scRNAseq-preprocessed Data - Ferroptosis as a Physiologic Vulnerability of Iron-Recycling Macrophages

Processed AnnData (.h5ad) objects underlying the single-cell RNA-seq analysis in "Ferroptosis as a Physiologic Vulnerability of Iron-Recycling Macrophages" (full citation to be added upon publication). Mouse spleen, CD45+/CD64+ sorted cells, 10x Genomics scRNA-seq, comparing Ctrl and DKO (NRF2/BVRA double-knockout) animals (3 Ctrl, 4 DKO), with a focus on red pulp macrophages (RPM) and other myeloid subsets. Files included: adata_full_after_coarse_annotation_v1.h5ad — full dataset (all sorted cells) after QC filtering, doublet removal, normalization, Leiden clustering, and coarse cell-type annotation (Myeloid / Lymphoid / Mixed). adata_sub_annotated.h5ad — myeloid/macrophage subset after subclustering and fine-grained cell-type annotation (RPM, transitional RPM stages, monocyte/macrophage subtypes). This is the object used directly by the downstream differential expression, enrichment analysis, and figure-generation code. Both objects are standard anndata.AnnData objects (format compatible with anndata>=0.10). Key fields: .obs["sample"] — per-animal sample ID (Ctrl_1–3, DKO_1–4) .obs["orig_ident"] — genotype/condition (Ctrl vs. DKO) .obs["celltype"] / .obs["leiden_sub"] — cell-type annotation / cluster identity (sub-annotated object only) .layers["counts"] — raw counts (HVG-restricted gene space) .raw — log-normalized expression, full gene space .obsm["X_umap"], .obsm["X_pca"] — embeddings Note: 01_clustering_QC_annotation_DE.ipynb labels the two groups "Ctrl"/"DKO" in .obs["orig_ident"] and its own plots; 02_paper_figures.ipynb derives its own "LOX ctrl"/"CRE DKO" labels directly from .obs["sample"] (independent of orig_ident). Both refer to the same animals/groups. Associated code: <GitHub repo URL> (notebooks 01_clustering_QC_annotation_DE.ipynb and 02_paper_figures.ipynb), archived at <Zenodo software DOI, if you create a separate Software record>. Note on reproducibility: Leiden clustering is not perfectly deterministic across package versions/hardware; these files represent the exact clustering and annotation used for the figures in the manuscript. Re-running the clustering code from raw data may produce minor differences in cluster boundaries.