An integrated skin cell atlas decodes the pilosebaceous unit

Single-cell and spatial transcriptomics have transformed the ability to chart human tissue organization at high resolution, enabling the construction of reference atlases and robust gene marker identification. The skin, as the largest human organ, lacks a comprehensive integrated atlas, particularly for the pilosebaceous unit, a key epithelial structure involved in homeostasis and disease. We present a framework for a scalable healthy Human Skin Cell Atlas that systematically integrates 34 publicly available single-cell RNA sequencing datasets comprising 818,951 cells, with harmonized metadata and standardized cell type nomenclature. In addition, we generated a new high-resolution Visium HD spatial transcriptomics dataset, demonstrating the benefits of integrating spatial information with single-cell data and enabling the mapping of hair follicle compartments and the identification of critical signaling hubs. The integrated atlas revealed cell types not detectable in individual datasets, including Merkel cells, and refined the classification of pilosebaceous unit subtypes. Importantly, the Human Skin Cell Atlas enables the rapid annotation of new datasets and the assessment of mapping uncertainties, facilitating the discovery of rare or disease-specific cell types. This resource provides a comprehensive and expandable reference for healthy human skin and illustrates the value of combining single-cell and spatial data for studies of tissue organization, disease mechanisms, and regenerative medicine.