Visualizing hierarchies in scRNA-seq data using a density tree-biased autoencoder.

MOTIVATION: Single-cell RNA sequencing (scRNA-seq) allows studying the development of cells in unprecedented detail. Given that many cellular differentiation processes are hierarchical, their scRNA-seq data are expected to be approximately tree-shaped in gene expression space. Inference and representation of this tree structure in two dimensions is highly desirable for biological interpretation and exploratory analysis. RESULTS: Our two contributions are an approach for identifying a meaningful tree structure from high-dimensional scRNA-seq data, and a visualization method respecting the tree structure. We extract the tree structure by means of a density-based maximum spanning tree on a vector quantization of the data and show that it captures biological information well. We then introduce density-tree biased autoencoder (DTAE), a tree-biased autoencoder that emphasizes the tree structure of the data in low dimensional space. We compare to other dimension reduction methods and demonstrate the success of our method both qualitatively and quantitatively on real and toy data. AVAILABILITY AND IMPLEMENTATION: Our implementation relying on PyTorch and Higra is available at github.com/hci-unihd/DTAE. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Simulation-based inference of differentiation trajectories from RNA velocity fields.

We report Cytopath, a method for trajectory inference that takes advantage of transcriptional activity information from the RNA velocity of single cells to perform trajectory inference. Cytopath performs this task by defining a Markov chain model, simulating an ensemble of possible differentiation trajectories, and constructing a consensus trajectory. We show that Cytopath can recapitulate the topological and molecular characteristics of the differentiation process under study. In our analysis, we include differentiation trajectories with varying bifurcated, circular, convergent, and mixed topologies studied in single-snapshot as well as time-series single-cell RNA sequencing experiments. We demonstrate the capability to reconstruct differentiation trajectories, assess the association of RNA velocity-based pseudotime with actually elapsed process time, and identify drawbacks in current state-of-the art trajectory inference approaches.

Exhausted CD8+ T cells exhibit low and strongly inhibited TCR signaling during chronic LCMV infection.

Chronic viral infections are often associated with impaired CD8+ T cell function, referred to as exhaustion. Although the molecular and cellular circuits involved in CD8+ T cell exhaustion are well defined, with sustained presence of antigen being one important parameter, how much T cell receptor (TCR) signaling is actually ongoing in vivo during established chronic infection is unclear. Here, we characterize the in vivo TCR signaling of virus-specific exhausted CD8+ T cells in a mouse model, leveraging TCR signaling reporter mice in combination with transcriptomics. In vivo signaling in exhausted cells is low, in contrast to their in vitro signaling potential, and despite antigen being abundantly present. Both checkpoint blockade and adoptive transfer of naïve target cells increase TCR signaling, demonstrating that engagement of co-inhibitory receptors curtails CD8+ T cell signaling and function in vivo.

Landscape of Exhausted Virus-Specific CD8 T Cells in Chronic LCMV Infection.

A hallmark of chronic infections is the presence of exhausted CD8 T cells, characterized by a distinct transcriptional program compared with functional effector or memory cells, co-expression of multiple inhibitory receptors, and impaired effector function, mainly driven by recurrent T cell receptor engagement. In the context of chronic lymphocytic choriomeningitis virus (LCMV) infection in mice, most studies focused on studying splenic virus-specific CD8 T cells. Here, we provide a detailed characterization of exhausted CD8 T cells isolated from six different tissues during established LCMV infection, using single-cell RNA sequencing. Our data reveal that exhausted cells are heterogeneous, adopt organ-specific transcriptomic profiles, and can be divided into five main functional subpopulations: advanced exhaustion, effector-like, intermediate, proliferating, or memory-like. Adoptive transfer experiments showed that these phenotypes are plastic, suggesting that the tissue microenvironment has a major impact in shaping the phenotype and function of virus-specific CD8 T cells during chronic infection.

The type IV pilin PilA couples surface attachment and cell-cycle initiation in Caulobacter crescentus.

Understanding how bacteria colonize surfaces and regulate cell-cycle progression in response to cellular adhesion is of fundamental importance. Here, we use transposon sequencing in conjunction with fluorescence resonance energy transfer (FRET) microscopy to uncover the molecular mechanism for how surface sensing drives cell-cycle initiation in Caulobacter crescentus We identify the type IV pilin protein PilA as the primary signaling input that couples surface contact to cell-cycle initiation via the second messenger cyclic di-GMP (c-di-GMP). Upon retraction of pili filaments, the monomeric pilin reservoir in the inner membrane is sensed by the 17-amino acid transmembrane helix of PilA to activate the PleC-PleD two-component signaling system, increase cellular c-di-GMP levels, and signal the onset of the cell cycle. We termed the PilA signaling sequence CIP for "cell-cycle initiating pilin" peptide. Addition of the chemically synthesized CIP peptide initiates cell-cycle progression and simultaneously inhibits surface attachment. The broad conservation of the type IV pili and their importance in pathogens for host colonization suggests that CIP peptide mimetics offer strategies to inhibit surface sensing, prevent biofilm formation and control persistent infections.

Quantitative Selection Analysis of Bacteriophage φCbK Susceptibility in Caulobacter crescentus.

Classical molecular genetics uses stringent selective conditions to identify mutants with distinct phenotypic responses. Mutations giving rise to less pronounced phenotypes are often missed. However, to gain systems-level insights into complex genetic interaction networks requires genome-wide assignment of quantitative phenotypic traits. In this paper, we present a quantitative selection approach coupled with transposon sequencing (QS-TnSeq) to globally identify the cellular components that orchestrate susceptibility of the cell cycle model bacterium Caulobacter crescentus toward bacteriophage φCbK infection. We found that 135 genes representing 3.30% of the Caulobacter genome exhibit significant accumulation of transposon insertions upon φCbK selection. More than 85% thereof consist of new factors not previously associated with phage φCbK susceptibility. Using hierarchical clustering of dose-dependent TnSeq datasets, we grouped these genes into functional modules that correlate with different stages of the φCbK infection process. We assign φCbK susceptibility to eight new genes that represent novel components of the pilus secretion machinery. Further, we demonstrate that, from 86 motility genes, only seven genes encoding structural and regulatory components of the flagellar hook increase phage resistance when disrupted by transposons, suggesting a link between flagellar hook assembly and pili biogenesis. In addition, we observe high recovery of Tn5 insertions within regulatory sequences of the genes encoding the essential NADH:ubiquinone oxidoreductase complex indicating that intact proton motive force is crucial for effective phage propagation. In sum, QS-TnSeq is broadly applicable to perform quantitative and genome-wide systems-genetics analysis of complex phenotypic traits.