Revealing the Critical Regulators of Cell Identity in the Mouse Cell Atlas.
Suo S, Zhu Q, Saadatpour A, Fei L, Guo G, Yuan GC. · 2018
Read Revealing the Critical Regulators of Cell Identity in the Mouse Cell Atlas. by Suo S, Zhu Q, Saadatpour A, Fei L, Guo G, Yuan GC. (2018) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc6281296.txt.
Introduction
Suo and co-authors mined the mouse cell atlas for the regulators that define each cell type. Their study appeared in Cell Reports. Single cell methods have now named all major mouse cell types. What keeps each type's identity was less clear. The team analysed the published atlas data by computer. They searched for gene control units whose activity varied most across types.
Two hundred two control units, called regulons, stood out. From them the team predicted a small set of key regulators per cell type. Automated checks against literature and data backed many predictions. A free web portal lets others browse the network. Cell identity seems to rest on few master switches. That finding echoes classic work in blood and muscle. The resource should help labs design cell conversion studies.
Catalog record
| CREATOR | Suo S, Zhu Q, Saadatpour A, Fei L, Guo G, Yuan GC. |
|---|---|
| YEAR | 2018 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Revealing the Critical Regulators of Cell Identity in the Mouse Cell Atlas By Shengbao Suo, Qian Zhu, Assieh Saadatpour, Lijiang Fei, Guoji Guo, Guo-Cheng Yuan Published in: Cell reports DOI: 10.1016/j.celrep.2018.10.045 ABSTRACT SUMMARY 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. INTRODUCTION A multi-cellular organism contains diverse cell types; each has its own functions and morphology. A fundamental goal in biology is to characterize the entire cell-type atlas in human and model organisms. With the rapid development of single-cell technologies, great strides have been made in the past few years (Svensson et al., 2018). Multiple groups have made tremendous progresses in mapping cell atlases in complex organs (such as mouse brain and immune system) (Rosenberg et al., 2018; Saunders et al., 2018; Stubbington et al., 2017; Zeisel et al., 2018), early embryos (such as in C. elegans and zebrafish) (Cao et al., 2017; Wagner et al., 2018), or even entire adult animals (such as Schmidtea mediterranea and mouse) (The Tabula Muris Consortium et al., 2018; Fincher et al., 2018; Han et al., 2018; Plass et al., 2018). International collaborative efforts are underway to map out the cell atlas in human (Regev et al., 2017). How do cells maintain their identity? While it is clear the maintenance of cell identity involves the coordinated action of many regulators, transcription factors (TFs) have been long recognized to play a central role. In several cases, the activity of a small number of key TFs, also known as the master regulators, are essential for cell identity maintenance: depletion of these regulators cause significant alteration of cell identity, while forced expression of these regulators can effectively reprogram cells to a different cell type (Han et al., 2012; Ieda et al., 2010; Riddell et al., 2014; Takahashi and Yamanaka, 2006). However, for most cell types, the underlying gene regulatory circuitry is incompletely understood. With the increasing diversity of gene expression programs being identified through single-cell analysis, an urgent need is to understand how these programs are established during development, and to identify the key regulators responsible for such processes. Systematic approaches for mapping gene regulatory networks (GRNs) have been well established. The most direct approach is through genome-wide occupancy analysis, using experimental assays such as chromatin immunoprecipitation sequencing (ChIP-seq), chromatin accessibility, or long-range chromatin interaction assays (ENCODE Project Consortium, 2012). However, this approach is not scalable to a large number of cell types, and its application is often limited by the number of cells that can be obtained in vivo. An alternative, more generalizable approach is to computationally reconstruct GRNs based on single-cell gene expression data (Fiers et al., 2018), followed by more focused experimental validations. In this study, we took this latter approach to build a comprehensive mouse cell network atlas. To this end, we took advantage of the recently mapped mouse cell atlas (MCA) derived from comprehensive single-cell transcriptomic analysis (Han et al., 2018), and combined with a computational algorithm to construct GRNs from single-cell transcriptomic data. Our analysis indicates that most cell types have distinct regulatory network structure and identifies regulators that are critical for cell identity. In addition, we provide an interactive web-based portal for exploring the mouse cell network atlas. RESULTS Reconstructing Gene Regulatory Networks Using the MCA To comprehensively reconstruct the gene regulatory networks for all major cell types, we applied the SCENIC pipeline (Aibar et al., 2017) to analyze the MCA data. In brief, SCENIC links cis-regulatory sequence information together with single-cell RNA sequencing (RNA-seq) data. SCENIC contains three main steps, including co-expression analysis, target gene motif enrichment analysis, and regulon activity evaluation. The main outcomes contain a list of regulons (each representing a TF along with a set of co-expressed and motif significantly enriched target genes), and the regulon activity scores (RAS) for each cell (Figure 1A). To improve computational efficiency and robustness, we modified the original pipeline to analyze pooled data instead. Specifically, we divided the entire cell population into small groups with similar cell states. This was achieved by random, non-overlapping sampling from the same tissue and cell type. Each group contains 20 cells. We applied SCENIC to infer regulons based on the group-averaged gene expression profiles, and the RAS scores were calculated at the single-cell level as in the original study (Aibar et al., 2017). By testing its performance on several representative tissues, we found that our modified approach separated cell types more effectively compared to the original implementation (Figure S1; see STAR Methods for details). We focused on a representative, well-annotated subset of MCA data (Han et al., 2018), containing 61,637 cells sampled from 43 tissues. Previous analysis has identified 98 main cell types (Han et al., 2018). By applying the modified SCENIC approach described above, we identified 202 significant regulons containing 8,461 genes (Table S1). The size of each regulon varies from 10–2,502 genes, with a median size of 73 genes. Strikingly, different cell types are well-separated using the RAS-based distance (Figures 1B and S2A; Table S2). Even within a single tissue, different cell types can be separated based on the regulon activities (Figures 1C and S2B). For example, each of four main cell types identified in liver occupies a distinct territory in the t-Distributed Stochastic Neighbor Embedding (tSNE) plot (Figure 1C). One question of interest is whether cells of the same type may have different regulatory circuitries across tissues. Such differences would be relevant for investigating cell-environment interactions (Figures 1D and S2C). To address this question, we focused on stromal cells, which can be found in a wide variety of tissues, providing support, structural and anchoring functions. The behavior of stromal cells is well known to be highly plastic, a necessary property for supporting a diverse range of tissue development (Lee et al., 2006). Indeed, we found that the stromal cells from different tissues tend to have distinct regulon activities (Figure 1D). While stromal cells are clustered together from the global view of t-SNE map, closer examination suggests the subpopulations from different tissues—such as uterus, mammary gland, bladder, and pancreas—are well separated. Similar refined structure can be found in various other cell types, such as T cells and epithelial cells (Figure S2C). Taken together, regulon-based activity score in different cell types provides a new avenue to investigate the potential regulatory mechanism in inter- and intra-cell type variations. Our analysis indicates that GRN differences are primarily driven by cell type differences but further modulated by tissue environment differences. Comparative Analysis Identified Essential Regulators for the Maintenance of Cell Identity Our comprehensive network analysis provides an opportunity to systematically identify