As a final demonstration, we considered a similar problem of community-wide integration for circulating human peripheral blood cells, which is one of the most widely profiled systems with diverse single-cell technologies. Exploring publicly available studies of either COVID-19 samples or healthy controls, we accumulated a collection of 14 studies with scRNA-seq measurements, representing a total of 3.46M cells from 639 individuals. Data from 11 of the studies was obtained from a recently published collection of standardized single-cell sequencing datasets62. We performed unsupervised atomic sketch integration, yielding a harmonized collection in which we annotated 30 cell states (Fig. 5a). As a subset of our samples were not depleted for granulocytes, our collection includes a distinct population of neutrophils that were absent in our previous Azimuth reference of human PBMC. Moreover, we identified specific populations of activated granulocytes and B cells that were specific to COVID-19 samples (Supplementary Fig. 7a). Consistent with previous reports, monocytes in COVID-19 samples sharply upregulated interferon response genes63,64, but were correctly harmonized with healthy monocytes (Fig. 5b and Supplementary Fig. 7b). By matching shared cell types across
