A pooled shRNA screen for regulators of primary mammary stem and progenitor cells identifies roles for Asap1 and Prox1

OmicNavigator: open-source software for the exploration, visualization, and archival of omic studies

BACKGROUND

The results of high-throughput biology ('omic') experiments provide insight into biological mechanisms but can be challenging to explore, archive and share. The scale of these challenges continues to grow as omic research volume expands and multiple analytical technologies, bioinformatic pipelines, and visualization preferences have emerged. Multiple software applications exist that support omic study exploration and/or archival. However, an opportunity remains for open-source software that can archive and present the results of omic analyses with broad accommodation of study-specific analytical approaches and visualizations with useful exploration features.

RESULTS

We present OmicNavigator, an R package for the archival, visualization and interactive exploration of omic studies. OmicNavigator enables bioinformaticians to create web applications that interactively display their custom visualizations and analysis results linked with app-derived analytical tools, graphics, and tables. Studies created with OmicNavigator can be viewed within an interactive R session or hosted on a server for shared access.

CONCLUSIONS

OmicNavigator can be found at https://github.com/abbvie-external/OmicNavigator.

Mcam inhibits macrophage-mediated development of mammary gland through non-canonical Wnt signaling

While canonical Wnt signaling is well recognized for its crucial regulatory functions in cell fate decisions, the role of non-canonical Wnt signaling in adult stem cells remains elusive and contradictory. Here, we identified Mcam, a potential member of the non-canonical Wnt signaling, as an important negative regulator of mammary gland epithelial cells (MECs) by genome-scale CRISPR-Cas9 knockout (GeCKO) library screening. Loss of Mcam increases the clonogenicity and regenerative capacity of MECs, and promotes the proliferation, differentiation, and ductal morphogenesis of mammary epithelial in knockout mice. Mechanically, Mcam knockout recruits and polarizes macrophages through the Il4-Stat6 axis, thereby promoting secretion of the non-canonical Wnt ligand Wnt5a and its binding to the non-canonical Wnt signaling receptor Ryk to induce the above phenotypes. These findings reveal Mcam roles in mammary gland development by orchestrating communications between MECs and macrophages via a Wnt5a/Ryk axis, providing evidences for non-canonical Wnt signaling in mammary development.

Prox1 Suppresses the Proliferation of Breast Cancer Cells via Direct Inhibition of c-Myc Gene Expression

Breast cancer is one of the most lethal malignancies in women worldwide and is characterized by rapid growth and low survival rates, despite advances in tumor biology and therapies. Novel therapeutic approaches require new insights into the molecular mechanisms of malignant transformation and progression. To this end, here, we identified Prox1 as a negative regulator of proliferation and tumor-related metabolism in breast cancer. In particular, we showed that breast tumors from human patients exhibited reduced levels of Prox1 expression, while high expression levels of Prox1 were associated with a favorable prognosis in breast cancer patients. Moreover, we experimentally demonstrated that Prox1 was sufficient to strongly suppress proliferation, migration, and the Warburg effect in human breast cancer cells without inducing apoptosis. Most importantly, over-expression of Prox1 inhibited breast tumor growth in vivo in both heterotopic and orthotopic xenograft mouse models. The anti-tumorigenic effect of Prox1 was mediated by the direct repression of c-Myc transcription and its downstream target genes. Consistently, c-Myc over-expression from an artificial promoter that was not targeted by Prox1 reversed Prox1's anti-tumor effects. These findings suggest that Prox1 has a tumor suppressive role via direct transcriptional regulation of c-Myc, making it a promising therapeutic gene for breast cancer.

