Tabula Sapiens Overview

Single Cell Transcriptomics of 25 Human Organs to Create a Tabula Sapiens

I. Overview

Our recent work developing and scaling single-cell RNA expression profiling technologies1–10 has produced cell expression databases covering the major organs of a mammal11 (mouse Mus musculus, Figure 1) and, more recently, a non-human primate (mouse lemur Microcebus murinus, Figures 3 and 4, our unpublished data). We propose to leverage our unique skills, experienced team (Figure 2), and data infrastructure obtained from these first organism-wide cell profiling projects to create a high-quality benchmark first-draft human cell atlas of two million cells from 25 organs in eight normal human subjects. To examine the feasibility of our approach for human organs, we recently tested it on the lungs of three human subjects. We obtained full mRNA expression profiles of 80,000 cells representing 39 (91%) of the 43 known human lung cell types, plus 15 newly discovered cell types or subtypes (Figure 5). The profiled cells obtained include rare cell types (e.g. neuroendocrine cells) comprising as few as 0.01% of all human lung cells; the only extant lung cell types not identified in this pilot study were two rare cell types (serous and mucous epithelial cells) that may have been missed because no specific markers are known, and two exceedingly rare cell types, intrinsic neurons and glia, which comprise only ~0.0005% of human lung cells so were not expected to be found at this level of cell coverage. The locations and histology of the newly identified cell types were established by immunostaining or in situ hybridization using cell type specific markers obtained by computational comparison of the obtained cell type expression profiles. Before our analysis, no full mRNA expression profiles were available for any human lung cell types. Eighteen (31%) of the 58 human lung cell types and subtypes have not been found in mice, including 13 (87%) of the 15 newly identified types (Figure 5). These results highlight the importance and urgency of applying the approach globally across human organs. Our plan is to train the organ-specific rapid-response teams that prepared, surface-stained and sorted the cell suspensions for each organ in the Tabula Muris and Tabula Microcebus projects in the best practices established in this successful pilot experiment for human lung. Best practices include compiling from the literature a comprehensive list of all known cell types in each organ and their relative abundance, and conducting in collaboration with our human tissue processing experts preliminary experiments to ensure that each organ can be dissociated and viable cell populations obtained and then balanced by FACS sorting to ensure optimal coverage of all known cell populations, including rare cell populations within each tissue compartment of the organ. These preliminary experiments will be completed at the beginning of the grant period, so that rapid response teams for all organs can be activated immediately when the first human subject becomes available, as done for the Tabula Microcebus project. This approach ensures high quality data and comparability of data across all organs of an individual.