The Human Dendritic Cell Atlas: An Integrated Transcriptional Tool to Study Human Dendritic Cell Biology.

Dendritic cells (DCs) are functionally diverse and are present in most adult tissues, but deep understanding of human DC biology is hampered by relatively small numbers of these in circulation and their short lifespan in human tissues. We built a transcriptional atlas of human DCs by combining samples from 14 expression profiling studies derived from 10 laboratories. We identified significant gene expression variation of DC subset-defining markers across tissue type and upon viral or bacterial stimulation. We further highlight critical gaps between in vitro-derived DC subsets and their in vivo counterparts and provide evidence that monocytes or cord blood progenitor in vitro-differentiated DCs fail to capture the repertoire of primary DC subsets or behaviors. In constructing a reference DC atlas, we provide an important resource for the community wishing to identify and annotate tissue-specific DC subsets from single-cell datasets, or benchmark new in vitro models of DC biology.

Sincast: a computational framework to predict cell identities in single-cell transcriptomes using bulk atlases as references.

Characterizing the molecular identity of a cell is an essential step in single-cell RNA sequencing (scRNA-seq) data analysis. Numerous tools exist for predicting cell identity using single-cell reference atlases. However, many challenges remain, including correcting for inherent batch effects between reference and query data andinsufficient phenotype data from the reference. One solution is to project single-cell data onto established bulk reference atlases to leverage their rich phenotype information. Sincast is a computational framework to query scRNA-seq data by projection onto bulk reference atlases. Prior to projection, single-cell data are transformed to be directly comparable to bulk data, either with pseudo-bulk aggregation or graph-based imputation to address sparse single-cell expression profiles. Sincast avoids batch effect correction, and cell identity is predicted along a continuum to highlight new cell states not found in the reference atlas. In several case study scenarios, we show that Sincast projects single cells into the correct biological niches in the expression space of the bulk reference atlas. We demonstrate the effectiveness of our imputation approach that was specifically developed for querying scRNA-seq data based on bulk reference atlases. We show that Sincast is an efficient and powerful tool for single-cell profiling that will facilitate downstream analysis of scRNA-seq data.

An integrated analysis of human myeloid cells identifies gaps in in vitro models of in vivo biology.

The Stemformatics myeloid atlas is an integrated transcriptome atlas of human macrophages and dendritic cells that systematically compares freshly isolated tissue-resident, cultured, and pluripotent stem cell-derived myeloid cells. Three classes of tissue-resident macrophage were identified: Kupffer cells and microglia; monocyte-associated; and tumor-associated macrophages. Culture had a major impact on all primary cell phenotypes. Pluripotent stem cell-derived macrophages were characterized by atypical expression of collagen and a highly efferocytotic phenotype. Myeloid subsets, and phenotypes associated with derivation, were reproducible across experimental series including data projected from single-cell studies, demonstrating that the atlas provides a robust reference for myeloid phenotypes. Implementation in Stemformatics.org allows users to visualize patterns of sample grouping or gene expression for user-selected conditions and supports temporary upload of your own microarray or RNA sequencing samples, including single-cell data, to benchmark against the atlas.

A simple, scalable approach to building a cross-platform transcriptome atlas.

Gene expression atlases have transformed our understanding of the development, composition and function of human tissues. New technologies promise improved cellular or molecular resolution, and have led to the identification of new cell types, or better defined cell states. But as new technologies emerge, information derived on old platforms becomes obsolete. We demonstrate that it is possible to combine a large number of different profiling experiments summarised from dozens of laboratories and representing hundreds of donors, to create an integrated molecular map of human tissue. As an example, we combine 850 samples from 38 platforms to build an integrated atlas of human blood cells. We achieve robust and unbiased cell type clustering using a variance partitioning method, selecting genes with low platform bias relative to biological variation. Other than an initial rescaling, no other transformation to the primary data is applied through batch correction or renormalisation. Additional data, including single-cell datasets, can be projected for comparison, classification and annotation. The resulting atlas provides a multi-scaled approach to visualise and analyse the relationships between sets of genes and blood cell lineages, including the maturation and activation of leukocytes in vivo and in vitro. In allowing for data integration across hundreds of studies, we address a key reproduciblity challenge which is faced by any new technology. This allows us to draw on the deep phenotypes and functional annotations that accompany traditional profiling methods, and provide important context to the high cellular resolution of single cell profiling. Here, we have implemented the blood atlas in the open access Stemformatics.org platform, drawing on its extensive collection of curated transcriptome data. The method is simple, scalable and amenable for rapid deployment in other biological systems or computational workflows.

HBO1 is required for the maintenance of leukaemia stem cells.