Inactivation of LATS1/2 drives luminal-basal plasticity to initiate basal-like mammary carcinomas

Basal-like breast cancers, an aggressive breast cancer subtype that has poor treatment options, are thought to arise from luminal mammary epithelial cells that undergo basal plasticity through poorly understood mechanisms. Using genetic mouse models and ex vivo primary organoid cultures, we show that conditional co-deletion of the LATS1 and LATS2 kinases, key effectors of Hippo pathway signaling, in mature mammary luminal epithelial cells promotes the development of Krt14 and Sox9-expressing basal-like carcinomas that metastasize over time. Genetic co-deletion experiments revealed that phenotypes resulting from the loss of LATS1/2 activity are dependent on the transcriptional regulators YAP/TAZ. Gene expression analyses of LATS1/2-deleted mammary epithelial cells notably revealed a transcriptional program that associates with human basal-like breast cancers. Our study demonstrates in vivo roles for the LATS1/2 kinases in mammary epithelial homeostasis and luminal-basal fate control and implicates signaling networks induced upon the loss of LATS1/2 activity in the development of basal-like breast cancer.

Loss of ASAP1 in the MMTV-PyMT model of luminal breast cancer activates AKT, accelerates tumorigenesis, and promotes metastasis

ASAP1 is a multi-domain adaptor protein that regulates cytoskeletal dynamics, receptor recycling and intracellular vesicle trafficking. Its expression is associated with poor prognosis in a variety of cancers, and can promote cell migration, invasion and metastasis. Although amplification and expression of ASAP1 has been associated with poor survival in breast cancer, we found that in the autochthonous MMTV-PyMT model of luminal breast cancer, ablation of ASAP1 resulted in an earlier onset of tumor initiation and increased metastasis. This was due to tumor cell-intrinsic effects of ASAP1 deletion, as ASAP1 deficiency in tumor, but not in stromal cells was sufficient to replicate the enhanced tumorigenicity and metastasis observed in the ASAP1-null MMTV-PyMT mice. Loss of ASAP1 in MMTV-PyMT mice had no effect on proliferation, apoptosis, angiogenesis or immune cell infiltration, but enhanced mammary gland hyperplasia and tumor cell invasion, indicating that ASAP1 can accelerate tumor initiation and promote dissemination. Mechanistically, these effects were associated with a potent activation of AKT. Importantly, lower ASAP1 levels correlated with poor prognosis and enhanced AKT activation in human ER+/luminal breast tumors, validating our findings in the MMTV-PyMT mouse model for this subtype of breast cancer. Taken together, our findings reveal that ASAP1 can have distinct functions in different tumor types and demonstrate a tumor suppressive activity for ASAP1 in luminal breast cancer.

In vivo genome-editing screen identifies tumor suppressor genes that cooperate with Trp53 loss during mammary tumorigenesis

Breast cancer is a heterogeneous disease that comprises multiple histological and molecular subtypes. To gain insight into mutations that drive breast tumorigenesis, we describe a pipeline for the identification and validation of tumor suppressor genes. Based on an in vivo genome‐wide CRISPR/Cas9 screen in Trp53^+/– heterozygous mice, we identified tumor suppressor genes that included the scaffold protein Axin1, the protein kinase A regulatory subunit gene Prkar1a, as well as the proof‐of‐concept genes Pten, Nf1, and Trp53 itself. Ex vivo editing of primary mammary epithelial organoids was performed to further interrogate the roles of Axin1 and Prkar1a. Increased proliferation and profound changes in mammary organoid morphology were observed for Axin1/Trp53 and Prkar1a/Trp53 double mutants compared to Pten/Trp53 double mutants. Furthermore, direct in vivo genome editing via intraductal injection of lentiviruses engineered to express dual short‐guide RNAs revealed that mutagenesis of Trp53 and either Prkar1a, Axin1, or Pten markedly accelerated tumor development compared to Trp53‐only mutants. This proof‐of‐principle study highlights the application of in vivo CRISPR/Cas9 editing for uncovering cooperativity between defects in tumor suppressor genes that elicit mammary tumorigenesis.

