Scanpy

For analyzing single-cell RNA-seq data with the standard Python toolkit. Covers the full workflow from QC and normalization through clustering, UMAP visualization, and marker gene identification. Built on AnnData, works with 10X data and h5ad files. Note that version 1.12 requires Python 3.12+ and deprecated the per-plot save parameter in favor of autosave settings. The docs correctly warn that rank_genes_groups is fine for exploratory cluster markers but shouldn't be used for rigorous differential expression between conditions, you need pseudobulk and proper DE tools for that. Includes a QC script for automated filtering. Good default choice for exploratory scRNA-seq work unless you need probabilistic models, then reach for scvi-tools instead.