Human Gut Cell Atlas (HGCA) v1.0

Highlights

• First official Human Cell Atlas reference for the intestine: 944,390 healthy and disease-adjacent cells from 502 samples and 265 donors across 24 studies, spanning duodenum to colorectum.
• A five-level community-defined cell type taxonomy, curated through annotation jamborees and CAP, resolves 94 homeostatic epithelial, stromal, and immune identities, including rare and transitional populations.
• Intestinal segment and radial sampling depth systematically shape composition and transcriptional state; independent Visium and Xenium data recover the inferred niches, including lymphoid follicles.
• The atlas is a transferable reference - mapping 987,466 TAURUS IBD cells identified 42 additional standardized healthy subtypes and states, and recovered inflammation and disease remission signatures as well as de novo annotation. Mapping HEOCA intestinal organoids expanded 16 epithelial labels to 28, and separated adult-state maturation from regional specification.

Overview

The Human Gut Cell Atlas (HGCA) v1.0 is the first official Human Cell Atlas reference for the human intestine. It comprises 944,390 healthy and disease-adjacent intestinal cells from 502 samples and 265 donors across 24 landmark studies, spanning duodenum to colorectum and covering both endoscopic biopsies and surgical resections. Harmonized anatomical, clinical, and technical metadata were collected with MetaManager, and the HCA Atlas Tracker.

A cornerstone of this effort is community-driven cell annotation: input from clinicians, stem cell biologists, immunologists, and computational experts across the Gut Biological Network created the most comprehensive and consensus-driven annotation of human gut cell identities to date. A five-level consensus taxonomy, developed through two international annotation jamborees and curated on the Cell Annotation Platform (CAP), resolves 94 homeostatic cell identities across epithelial, stromal, and immune lineages. We provide scANVI models for label transfer. The atlas will continue to be versioned so that new studies, technologies, and population cohorts can be added within a shared cellular, anatomical, and metadata coordinate system.

Integrated Objects

HGCA All Cells — The HGCA All Cells object contains an scVI integration with sample as batch for all cells in the HGCA. The HGCA all cells can be used as a reference to find cross-lineage gene programs and patterns as well as covariate effects. The HGCA is further split into lineages where lineage-specific scVI integration was performed for context-specific clustering and further sub-clustering.

HGCA Lymphoid — The HGCA Lymphoid lineage contains an scVI embedding of all lymphoid cells in the atlas for lineage-specific label transfer and examination of lymphoid gene programs in context.

HGCA Myeloid — The HGCA Myeloid lineage contains an scVI embedding of all myeloid cells in the atlas for lineage-specific label transfer and examination of lymphoid gene programs in context.

HGCA Epithelial — The HGCA Epithelial lineage contains an scVI embedding of all epithelial cells in the atlas for lineage-specific label transfer and examination of lymphoid gene programs in context.

HGCA Stromal — The HGCA Stromal lineage contains an scVI embedding of all stromal cells in the atlas for lineage-specific label transfer and examination of lymphoid gene programs in context.