Dashboard-style interactive plots for RNA-seq analysis are R Markdown ready with Glimma 2.0

Glimma 1.0 introduced intuitive, point-and-click interactive graphics for differential gene expression analysis. Here, we present a major update to Glimma that brings improved interactivity and reproducibility using high-level visualization frameworks for R and JavaScript. Glimma 2.0 plots are now readily embeddable in R Markdown, thus allowing users to create reproducible reports containing interactive graphics. The revamped multidimensional scaling plot features dashboard-style controls allowing the user to dynamically change the colour, shape and size of sample points according to different experimental conditions. Interactivity was enhanced in the MA-style plot for comparing differences to average expression, which now supports selecting multiple genes, export options to PNG, SVG or CSV formats and includes a new volcano plot function. Feature-rich and user-friendly, Glimma makes exploring data for gene expression analysis more accessible and intuitive and is available on Bioconductor and GitHub.

A functional genetic screen identifies aurora kinase b as an essential regulator of Sox9-positive mouse embryonic lung progenitor cells

Development of a branching tree in the embryonic lung is crucial for the formation of a fully mature functional lung at birth. Sox9+ cells present at the tip of the primary embryonic lung endoderm are multipotent cells responsible for branch formation and elongation. We performed a genetic screen in murine primary cells and identified aurora kinase b (Aurkb) as an essential regulator of Sox9+ cells ex vivo. In vivo conditional knockout studies confirmed that Aurkb was required for lung development but was not necessary for postnatal growth and the repair of the adult lung after injury. Deletion of Aurkb in embryonic Sox9+ cells led to the formation of a stunted lung that retained the expression of Sox2 in the proximal airways, as well as Sox9 in the distal tips. Although we found no change in cell polarity, we showed that loss of Aurkb or chemical inhibition of Aurkb caused Sox9+ cells to arrest at G2/M, likely responsible for the lack of branch bifurcation. This work demonstrates the power of genetic screens in identifying novel regulators of Sox9+ progenitor cells and lung branching morphogenesis.

Single cell transcriptome atlas of mouse mammary epithelial cells across development

Background

Heterogeneity within the mouse mammary epithelium and potential lineage relationships have been recently explored by single-cell RNA profiling. To further understand how cellular diversity changes during mammary ontogeny, we profiled single cells from nine different developmental stages spanning late embryogenesis, early postnatal, prepuberty, adult, mid-pregnancy, late-pregnancy, and post-involution, as well as the transcriptomes of micro-dissected terminal end buds (TEBs) and subtending ducts during puberty.

Methods

The single cell transcriptomes of 132,599 mammary epithelial cells from 9 different developmental stages were determined on the 10x Genomics Chromium platform, and integrative analyses were performed to compare specific time points.

Results

The mammary rudiment at E18.5 closely aligned with the basal lineage, while prepubertal epithelial cells exhibited lineage segregation but to a less differentiated state than their adult counterparts. Comparison of micro-dissected TEBs versus ducts showed that luminal cells within TEBs harbored intermediate expression profiles. Ductal basal cells exhibited increased chromatin accessibility of luminal genes compared to their TEB counterparts suggesting that lineage-specific chromatin is established within the subtending ducts during puberty. An integrative analysis of five stages spanning the pregnancy cycle revealed distinct stage-specific profiles and the presence of cycling basal, mixed-lineage, and 'late' alveolar intermediates in pregnancy. Moreover, a number of intermediates were uncovered along the basal-luminal progenitor cell axis, suggesting a continuum of alveolar-restricted progenitor states.

Conclusions

This extended single cell transcriptome atlas of mouse mammary epithelial cells provides the most complete coverage for mammary epithelial cells during morphogenesis to date. Together with chromatin accessibility analysis of TEB structures, it represents a valuable framework for understanding developmental decisions within the mouse mammary gland.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13058-021-01445-4.

