Revealing the Critical Regulators of Cell Identity in the Mouse Cell Atlas

Abstract

Recent progress in single-cell technologies has enabled the identification of all major cell types in mouse. However, for most cell types, the regulatory mechanism underlying their identity remains poorly understood. By computational analysis of the recently published mouse cell atlas data, we have identified 202 regulons whose activities are highly variable across different cell types, and more importantly, predicted a small set of essential regulators for each major cell type in mouse. Systematic validation by automated literature and data mining provides strong additional support for our predictions. Thus, these predictions serve as a valuable resource that would be useful for the broad biological community. Finally, we have built a user-friendly, interactive web portal to enable users to navigate this mouse cell network atlas.

Keywords: cell type; co-expression analysis; database; gene regulatory network; mouse cell atlas; single-cell RNA-seq; text mining; transcription factor; visualization; web application.

Figure 1.. Mapping Mouse Cell Network Atlas with Regulon Activity

(A) Schematic overview of the computational approach in this study. A modified SCENIC pipeline is used to infer cell-type-specific gene regulatory networks. (B–D) t-SNE map for all sampled single cells (~61 k) based on regulon activity scores (RAS), each cell is color-coded based on major cell-type assignment. (B) All sampled cells (~61 k) are highlighted. (C) Liver cells are highlighted. (D) Stromal cells are highlighted. See also Figures S1 and S2 and Tables S1, S2, and S3.

Figure 2.. Cell-Type-Specific Regulon Activity Analysis

(A–D) Erythroblast. (A) Rank for regulons in erythroblast cell based on regulon specificity score (RSS). (B) Erythroblast cells are highlighted in the t-SNE map (red dots). (C) Binarized regulon activity scores (RAS) (do Z score normalization across all samples, and set 2.5 as cutoff to convert to 0 and 1) for top regulon Lmo2 on t-SNE map (dark green dots). (D) SEEK co-expression result for target genes of top regulon Lmo2 in different GEO datasets. The x axis represents different datasets, and the y axis represents the co-expression significance of target genes in each dataset. Erythroblast related datasets with significant correlation (p value < 0.01) are highlighted by yellow dots. (E–H) Same as (A)–(D) but for B cells. (I–L) Same as (A)–(D) but for oligodendrocytes. (M–P) Same as (A)–(D) but for alveolar type II cells. See also Figure S2 and Table S4.

Figure 3.. Identification of Combinatorial Regulon Modules

(A) Identified regulon modules based on regulon connection specificity index (CSI) matrix, along with representative transcription factors, corresponding binding motifs, and associated cell types. (B) Zoomed-in view of module M7 identifies sub-module structures. (C) Different sub-modules in M7 are associated with distinct immune cell types and regulon activities. See also Figure S3.

Figure 4.. A Summary View of the Mouse Cell Network Atlas

(A) Relatedness network for the 818 cell types based on similarity of regulon activities. Each group represents a set of highly related cell types. (B) Sankey plot shows relationship between cell-type groups G1–G9 and regulon modules M1–M8, the thematic cell type composition within each cluster is indicated by the corresponding wordcloud plot. (C) A representative screenshot of the web portal obtained by querying “cumulus cells.” See also Figure S4 and Tables S2 and S3.