Ultra-fast scalable estimation of single-cell differentiation potency from scRNA-Seq data

Profiling cell identity and tissue architecture with single-cell and spatial transcriptomics

Single-cell transcriptomics has broadened our understanding of cellular diversity and gene expression dynamics in healthy and diseased tissues. Recently, spatial transcriptomics has emerged as a tool to contextualize single cells in multicellular neighbourhoods and to identify spatially recurrent phenotypes, or ecotypes. These technologies have generated vast datasets with targeted-transcriptome and whole-transcriptome profiles of hundreds to millions of cells. Such data have provided new insights into developmental hierarchies, cellular plasticity and diverse tissue microenvironments, and spurred a burst of innovation in computational methods for single-cell analysis. In this Review, we discuss recent advancements, ongoing challenges and prospects in identifying and characterizing cell states and multicellular neighbourhoods. We discuss recent progress in sample processing, data integration, identification of subtle cell states, trajectory modelling, deconvolution and spatial analysis. Furthermore, we discuss the increasing application of deep learning, including foundation models, in analysing single-cell and spatial transcriptomics data. Finally, we discuss recent applications of these tools in the fields of stem cell biology, immunology, and tumour biology, and the future of single-cell and spatial transcriptomics in biological research and its translation to the clinic.

Cell Fate Dynamics Reconstruction Identifies TPT1 and PTPRZ1 Feedback Loops as Master Regulators of Differentiation in Pediatric Glioblastoma-Immune Cell Networks

Pediatric glioblastoma is a complex dynamical disease that is difficult to treat due to its multiple adaptive behaviors driven largely by phenotypic plasticity. Integrated data science and network theory pipelines offer novel approaches to studying glioblastoma cell fate dynamics, particularly phenotypic transitions over time. Here we used various single-cell trajectory inference algorithms to infer signaling dynamics regulating pediatric glioblastoma-immune cell networks. We identified GATA2, PTPRZ1, TPT1, MTRNR2L1/2, OLIG1/2, SOX11, FXYD6, SEZ6L, PDGFRA, EGFR, S100B, WNT, TNF α , and NF-kB as critical transition genes or signals regulating glioblastoma-immune network dynamics, revealing potential clinically relevant targets. Further, we reconstructed glioblastoma cell fate attractors and found complex bifurcation dynamics within glioblastoma phenotypic transitions, suggesting that a causal pattern may be driving glioblastoma evolution and cell fate decision-making. Together, our findings have implications for developing targeted therapies against glioblastoma, and the continued integration of quantitative approaches and artificial intelligence (AI) to understand pediatric glioblastoma tumor-immune interactions.

Targeting axonal guidance dependencies in glioblastoma with ROBO1 CAR T cells

Resistance to genotoxic therapies and tumor recurrence are hallmarks of glioblastoma (GBM), an aggressive brain tumor. In this study, we investigated functional drivers of post-treatment recurrent GBM through integrative genomic analyses, genome-wide genetic perturbation screens in patient-derived GBM models and independent lines of validation. Specific genetic dependencies were found consistent across recurrent tumor models, accompanied by increased mutational burden and differential transcript and protein expression compared to its primary GBM predecessor. Our observations suggest a multi-layered genetic response to drive tumor recurrence and implicate PTP4A2 (protein tyrosine phosphatase 4A2) as a modulator of self-renewal, proliferation and tumorigenicity in recurrent GBM. Genetic perturbation or small-molecule inhibition of PTP4A2 acts through a dephosphorylation axis with roundabout guidance receptor 1 (ROBO1) and its downstream molecular players, exploiting a functional dependency on ROBO signaling. Because a pan-PTP4A inhibitor was limited by poor penetrance across the blood-brain barrier in vivo, we engineered a second-generation chimeric antigen receptor (CAR) T cell therapy against ROBO1, a cell surface receptor enriched across recurrent GBM specimens. A single dose of ROBO1-targeted CAR T cells doubled median survival in cell-line-derived xenograft (CDX) models of recurrent GBM. Moreover, in CDX models of adult lung-to-brain metastases and pediatric relapsed medulloblastoma, ROBO1 CAR T cells eradicated tumors in 50-100% of mice. Our study identifies a promising multi-targetable PTP4A-ROBO1 signaling axis that drives tumorigenicity in recurrent GBM, with potential in other malignant brain tumors.