Single-cell transcriptomics reveals involution mimicry during the specification of the basal breast cancer subtype

Basal breast cancer is associated with younger age, early relapse, and a high mortality rate. Here, we use unbiased droplet-based single-cell RNA sequencing (RNA-seq) to elucidate the cellular basis of tumor progression during the specification of the basal breast cancer subtype from the luminal progenitor population in the MMTV-PyMT (mouse mammary tumor virus-polyoma middle tumor-antigen) mammary tumor model. We find that basal-like cancer cells resemble the alveolar lineage that is specified upon pregnancy and encompass the acquisition of an aberrant post-lactation developmental program of involution that triggers remodeling of the tumor microenvironment and metastatic dissemination. This involution mimicry is characterized by a highly interactive multicellular network, with involution cancer-associated fibroblasts playing a pivotal role in extracellular matrix remodeling and immunosuppression. Our results may partially explain the increased risk and poor prognosis of breast cancer associated with childbirth.

Optimal, Large-Scale Propagation of Mouse Mammary Tumor Organoids

Tumor organoids mimic the architecture and heterogeneity of in vivo tumors and enable studies of collective interactions between tumor cells as well as with their surrounding microenvironment. Although tumor organoids hold significant promise as cancer models, they are also more costly and labor-intensive to cultivate than traditional 2D cell culture. We sought to identify critical factors regulating organoid growth ex vivo, and to use these observations to develop a more efficient organoid expansion method. Using time-lapse imaging of mouse mammary tumor organoids in 3D culture, we observed that outgrowth potential varies non-linearly with initial organoid size. Maximal outgrowth occurred in organoids with a starting size between ~10 to 1000 cells. Based on these observations, we developed a suspension culture method that maintains organoids in the ideal size range, enabling expansion from 1 million to over 100 million cells in less than 2 weeks and less than 3 hours of hands-on time. Our method facilitates the rapid, cost-effective expansion of organoids for CRISPR based studies and other assays requiring a large amount of organoid starting material.

Covering all your bases: incorporating intron signal from RNA-seq data

RNA-seq datasets can contain millions of intron reads per library that are typically removed from downstream analysis. Only reads overlapping annotated exons are considered to be informative since mature mRNA is assumed to be the major component sequenced, especially for poly(A) RNA libraries. In this study, we show that intron reads are informative, and through exploratory data analysis of read coverage that intron signal is representative of both pre-mRNAs and intron retention. We demonstrate how intron reads can be utilized in differential expression analysis using our index method where a unique set of differentially expressed genes can be detected using intron counts. In exploring read coverage, we also developed the superintronic software that quickly and robustly calculates user-defined summary statistics for exonic and intronic regions. Across multiple datasets, superintronic enabled us to identify several genes with distinctly retained introns that had similar coverage levels to that of neighbouring exons. The work and ideas presented in this paper is the first of its kind to consider multiple biological sources for intron reads through exploratory data analysis, minimizing bias in discovery and interpretation of results. Our findings open up possibilities for further methods development for intron reads and RNA-seq data in general.

TET2 directs mammary luminal cell differentiation and endocrine response

Epigenetic regulation plays an important role in governing stem cell fate and tumorigenesis. Lost expression of a key DNA demethylation enzyme TET2 is associated with human cancers and has been linked to stem cell traits in vitro; however, whether and how TET2 regulates mammary stem cell fate and mammary tumorigenesis in vivo remains to be determined. Here, using our recently established mammary specific Tet2 deletion mouse model, the data reveals that TET2 plays a pivotal role in mammary gland development and luminal lineage commitment. We show that TET2 and FOXP1 form a chromatin complex that mediates demethylation of ESR1, GATA3, and FOXA1, three key genes that are known to coordinately orchestrate mammary luminal lineage specification and endocrine response, and also are often silenced by DNA methylation in aggressive breast cancers. Furthermore, Tet2 deletion-PyMT breast cancer mouse model exhibits enhanced mammary tumor development with deficient ERα expression that confers tamoxifen resistance in vivo. As a result, this study elucidates a role for TET2 in governing luminal cell differentiation and endocrine response that underlies breast cancer resistance to anti-estrogen treatments.

TET2 loss is associated with human cancers but its role in the mammary gland development and tumorigenesis is unclear. Here, the authors show that TET2–FOXP1 complex mediates demethylation of genes involved in luminal lineage commitment and endocrine response, underlying a role of TET2 loss in endocrine resistant breast cancer.

