Single-Cell RNA-seq Workshop: Processed Heart Development snRNA-seq Checkpoints

This record hosts pre-computed Seurat checkpoints for the Single-Cell RNA-seq Workshop, a methods-focused tutorial that analyses single-nucleus RNA-seq data from human heart tissue across foetal, young, and adult developmental stages. Source data All checkpoints are derived from the raw count matrix and cell metadata published in [Zenodo record 18237749](https://zenodo.org/records/18237749), which in turn comes from [Sim et al. (2021), Circulation 143(10):1614–1628](https://doi.org/10.1161/CIRCULATIONAHA.120.051921). The four disease samples were excluded; nine healthy samples were used (3 foetal, 3 young, 3 adult). Files 01_qc_filtered.rds — Module 1 output. All 9 samples merged into a single Seurat (v5) object after per-cell quality control (`nFeature_RNA ≥ 500`, `nCount_RNA ∈ [2500, 40000]`, `percent.mt ≤ 20`) and gene-level filtering (mitochondrial, ribosomal, sex-chromosome, and unannotated genes removed). 02_integrated_clustered.rds — Module 2 output. 10,000-cell stratified downsample (proportional to per-sample cell counts), then: SCTransform v2 normalisation with `glmGamPoi` PCA on the top 20 components Harmony integration on `sample` UMAP on the Harmony embedding The active identity is resolution 0.4, producing 17 clusters. 03_annotated.rds— Module 3 output. Cluster-level marker discovery via Seurat::FindAllMarkers, plus per-cell `cell_type` and `cell_type_broad` columns from a hand-curated cluster → cell-type mapping (Cardiomyocyte subtypes, Fibroblasts, Endothelial, Macrophages, Epicardial, Neurons, T cells; broad categories also included). afternoonSession.zip — Afternoon session (Module 5, 6, 7) input and intermediate results. Download, unzip and add files to data or results folder according to instructions. Reproducibility Generated with R 4.5.2, Bioconductor 3.22, Seurat 5.4.0, SeuratObject 5.3.0, harmony 1.2.4, glmGamPoi 1.22.0, sctransform 0.4.3, edgeR 4.8.2, limma 3.66.0, speckle 1.10.0. Exact package versions are pinned in [`renv.lock`](https://github.com/phipsonlab/single_cell_workshop/blob/main/renv.lock) in the workshop repository. Original analysis workflow: https://bphipson.github.io/Human_Development_snRNAseq/ Sim CB, Phipson B, Ziemann M, et al. Sex-Specific Control of Human Heart Maturation by the Progesterone Receptor. Circulation. 2021;143(10):1614-1628. doi:https://doi.org/10.1161/CIRCULATIONAHA.120.051921本数据集包含为**单细胞RNA测序工作坊（Single-Cell RNA-seq Workshop）**预计算的Seurat检查点，该工作坊是一门聚焦方法学的教程，用于分析覆盖胎儿期、青年期与成年期发育阶段的人类心脏组织单细胞核RNA测序（single-nucleus RNA-seq）数据。 ### 源数据 所有检查点均源自[Zenodo记录18237749](https://zenodo.org/records/18237749)发布的原始计数矩阵与细胞元数据，而该记录的数据来自[Sim等人（2021年），《循环（Circulation）》143(10):1614–1628](https://doi.org/10.1161/CIRCULATIONAHA.120.051921)。本数据集剔除了4例疾病样本，仅使用9例健康样本（胎儿期、青年期、成年期各3例）。 ### 文件 1. `01_qc_filtered.rds` — 模块1输出结果。经单细胞质量控制（`nFeature_RNA ≥ 500`，`nCount_RNA ∈ [2500, 40000]`，`percent.mt ≤ 20`）与基因水平过滤（移除线粒体、核糖体、性染色体及未注释基因）后，将9例样本合并为单个Seurat（v5）对象。 2. `02_integrated_clustered.rds` — 模块2输出结果。先按样本细胞数比例进行分层下采样至10000个细胞，随后： - 采用`glmGamPoi`进行SCTransform v2标准化 - 对前20个主成分进行主成分分析（PCA） - 基于`sample`分组进行Harmony整合 - 基于Harmony嵌入结果进行UMAP降维 当前聚类标识为分辨率0.4，共生成17个聚类。 3. `03_annotated.rds` — 模块3输出结果。通过Seurat::FindAllMarkers进行聚类水平的标志物筛选，并结合人工手动整理的聚类-细胞类型映射表，新增了单细胞水平的`cell_type`与`cell_type_broad`列（涵盖心肌细胞亚型、成纤维细胞、内皮细胞、巨噬细胞、心外膜细胞、神经元、T细胞等类别，并包含更宽泛的分类）。 4. `afternoonSession.zip` — 下午课程（模块5、6、7）的输入与中间结果文件。请下载解压后，按照操作说明将文件添加至数据或结果文件夹。 ### 可复现性 本数据集基于R 4.5.2、Bioconductor 3.22、Seurat 5.4.0、SeuratObject 5.3.0、harmony 1.2.4、glmGamPoi 1.22.0、sctransform 0.4.3、edgeR 4.8.2、limma 3.66.0、speckle 1.10.0生成。精确的软件包版本已在工作坊仓库的[`renv.lock`](https://github.com/phipsonlab/single_cell_workshop/blob/main/renv.lock)文件中锁定。 原始分析流程：https://bphipson.github.io/Human_Development_snRNAseq/ 参考文献： Sim CB, Phipson B, Ziemann M, 等. 孕激素受体对人类心脏成熟的性别特异性调控. 循环. 2021;143(10):1614-1628. doi:https://doi.org/10.1161/CIRCULATIONAHA.120.051921