Efficient Integration of Human Skin Single-Cell RNA Sequencing Data

This document presents a comprehensive effort to create an integrated human skin cell atlas, aiming to establish a consensus nomenclature for normal human skin. The study reanalyzed 24 individual scRNA-seq datasets, harmonizing metadata and cell annotations, and made the data accessible through "Skin Explorer," an interactive tool. A molecular and spatial atlas of human skin fibroblasts was developed, revealing shared disease-related fibroblast subtypes across tissues. The integration of 463,503 cells from 145 donors and 24 studies resulted in the integrated human skin cell atlas (iHSCA), serving as a reference for healthy human skin. Additionally, a deep learning-based method, scSEA, was developed for unbiased reference mapping, allowing for the discovery of new cell types.