Acute myeloid leukaemia (AML) is a heterogeneous disease characterized by transcriptional dysregulation that results in a block in differentiation and increased malignant self-renewal. Various epigenetic therapies aimed at reversing these hallmarks of AML have progressed into clinical trials, but most show only modest efficacy owing to an inability to effectively eradicate leukaemia stem cells (LSCs)1. Here, to specifically identify novel dependencies in LSCs, we screened a bespoke library of small hairpin RNAs that target chromatin regulators in a unique ex vivo mouse model of LSCs. We identify the MYST acetyltransferase HBO1 (also known as KAT7 or MYST2) and several known members of the HBO1 protein complex as critical regulators of LSC maintenance. Using CRISPR domain screening and quantitative mass spectrometry, we identified the histone acetyltransferase domain of HBO1 as being essential in the acetylation of histone H3 at K14. H3 acetylated at K14 (H3K14ac) facilitates the processivity of RNA polymerase II to maintain the high expression of key genes (including Hoxa9 and Hoxa10) that help to sustain the functional properties of LSCs. To leverage this dependency therapeutically, we developed a highly potent small-molecule inhibitor of HBO1 and demonstrate its mode of activity as a competitive analogue of acetyl-CoA. Inhibition of HBO1 phenocopied our genetic data and showed efficacy in a broad range of human cell lines and primary AML cells from patients. These biological, structural and chemical insights into a therapeutic target in AML will enable the clinical translation of these findings.

Transcriptional Profiling of Stem Cells: Moving from Descriptive to Predictive Paradigms.

Transcriptional profiling is a powerful tool commonly used to benchmark stem cells and their differentiated progeny. As the wealth of stem cell data builds in public repositories, we highlight common data traps, and review approaches to combine and mine this data for new cell classification and cell prediction tools. We touch on future trends for stem cell profiling, such as single-cell profiling, long-read sequencing, and improved methods for measuring molecular modifications on chromatin and RNA that bring new challenges and opportunities for stem cell analysis.

TCF-1 limits the formation of Tc17 cells via repression of the MAF-RORγt axis.

Interleukin (IL)-17-producing CD8+ T (Tc17) cells have emerged as key players in host-microbiota interactions, infection, and cancer. The factors that drive their development, in contrast to interferon (IFN)-γ-producing effector CD8+ T cells, are not clear. Here we demonstrate that the transcription factor TCF-1 (Tcf7) regulates CD8+ T cell fate decisions in double-positive (DP) thymocytes through the sequential suppression of MAF and RORγt, in parallel with TCF-1-driven modulation of chromatin state. Ablation of TCF-1 resulted in enhanced Tc17 cell development and exposed a gene set signature to drive tissue repair and lipid metabolism, which was distinct from other CD8+ T cell subsets. IL-17-producing CD8+ T cells isolated from healthy humans were also distinct from CD8+IL-17- T cells and enriched in pathways driven by MAF and RORγt Overall, our study reveals how TCF-1 exerts central control of T cell differentiation in the thymus by normally repressing Tc17 differentiation and promoting an effector fate outcome.

Stemformatics: visualize and download curated stem cell data.

Stemformatics is an established gene expression data portal containing over 420 public gene expression datasets derived from microarray, RNA sequencing and single cell profiling technologies. Developed for the stem cell community, it has a major focus on pluripotency, tissue stem cells, and staged differentiation. Stemformatics includes curated 'collections' of data relevant to cell reprogramming, as well as hematopoiesis and leukaemia. Rather than simply rehosting datasets as they appear in public repositories, Stemformatics uses a stringent set of quality control metrics and its own pipelines to process handpicked datasets from raw files. This means that about 30% of datasets processed by Stemformatics fail the quality control metrics and never make it to the portal, ensuring that Stemformatics data are of high quality and have been processed in a consistent manner. Stemformatics provides easy-to-use and intuitive tools for biologists to visually explore the data, including interactive gene expression profiles, principal component analysis plots and hierarchical clusters, among others. The addition of tools that facilitate cross-dataset comparisons provides users with snapshots of gene expression in multiple cell and tissues, assisting the identification of cell-type restricted genes, or potential housekeeping genes. Stemformatics is freely available at stemformatics.org.

Haemopedia RNA-seq: a database of gene expression during haematopoiesis in mice and humans.

During haematopoiesis, haematopoietic stem cells differentiate into restricted potential progenitors before maturing into the many lineages required for oxygen transport, wound healing and immune response. We have updated Haemopedia, a database of gene-expression profiles from a broad spectrum of haematopoietic cells, to include RNA-seq gene-expression data from both mice and humans. The Haemopedia RNA-seq data set covers a wide range of lineages and progenitors, with 57 mouse blood cell types (flow sorted populations from healthy mice) and 12 human blood cell types. This data set has been made accessible for exploration and analysis, to researchers and clinicians with limited bioinformatics experience, on our online portal Haemosphere: https://www.haemosphere.org. Haemosphere also includes nine other publicly available high-quality data sets relevant to haematopoiesis. We have added the ability to compare gene expression across data sets and species by curating data sets with shared lineage designations or to view expression gene vs gene, with all plots available for download by the user.