Cell cycle expression heterogeneity predicts degree of differentiation

Methods that predict fate potential or degree of differentiation from transcriptomic data have identified rare progenitor populations and uncovered developmental regulatory mechanisms. However, some state-of-the-art methods are too computationally burdensome for emerging large-scale data and all methods make inaccurate predictions in certain biological systems. We developed a method in R (stemFinder) that predicts single cell differentiation time based on heterogeneity in cell cycle gene expression. Our method is computationally tractable and is as good as or superior to competitors. As part of our benchmarking, we implemented four different performance metrics to assist potential users in selecting the tool that is most apt for their application. Finally, we explore the relationship between differentiation time and cell fate potential by analyzing a lineage tracing dataset with clonally labelled hematopoietic cells, revealing that metrics of differentiation time are correlated with the number of downstream lineages.

From single-cell to spatial transcriptomics: decoding the glioma stem cell niche and its clinical implications

Background

Gliomas are aggressive brain tumors associated with a poor prognosis. Cancer stem cells (CSCs) play a significant role in tumor recurrence and resistance to therapy. This study aimed to identify and characterize glioma stem cells (GSCs), analyze their interactions with various cell types, and develop a prognostic signature.

Methods

Single-cell RNA sequencing data from 44 primary glioma samples were analyzed to identify GSC populations. Spatial transcriptomics and gene regulatory network analyses were performed to investigate GSC localization and transcription factor activity. CellChat analysis was conducted to infer cell-cell communication patterns. A GSC signature (GSCS) was developed using machine learning algorithms applied to bulk RNA sequencing data from multiple cohorts. In vitro and in vivo experiments were conducted to validate the role of TUBA1C, a key gene within the signature.

Results

A distinct GSC population was identified, characterized by high proliferative potential and an enrichment of E2F1, E2F2, E2F7, and BRCA1 regulons. GSCs exhibited spatial proximity to myeloid-derived suppressor cells (MDSCs). CellChat analysis revealed an active MIF signaling pathway between GSCs and MDSCs. A 26-gene GSCS demonstrated superior performance compared to existing prognostic models. Knockdown of TUBA1C significantly inhibited glioma cell migration, and invasion in vitro, and reduced tumor growth in vivo.

Conclusion

This study offers a comprehensive characterization of GSCs and their interactions with MDSCs, while presenting a robust GSCS. The findings offer new insights into glioma biology and identify potential therapeutic targets, particularly TUBA1C, aimed at improving patient outcomes.

Cell cycle expression heterogeneity predicts degree of differentiation

Methods that predict fate potential or degree of differentiation from transcriptomic data have identified rare progenitor populations and uncovered developmental regulatory mechanisms. However, some state-of-the-art methods are too computationally burdensome for emerging large-scale data and all methods make inaccurate predictions in certain biological systems. We developed a method in R (stemFinder) that predicts single cell differentiation time based on heterogeneity in cell cycle gene expression. Our method is computationally tractable and is as good as or superior to competitors. As part of our benchmarking, we implemented four different performance metrics to assist potential users in selecting the tool that is most apt for their application. Finally, we explore the relationship between differentiation time and cell fate potential by analyzing a lineage tracing dataset with clonally labelled hematopoietic cells, revealing that metrics of differentiation time are correlated with the number of downstream lineages.

Computational single-cell methods for predicting cancer risk

Despite recent biotechnological breakthroughs, cancer risk prediction remains a formidable computational and experimental challenge. Addressing it is critical in order to improve prevention, early detection and survival rates. Here, I briefly summarize some key emerging theoretical and computational challenges as well as recent computational advances that promise to help realize the goals of cancer-risk prediction. The focus is on computational strategies based on single-cell data, in particular on bottom-up network modeling approaches that aim to estimate cancer stemness and dedifferentiation at single-cell resolution from a systems-biological perspective. I will describe two promising methods, a tissue and cell-lineage independent one based on the concept of diffusion network entropy, and a tissue and cell-lineage specific one that uses transcription factor regulons. Application of these tools to single-cell and single-nucleus RNA-seq data from stages prior to invasive cancer reveal that they can successfully delineate the heterogeneous inter-cellular cancer-risk landscape, identifying those cells that are more likely to turn cancerous. Bottom-up systems biological modeling of single-cell omic data is a novel computational analysis paradigm that promises to facilitate the development of preventive, early detection and cancer-risk prediction strategies.

