SampleDisco tutorial demo data: scRNA-seq and scATAC-seq COVID-19 PBMC subsets

Demo input data for the SampleDisco documentation tutorials (https://j041120h.github.io/SampleDisco_tutorial/). Single-cell AnnData (.h5ad) datasets, COVID-19 and healthy PBMC subsets: test_RNA.h5ad – scRNA-seq, 29,989 cells × 14,881 genes, 8 samples (CoV/HD donors). test_ATAC.h5ad – scATAC-seq, 29,021 cells × 230,356 peaks, 8 samples (SRA accessions). test_multiomics_integrated.h5ad – the scGLUE-integrated RNA+ATAC object (59,010 cells: 29,989 RNA + 29,021 ATAC), carrying the joint cell embedding in obsm['X_glue'] and the sample-removed obsm['Z_clust'], plus joint cell-type labels. Use it to skip scGLUE training: in the multi-omics pipeline set multiomics_integration: false and point multiomics_integrated_h5ad_path at this file. The RNA and ATAC modalities are unpaired (no shared donors). Each cell carries a sample label and a sev.level severity phenotype (1 = healthy, 4 = severe COVID-19). These files drive the RNA, ATAC, and multi-omics tutorial examples of SampleDisco, a cross-omics sample-embedding tool for single-cell data.