Skin sc-RNASeq from seven body sites (face, scalp, axilla, palmoplantar, arm, leg, and back)

This sc-RNAseq dataset is composed of disease-unaffected epidermal samples from 96 skin biopsies: 18 from published datasets - GSE173706, GSE249279 – and 78 newly generated ones. Biopsy sample and protocol details, and curated cell-type signature genes, are available in the scRNASeq_source_info_FigShare spreadsheet of this dataset. Processed Seurat object are provided herein. Raw data are available in SRA (id PRJNA1054546). Biopsies originated from seven body sites (face, scalp, axilla, palmoplantar, arm, leg, and back). The skin biopsies were separated into epidermis and dermis before dissociated and enriched for various cell fractions (keratinocytes, fibroblasts, and endothelial cells) and immune cells (myeloid and lymphoid cells) to up sample rare cell types. In total, across body sites, 274,834 cells were profiled, including 96,194 keratinocytes. <i>Seurat v3.0.</i> was utilized to normalize, scale, and reduce the dimensionality of the data. Low quality cells containing less than 200 genes per cell as well as greater than 5,000 genes per cell were filtered out. Cells containing more mitochondrial genes than the permitted quantile of 0.05 were removed. Ambient RNA was removed using R package <i>SoupX</i> v1.6.2. Doublets were removed using <i>scDblFinder</i> v1.12.0. Principal components (PC) were obtained from the topmost 2,000 variable genes, and the Uniform Manifold Approximation and Projection (UMAP) dimensional reduction technique was applied to the 30 topmost variable PC-reduced dataset. Batch effect correction was performed utilizing <i>harmony</i> v1.0, using donor as batch. After batch correction, cells were clustered using shared nearest neighbor modularity optimization-based clustering. Cluster marker genes were identified with <i>FindAllMarkers</i>; cluster corresponding cell type was identified by comparing marker genes to curated cell-type signature genes. Differential expression by keratinocyte subtype was performed with Seurat (v4.3.0) <i>FindMarkers</i> function by comparing keratinocyte subtype to non-keratinocyte clusters. The log fold-change of the average expression between a keratinocyte subtype cluster compared to the rest of clusters is utilized as keratinocyte-subtype gene expression statistic.

本单细胞RNA测序（sc-RNAseq）数据集由96份未患病的表皮样本组成，其中18份来自已发表数据集GSE173706与GSE249279，剩余78份为新生成的样本。活检样本的相关信息、实验流程细节以及经过整理的细胞类型特征基因，可在本数据集的scRNASeq_source_info_FigShare表格中获取。本研究提供了处理完成的Seurat对象，原始数据则存储于SRA数据库（编号PRJNA1054546）。本次活检样本取自7个身体部位，分别为面部、头皮、腋窝、掌跖、手臂、腿部及背部。在进行解离并富集各类细胞组分（角质形成细胞、成纤维细胞与内皮细胞）和免疫细胞（髓系与淋巴系细胞）以提升稀有细胞类型的采样量前，研究人员先将皮肤活检样本分离为表皮层与真皮层。最终跨所有身体部位共完成274834个细胞的测序分析，其中包含96194个角质形成细胞。本研究使用<i>Seurat v3.0</i>对数据进行标准化、缩放与降维处理。首先过滤掉每细胞基因数少于200或多于5000的低质量细胞，同时移除线粒体基因占比超过0.05分位数阈值的细胞。使用R包<i>SoupX v1.6.2</i>去除环境RNA污染，通过<i>scDblFinder v1.12.0</i>移除双细胞。基于前2000个高可变基因获取主成分（Principal components，PC），并对经过主成分降维的前30个高可变数据集应用均匀流形近似与投影（Uniform Manifold Approximation and Projection，UMAP）降维技术。使用<i>harmony v1.0</i>以供体为批次单位进行批次效应校正。批次校正完成后，基于共享最近邻模块化优化算法对细胞进行聚类。使用<i>FindAllMarkers</i>识别聚类标记基因，通过将标记基因与已整理的细胞类型特征基因进行比对，确定每个聚类对应的细胞类型。针对角质形成细胞亚型的差异表达分析采用Seurat（v4.3.0）的<i>FindMarkers</i>函数完成，即将角质形成细胞亚型聚类与非角质形成细胞聚类进行比较。本研究将角质形成细胞亚型聚类与其余聚类的平均表达量的对数倍数变化作为角质形成细胞亚型的基因表达统计量。