50 enrichment scores against Gene Ontology (GO) Biological Process gene sets.

   The resulting m bin50 × n pathways matrix of enrichment scores was compiled into h5ad files for interactive visualization.

   

   Example: explore organ pathways

   1. Open Organ Atlas → Organs-pathway.
   2. Select an organ file (for example Liver.pathway.h5ad).
   3. Choose a pathway of interest and examine its spatial enrichment on bin50 spots.

IV. snRNA – Single-nucleus RNA-seq Visualization

The snRNA module provides visualization of matching single-nucleus RNA-seq datasets.

Supports UMAP/t-SNE plots and cell-type annotation.

Users can explore gene expression across clusters and compare with spatial maps in Organ Atlas.

V. Analysis – Cross-dataset Analysis Tools

The Analysis module provides several tools for cross-dataset and cross-species analysis.

1. Gene Overview

   Search a gene and browse seven developmental stages in one row. Under each stage, users can switch section and view section spatial maps, gene expression maps, and gene violin plots across tissues.

   A stage bar supports jumping between stages and custom stage order (drag or move buttons), with Restore original order to reset the default CS12–13 → CS23 sequence.

   

   Example: browse a gene across all developmental stages

   Goal: follow one gene from CS12–13 through CS23 in a single view.

   1. Open Analysis → Gene overview.
   2. In the Gene search box, type a symbol (for example PAX6 or AFP), select it from the list, then click Confirm.
   3. Scroll across the seven stage panels. Each stage shows (1) section spatial, (2) gene expression, and (3) gene violin by tissue; change the section dropdown under a stage to switch slices.
   4. Use the stage bar chips to jump between stages, or drag chips to reorder. Click Restore original order to reset CS12–13 → CS23.

2. Cross-Species Comparison

   Compare human (HESTA) and mouse (MOSTA) spatial gene expression along a shared developmental-stage timeline. Users select stage-matched files, orthologous gene pairs, and metadata fields, then confirm to refresh spatial visualization, spatial gene expression, and expression distribution for both species.

   

   Example: compare human and mouse spatial expression

   Goal: view a human–mouse ortholog pair at matched developmental stages (HESTA vs MOSTA).

   1. Open Analysis → Cross-Species Comparison.
   2. Select a developmental stage on the shared human↔mouse timeline.
   3. Choose a Human file (HESTA) and a Mouse file (MOSTA), then search and pick a gene pair (for example PAX6 / Pax6).
   4. Optionally set Human / Mouse cell metadata fields, then click Confirm to refresh the three visualization blocks for both species.

3. Allele

   Visualizes the spatial expression distribution of genes with significant allelic imbalance.

   Users can select different sections and use the dropdown list to select genes with significant allelic imbalance expression, displaying their spatial expression distribution on the section.

   

   Example: view allelic imbalance on a section

   1. Open Analysis → Allele.
   2. Select a section, then choose a gene with significant allelic imbalance from the gene list.
   3. Inspect the spatial expression distribution of the selected allele-imbalanced gene on that section.

VI. Download – Data and Metadata Access

The Download page lists HESTA data files (spatial clustering, substructure clustering, snRNA, organs-gene / regulon / pathway, and related resources). Use the type filter and search box to find files, then click the download icon for a single file. You can also export the current or full table for offline reference.

HESTA data are hosted on CNGBdb FTP: https://ftp.cngb.org/pub/SciRAID/stomics/STDS0000394/.

For faster and more stable large-scale data transfer, we strongly recommend using RaySync with the following steps:

1. Download RaySync

Get the tool from: https://ftp.cngb.org/pub/course/tool/raysync/

2. Use RaySync to Download Data

To access and download specific files or directories in the HESTA resource directory (https://ftp.cngb.org/pub/SciRAID/stomics/STDS0000394/), open the FTP page in a browser and use the RaySync icon to copy and execute the corresponding command. Start the RaySync Client first (Windows / Mac), then refresh the FTP page if needed.

3. Batch Download

To download the entire HESTA dataset or directory, click “ Wget” or “ RaySync” to copy the command, then run it in your terminal.

Wget (resume-friendly recursive download)

RaySync command-line (rayfile-c)

For additional details and guidance, please refer to the Readme: Readme.txt.

If you encounter any issues during the download process, please contact us at CNGBdb@genomics.cn.

VII. DCS Cloud – Customized Online Analysis

In addition to browsing HESTA on this website and downloading data from CNGBdb FTP, we have uploaded the HESTA dataset to DCS Cloud (https://www.dcs.cloud/), where users can run customized analyses on the platform.

DCS Cloud (DCS智能云平台) is a spatiotemporal-focused multi-omics intelligent analysis platform for managing and analyzing multi-omics data, and for sharing and publishing analysis results and script tools. Built around projects, DCS Cloud helps users quickly integrate data and tools into a project so that analysis is traceable, results are reproducible, knowledge is shareable, and projects are collaborative—closing the loop of the research workflow and accelerating discovery.

Through a friendly online portal, DCS Cloud provides flexible no-code workflow analysis and interactive analysis, high-resolution visualization tools, and personalized code-based analysis services. Any registered and signed-in user can easily access, analyze, use, and share data and tools on the platform.