Cell and UBERON ontology term mappings were extracted from the FANTOM5 sample ontology 6 for primary cell and tissue samples, respectively, using indirect and direct ‘is_a’ and ‘part_of’ relationships. Ontology terms were manually selected to construct groups (facets) of samples that were mutually exclusive and to cover as broad histological and functional annotations as possible. 362 primary cell samples and 138 tissue and whole blood samples were grouped into 69 cell type facets and 41 organ/tissue facets, respectively (the groupings of samples into facets are provided in Supplementary Tables 10 and 11). A few samples were ignored because they were difficult to assign to a facet with certainty, which means that the number of samples within facets is slightly lower than the total number of samples.