Tissue-resident ductal macrophages survey the mammary epithelium and facilitate tissue remodelling

Macrophages are diverse immune cells that reside in all tissues. Although macrophages have been implicated in mammary-gland function, their diversity has not been fully addressed. By exploiting high-resolution three-dimensional imaging and flow cytometry, we identified a unique population of tissue-resident ductal macrophages that form a contiguous network between the luminal and basal layers of the epithelial tree throughout postnatal development. Ductal macrophages are long lived and constantly survey the epithelium through dendrite movement, revealed via advanced intravital imaging. Although initially originating from embryonic precursors, ductal macrophages derive from circulating monocytes as they expand during puberty. Moreover, they undergo proliferation in pregnancy to maintain complete coverage of the epithelium in lactation, when they are poised to phagocytose milk-producing cells post-lactation and facilitate remodelling. Interestingly, ductal macrophages strongly resemble mammary tumour macrophages and form a network that pervades the tumour. Thus, the mammary epithelium programs specialized resident macrophages in both physiological and tumorigenic contexts.

Intraclonal Plasticity in Mammary Tumors Revealed through Large-Scale Single-Cell Resolution 3D Imaging

Breast tumors are inherently heterogeneous, but the evolving cellular organization through neoplastic progression is poorly understood. Here we report a rapid, large-scale single-cell resolution 3D imaging protocol based on a one-step clearing agent that allows visualization of normal tissue architecture and entire tumors at cellular resolution. Imaging of multicolor lineage-tracing models of breast cancer targeted to either basal or luminal progenitor cells revealed profound clonal restriction during progression. Expression profiling of clones arising in Pten/Trp53-deficient tumors identified distinct molecular signatures. Strikingly, most clones harbored cells that had undergone an epithelial-to-mesenchymal transition, indicating widespread, inherent plasticity. Hence, an integrative pipeline that combines lineage tracing, 3D imaging, and clonal RNA sequencing technologies offers a comprehensive path for studying mechanisms underlying heterogeneity in whole tumors.

Isolation and Propagation of Mammary Epithelial Stem and Progenitor Cells

Several methods of mammary gland dissociation have been described that utilize a combined strategy of mechanical and enzymatic dissociation to isolate mammary epithelial cells (MECs) from intact tissue (Smalley et al., J Mammary Gland Biol Neoplasia 17:91-97, 2012). Here we detail a robust method that enables the isolation of all major stem and progenitor MEC populations, which has been successfully used to study stem cell behavior when coupled with transplantation and in vitro assays (Shackleton et al., Nature 439:84-88, 2006; Bouras et al., Cell Stem Cell 3:429-441, 2008; Sheridan et al., BMC Cancer 15:221, 2015; Jamieson et al., Development 144:1065-1071, 2017). Furthermore, we outline two prominent methods for culturing MECs for the purposes of ex vivo manipulation or study: 2D feeder layer cultures and 3D Matrigel colony assays. Importantly, all outlined methods retain stem and progenitor cell behaviors and can be used in combination with downstream in vivo, in vitro, or in silico analyses.

In Utero Heat Stress Alters the Offspring Epigenome

Exposure to intrauterine heat stress during late gestation affects offspring performance into adulthood. However, underlying mechanistic links between thermal insult in fetal life and postnatal outcomes are not completely understood. We examined morphology, DNA methylation, and gene expression of liver and mammary gland for bull calves and heifers that were gestated under maternal conditions of heat stress or cooling (i.e. in utero heat stressed vs. in utero cooled calves). Mammary tissue was harvested from dairy heifers during their first lactation and liver from bull calves at birth. The liver of in utero heat stressed bull calves contained more cells and the mammary glands of in utero heat stressed heifers were comprised of smaller alveoli. We identified more than 1,500 CpG sites differently methylated between maternal treatment groups. These CpGs were associated with approximately 400 genes, which play a role in processes, such as development, innate immune defense, cell signaling, and transcription and translation. We also identified over 100 differentially expressed genes in the mammary gland with similar functions. Interestingly, fifty differentially methylated genes were shared by both bull calf liver and heifer mammary gland. Intrauterine heat stress alters the methylation profile of liver and mammary DNA and programs their morphology in postnatal life, which may contribute to the poorer performance of in utero heat stressed calves.

