Single cell data used in the BENGAL pipeline

Single cell or single nucleus RNA-seq data used in the study: Song, Y., Miao, Z., Brazma, A. <i>et al.</i> Benchmarking strategies for cross-species integration of single-cell RNA sequencing data. <i>Nat Commun</i> <b>14</b>, 6495 (2023). https://doi.org/10.1038/s41467-023-41855-woriginal_data_curated/: raw counts data were first collected from public repositories (ref. data availability in the paper). They were made into .h5ad files, and were processed in terms of 1) placing the raw count matrices into the .X slot; 2) converting author cell type annotations into cell ontology terms; 3) matching gene identifiers to ENSEMBL gene ids and setting them as var_names. cell_ontology_mapped/: datasets containing this subfolder means that the scOntoMatch method was run to align cell ontology annotations across species (ref. Methods in the paper). Those without this subfolder mean that this step was not necessary, or specific annotations were used. *Note that in the heart dataset, the <i>X.laevis</i> data name "xlaevis_heart_raw_count_onto_anno_gene_name.h5ad" was the original data with curation except for converting gene IDs. Since only <i>X.tropicalis</i> was available in ENSEMBL, gene names were then converted to <i>X.tropicalis</i> gene IDs by summing the counts of duplicated genes, generating "xlaevis_heart_onto_anno_xtropicalis_gene_id.h5ad". <i>X. laevis </i>is a tetraploid (2N=36) while <i>X. tropicalis</i> is diploid (2N=20).<br>