Age-specific biological and molecular profiling distinguishes paediatric from adult acute myeloid leukaemias.

Acute myeloid leukaemia (AML) affects children and adults of all ages. AML remains one of the major causes of death in children with cancer and for children with AML relapse is the most common cause of death. Here, by modelling AML in vivo we demonstrate that AML is discriminated by the age of the cell of origin. Young cells give rise to myeloid, lymphoid or mixed phenotype acute leukaemia, whereas adult cells give rise exclusively to AML, with a shorter latency. Unlike adult, young AML cells do not remodel the bone marrow stroma. Transcriptional analysis distinguishes young AML by the upregulation of immune pathways. Analysis of human paediatric AML samples recapitulates a paediatric immune cell interaction gene signature, highlighting two genes, RGS10 and FAM26F as prognostically significant. This work advances our understanding of paediatric AML biology, and provides murine models that offer the potential for developing paediatric specific therapeutic strategies.

Building online genomics applications using BioPyramid.

Summary: BioPyramid is a Python package, which serves as a scaffold for building an online application for the exploration of gene expression data. It is designed for bioinformaticians wishing to quickly share transformed data and interactive analyses with collaborators. Current R-based tools similarly address the need to quickly share 'omics'-data in an exploratory format, but these are generally small-scale, single-dataset solutions. Biopyramid is written in Python pyramid framework and scalable to address longer-term or more complex projects. It contains a number of components designed to reduce the time and effort in building such an application from scratch, including gene annotation, dataset models and visualization tools. Availability and implementation: Freely available at http://github.com/jarny/biopyramid. Implemented in python and javascript.

Haemopedia: An Expression Atlas of Murine Hematopoietic Cells.

Hematopoiesis is a multistage process involving the differentiation of stem and progenitor cells into distinct mature cell lineages. Here we present Haemopedia, an atlas of murine gene-expression data containing 54 hematopoietic cell types, covering all the mature lineages in hematopoiesis. We include rare cell populations such as eosinophils, mast cells, basophils, and megakaryocytes, and a broad collection of progenitor and stem cells. We show that lineage branching and maturation during hematopoiesis can be reconstructed using the expression patterns of small sets of genes. We also have identified genes with enriched expression in each of the mature blood cell lineages, many of which show conserved lineage-enriched expression in human hematopoiesis. We have created an online web portal called Haemosphere to make analyses of Haemopedia and other blood cell transcriptional datasets easier. This resource provides simple tools to interrogate gene-expression-based relationships between hematopoietic cell types and genes of interest.

Guide: a desktop application for analysing gene expression data.

BACKGROUND: Multiplecompeting bioinformatics tools exist for next-generation sequencing data analysis. Many of these tools are available as R/Bioconductor modules, and it can be challenging for the bench biologist without any programming background to quickly analyse genomics data. Here, we present an application that is designed to be simple to use, while leveraging the power of R as the analysis engine behind the scenes. RESULTS: Genome Informatics Data Explorer (Guide) is a desktop application designed for the bench biologist to analyse RNA-seq and microarray gene expression data. It requires a text file of summarised read counts or expression values as input data, and performs differential expression analyses at both the gene and pathway level. It uses well-established R/Bioconductor packages such as limma for its analyses, without requiring the user to have specific knowledge of the underlying R functions. Results are presented in figures or interactive tables which integrate useful data from multiple sources such as gene annotation and orthologue data. Advanced options include the ability to edit R commands to customise the analysis pipeline. CONCLUSIONS: Guide is a desktop application designed to query gene expression data in a user-friendly way while automatically communicating with R. Its customisation options make it possible to use different bioinformatics tools available through R/Bioconductor for its analyses, while keeping the core usage simple. Guide is written in the cross-platform framework of Qt, and is freely available for use from http://guide.wehi.edu.au.

Stemformatics: visualisation and sharing of stem cell gene expression.

Genome-scale technologies are increasingly adopted by the stem cell research community, because of the potential to uncover the molecular events most informative about a stem cell state. These technologies also present enormous challenges around the sharing and visualisation of data derived from different laboratories or under different experimental conditions. Stemformatics is an easy to use, publicly accessible portal that hosts a large collection of exemplar stem cell data. It provides fast visualisation of gene expression across a range of mouse and human datasets, with transparent links back to the original studies. One difficulty in the analysis of stem cell signatures is the paucity of public pathways/gene lists relevant to stem cell or developmental biology. Stemformatics provides a simple mechanism to create, share and analyse gene sets, providing a repository of community-annotated stem cell gene lists that are informative about pathways, lineage commitment, and common technical artefacts. Stemformatics can be accessed at stemformatics.org.