PROX1 is a transcriptional regulator of MMP14

The transcription factor PROX1 is essential for development and cell fate specification. Its function in cancer is context-dependent since PROX1 has been shown to play both oncogenic and tumour suppressive roles. Here, we show that PROX1 suppresses the transcription of MMP14, a metalloprotease involved in angiogenesis and cancer invasion, by binding and suppressing the activity of MMP14 promoter. Prox1 deletion in murine dermal lymphatic vessels in vivo and in human LECs increased MMP14 expression. In a hepatocellular carcinoma cell line expressing high endogenous levels of PROX1, its silencing increased both MMP14 expression and MMP14-dependent invasion in 3D. Moreover, PROX1 ectopic expression reduced the MMP14-dependent 3D invasiveness of breast cancer cells and angiogenic sprouting of blood endothelial cells in conjunction with MMP14 suppression. Our study uncovers a new transcriptional regulatory mechanism of cancer cell invasion and endothelial cell specification.

Transcription factor compensation during mammary gland development in E2F knockout mice

The E2F transcription factors control key elements of development, including mammary gland branching morphogenesis, with several E2Fs playing essential roles. Additional prior data has demonstrated that loss of individual E2Fs can be compensated by other E2F family members, but this has not been tested in a mammary gland developmental context. Here we have explored the role of the E2Fs and their ability to functionally compensate for each other during mammary gland development. Using gene expression from terminal end buds and chromatin immunoprecipitation data for E2F1, E2F2 and E2F3, we noted both overlapping and unique mammary development genes regulated by each of the E2Fs. Based on our computational findings and the fact that E2Fs share a common binding motif, we hypothesized that E2F transcription factors would compensate for each other during mammary development and function. To test this hypothesis, we generated RNA from E2F1^-/-, E2F2^-/- and E2F3^+/- mouse mammary glands. QRT-PCR on mammary glands during pregnancy demonstrated increases in E2F2 and E2F3a in the E2F1^-/- mice and an increase in E2F2 levels in E2F3^+/- mice. During lactation we noted that E2F3b transcript levels were increased in the E2F2^-/- mice. Given that E2Fs have previously been noted to have the most striking effects on development during puberty, we hypothesized that loss of individual E2Fs would be compensated for at that time. Double mutant mice were generated and compared with the single knockouts. Loss of both E2F1 and E2F2 revealed a more striking phenotype than either knockout alone, indicating that E2F2 was compensating for E2F1 loss. Interestingly, while E2F2 was not able to functionally compensate for E2F3^+/- during mammary outgrowth, increased E2F2 expression was observed in E2F3^+/- mammary glands during pregnancy day 14.5 and lactation day 5. Together, these findings illustrate the specificity of E2F family members to compensate during development of the mammary gland.

Frequent overexpression of AMAP1, an Arf6 effector in cell invasion, is characteristic of the MMTV-PyMT rather than the MMTV-Neu human breast cancer model

Background

The small GTPase Arf6 and its downstream effector AMAP1 (also called ASAP1/DDEF1) constitute a signaling pathway promoting cell invasion, in which AMAP1 interacts with several different proteins, including PRKD2, EPB41L5, paxillin, and cortactin. Components of this pathway are often overexpressed in human breast cancer cells, to be correlated with poor prognosis of the patients, whereas overexpression of the Arf6 pathway did not correlate with the four main molecular classes of human breast tumors. In this pathway, receptor tyrosine kinases, including EGFR and Her2, activate Arf6 via GEP100. MMTV-PyMT mice and MMTV-Neu mice are well-established models of human breast cancer, and exhibit the early dissemination and the lung metastasis, by utilizing protein tyrosine phosphorylation for oncogenesis. PyMT-tumors and Neu-tumors are known to have overlapping gene expression profiles, which primarily correspond to the luminal B-type of human mammary tumors, although they differ in the time necessary for tumor onset and metastasis. Given the common usage of protein tyrosine phosphorylation, as well as the frequent use of these animal models for studying breast cancer at the molecular level, we here investigated whether mammary tumors in these mouse models utilize the Arf6-based pathway for invasion.

