Skip to main content
Premium Trial:

Request an Annual Quote

Researchers Generate Single-Cell Transcriptome Atlas for 20 Mouse Organs

NEW YORK (GenomeWeb) – Researchers have generated what they've dubbed a Tabula Muris — an atlas of mouse cell types and their gene expression patterns.

Researchers from the Tabula Muris Consortium isolated more than 100,000 cells from 20 different organs of three female and four male mice to generate their atlas. As they reported today in Nature, the investigators used booth FACS-based cell capture and microfluidic-droplet-based capture to isolate the cells for single-cell RNA sequencing. Their data allows for the comparison of gene expression across cell types and tissues, and represents an attempt to capture organism-wide cellular diversity.

"We believe that this rich collection of annotated cells will be a useful resource for biomedical research," co-corresponding author Spyros Darmanis from the Chan Zuckerberg Biohub said in a statement. "Through this project we uncovered gene expression patterns that allow the identification of distinct cell types originating from a wide range of tissues across an entire organism, which can be used for cell selection, targeting, and reprogramming."

The researchers isolated 44,949 cells using FACS and 55,656 cells using a microfluidic droplet approach from the seven mice for analysis. The mice were three months old — the equivalent of a 20 year-old human. To determine and assign cell types, the researchers analyzed each organ independently and used clustering analyses along with known markers.

They teased out these cells from numerous organs — including the brain, heart, pancreas, and thymus, among others — and sequenced the single-cell transcriptomes to an average depth of 814,488 reads per cell for FACS and 7,709 unique molecular identifiers for the microfluidic approach.

The researchers noted that FACS captured fewer cells, though more molecules per cell, than the microfluidic approach. However, they also found that the two methods largely agreed in their bulk gene expression profiles.

When the researchers used the t-SNE algorithm to visualize the relationship between different cell types from various organs isolated via FACS, they discovered that cells from different organs often mixed. For instance, clusters 2 and 48 both contained endothelial cells from five or more organs, while clusters 1 and 24 contained mesenchymal and stromal cells from four or more organs. The researchers also generated heterogeneity scores for each cluster to gauge whether they were made up of related or unrelated cell types.

When they particularly focused on cells annotated as T cells to show that they could investigate cell types across organ types, the researchers uncovered five clusters. Cluster 0, for instance, contained thymic cells undergoing VDJ recombination, while cluster 2 encompassed non-thymic T cells that expressed the IL2 receptor, suggesting they were activated.

The researchers further found that transcription factor expression could broadly define cell type. When they clustered cells by the average expression of the 1,016 transcription factors expressed in their dataset, the investigators saw that the dendogram they generated was largely similar to the one they'd generated using all expressed genes. That was not the case when they repeated their analysis with cell-surface markers, RNA splicing factors, or both, indicating that transcription factors could better determine cell type than those molecular factors.

This suggested to the researchers that their atlas could inform cell-reprogramming protocols. In addition, they said that their mouse transcriptome atlas could help uncover new cell types, aid in uncovering novel gene expression in different cell types, and be used to compare cells across organs.