Methods

Expression levels of Arf6, AMAP1, and GEP100 were analyzed in PyMT-tumors and Neu-tumors by western blotting. Expression of Arf6 and AMAP1 was also analyzed by immunohistochemistry. The involvement of AMAP1 in invasion, and the possible correlation of its high expression levels with cancer mesenchymal properties were also investigated.

Results

We found that PyMT-tumors, but not Neu-tumors, frequently overexpress AMAP1 and use it for invasion, whereas both types of tumors expressed Arf6 and GEP100 at different levels. High levels of the AMAP1 expression among PyMT-tumor cells were frequently correlated with loss of the epithelial marker CK8 and also with expression of the mesenchymal marker vimentin both at the primary sites and at sites of the lung metastases.

Conclusions

PyMT-tumors appear to frequently utilize the Arf6-based invasive machinery, whereas Neu-tumors do not. Our results suggest that MMTV-PyMT mice, rather than MMTV-Neu mice, are useful to study the Arf6-based mammary tumor malignancies, as a representative model of human breast cancer.

Electronic supplementary material

The online version of this article (10.1186/s12964-017-0212-z) contains supplementary material, which is available to authorized users.

Construction of developmental lineage relationships in the mouse mammary gland by single-cell RNA profiling

The mammary epithelium comprises two primary cellular lineages, but the degree of heterogeneity within these compartments and their lineage relationships during development remain an open question. Here we report single-cell RNA profiling of mouse mammary epithelial cells spanning four developmental stages in the post-natal gland. Notably, the epithelium undergoes a large-scale shift in gene expression from a relatively homogeneous basal-like program in pre-puberty to distinct lineage-restricted programs in puberty. Interrogation of single-cell transcriptomes reveals different levels of diversity within the luminal and basal compartments, and identifies an early progenitor subset marked by CD55. Moreover, we uncover a luminal transit population and a rare mixed-lineage cluster amongst basal cells in the adult mammary gland. Together these findings point to a developmental hierarchy in which a basal-like gene expression program prevails in the early post-natal gland prior to the specification of distinct lineage signatures, and the presence of cellular intermediates that may serve as transit or lineage-primed cells.

Easy and efficient ensemble gene set testing with EGSEA

Gene set enrichment analysis is a popular approach for prioritising the biological processes perturbed in genomic datasets. The Bioconductor project hosts over 80 software packages capable of gene set analysis. Most of these packages search for enriched signatures amongst differentially regulated genes to reveal higher level biological themes that may be missed when focusing only on evidence from individual genes. With so many different methods on offer, choosing the best algorithm and visualization approach can be challenging. The EGSEA package solves this problem by combining results from up to 12 prominent gene set testing algorithms to obtain a consensus ranking of biologically relevant results.This workflow demonstrates how EGSEA can extend limma-based differential expression analyses for RNA-seq and microarray data using experiments that profile 3 distinct cell populations important for studying the origins of breast cancer. Following data normalization and set-up of an appropriate linear model for differential expression analysis, EGSEA builds gene signature specific indexes that link a wide range of mouse or human gene set collections obtained from MSigDB, GeneSetDB and KEGG to the gene expression data being investigated. EGSEA is then configured and the ensemble enrichment analysis run, returning an object that can be queried using several S4 methods for ranking gene sets and visualizing results via heatmaps, KEGG pathway views, GO graphs, scatter plots and bar plots. Finally, an HTML report that combines these displays can fast-track the sharing of results with collaborators, and thus expedite downstream biological validation. EGSEA is simple to use and can be easily integrated with existing gene expression analysis pipelines for both human and mouse data.

Combining multiple tools outperforms individual methods in gene set enrichment analyses

Motivation

Gene set enrichment (GSE) analysis allows researchers to efficiently extract biological insight from long lists of differentially expressed genes by interrogating them at a systems level. In recent years, there has been a proliferation of GSE analysis methods and hence it has become increasingly difficult for researchers to select an optimal GSE tool based on their particular dataset. Moreover, the majority of GSE analysis methods do not allow researchers to simultaneously compare gene set level results between multiple experimental conditions.

Results

The ensemble of genes set enrichment analyses (EGSEA) is a method developed for RNA-sequencing data that combines results from twelve algorithms and calculates collective gene set scores to improve the biological relevance of the highest ranked gene sets. EGSEA’s gene set database contains around 25 000 gene sets from sixteen collections. It has multiple visualization capabilities that allow researchers to view gene sets at various levels of granularity. EGSEA has been tested on simulated data and on a number of human and mouse datasets and, based on biologists’ feedback, consistently outperforms the individual tools that have been combined. Our evaluation demonstrates the superiority of the ensemble approach for GSE analysis, and its utility to effectively and efficiently extrapolate biological functions and potential involvement in disease processes from lists of differentially regulated genes.

Availability and Implementation

EGSEA is available as an R package at http://www.bioconductor.org/packages/EGSEA/. The gene sets collections are available in the R package EGSEAdata from http://www.bioconductor.org/packages/EGSEAdata/.

Supplementary information

Supplementary data are available at Bioinformatics online.

RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR

The ability to easily and efficiently analyse RNA-sequencing data is a key strength of the Bioconductor project. Starting with counts summarised at the gene-level, a typical analysis involves pre-processing, exploratory data analysis, differential expression testing and pathway analysis with the results obtained informing future experiments and validation studies. In this workflow article, we analyse RNA-sequencing data from the mouse mammary gland, demonstrating use of the popular edgeR package to import, organise, filter and normalise the data, followed by the limma package with its voom method, linear modelling and empirical Bayes moderation to assess differential expression and perform gene set testing. This pipeline is further enhanced by the Glimma package which enables interactive exploration of the results so that individual samples and genes can be examined by the user. The complete analysis offered by these three packages highlights the ease with which researchers can turn the raw counts from an RNA-sequencing experiment into biological insights using Bioconductor.

RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR

The ability to easily and efficiently analyse RNA-sequencing data is a key strength of the Bioconductor project. Starting with counts summarised at the gene-level, a typical analysis involves pre-processing, exploratory data analysis, differential expression testing and pathway analysis with the results obtained informing future experiments and validation studies. In this workflow article, we analyse RNA-sequencing data from the mouse mammary gland, demonstrating use of the popular edgeR package to import, organise, filter and normalise the data, followed by the limma package with its voom method, linear modelling and empirical Bayes moderation to assess differential expression and perform gene set testing. This pipeline is further enhanced by the Glimma package which enables interactive exploration of the results so that individual samples and genes can be examined by the user. The complete analysis offered by these three packages highlights the ease with which researchers can turn the raw counts from an RNA-sequencing experiment into biological insights using Bioconductor.

RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR

The ability to easily and efficiently analyse RNA-sequencing data is a key strength of the Bioconductor project. Starting with counts summarised at the gene-level, a typical analysis involves pre-processing, exploratory data analysis, differential expression testing and pathway analysis with the results obtained informing future experiments and validation studies. In this workflow article, we analyse RNA-sequencing data from the mouse mammary gland, demonstrating use of the popular edgeR package to import, organise, filter and normalise the data, followed by the limma package with its voom method, linear modelling and empirical Bayes moderation to assess differential expression and perform gene set testing. This pipeline is further enhanced by the Glimma package which enables interactive exploration of the results so that individual samples and genes can be examined by the user. The complete analysis offered by these three packages highlights the ease with which researchers can turn the raw counts from an RNA-sequencing experiment into biological insights using Bioconductor.