SCTC: inference of developmental potential from single-cell transcriptional complexity

Inferring the developmental potential of single cells from scRNA-Seq data and reconstructing the pseudo-temporal path of cell development are fundamental but challenging tasks in single-cell analysis. Although single-cell transcriptional diversity (SCTD) measured by the number of expressed genes per cell has been widely used as a hallmark of developmental potential, it may lead to incorrect estimation of differentiation states in some cases where gene expression does not decrease monotonously during the development process. In this study, we propose a novel metric called single-cell transcriptional complexity (SCTC), which draws on insights from the economic complexity theory and takes into account the sophisticated structure information of scRNA-Seq count matrix. We show that SCTC characterizes developmental potential more accurately than SCTD, especially in the early stages of development where cells typically have lower diversity but higher complexity than those in the later stages. Based on the SCTC, we provide an unsupervised method for accurate, robust, and transferable inference of single-cell pseudotime. Our findings suggest that the complexity emerging from the interplay between cells and genes determines the developmental potential, providing new insights into the understanding of biological development from the perspective of complexity theory.

On epigenetic stochasticity, entropy and cancer risk

Epigenetic changes are known to accrue in normal cells as a result of ageing and cumulative exposure to cancer risk factors. Increasing evidence points towards age-related epigenetic changes being acquired in a quasi-stochastic manner, and that they may play a causal role in cancer development. Here, I describe the quasi-stochastic nature of DNA methylation (DNAm) changes in ageing cells as well as in normal cells at risk of neoplastic transformation, discussing the implications of this stochasticity for developing cancer risk prediction strategies, and in particular, how it may require a conceptual paradigm shift in how we select cancer risk markers. I also describe the mounting evidence that a significant proportion of DNAm changes in ageing and cancer development are related to cell proliferation, reflecting tissue-turnover and the opportunity this offers for predicting cancer risk via the development of epigenetic mitotic-like clocks. Finally, I describe how age-associated DNAm changes may be causally implicated in cancer development via an irreversible suppression of tissue-specific transcription factors that increases epigenetic and transcriptomic entropy, promoting a more plastic yet aberrant cancer stem-cell state. This article is part of a discussion meeting issue 'Causes and consequences of stochastic processes in development and disease'.

Mapping single-cell developmental potential in health and disease with interpretable deep learning

Single-cell RNA sequencing (scRNA-seq) has transformed our understanding of cell fate in developmental systems. However, identifying the molecular hallmarks of potency - the capacity of a cell to differentiate into other cell types - has remained challenging. Here, we introduce CytoTRACE 2, an interpretable deep learning framework for characterizing potency and differentiation states on an absolute scale from scRNA-seq data. Across 31 human and mouse scRNA-seq datasets encompassing 28 tissue types, CytoTRACE 2 outperformed existing methods for recovering experimentally determined potency levels and differentiation states covering the entire range of cellular ontogeny. Moreover, it reconstructed the temporal hierarchy of mouse embryogenesis across 62 timepoints; identified pan-tissue expression programs that discriminate major potency levels; and facilitated discovery of cellular phenotypes in cancer linked to survival and immunotherapy resistance. Our results illuminate a fundamental feature of cell biology and provide a broadly applicable platform for delineating single-cell differentiation landscapes in health and disease.

THBS1+ myeloid cells expand in SLD hepatocellular carcinoma and contribute to immunosuppression and unfavorable prognosis through TREM1

Hepatocellular carcinoma (HCC) is an inflammation-associated cancer arising from viral or non-viral etiologies including steatotic liver diseases (SLDs). Expansion of immunosuppressive myeloid cells is a hallmark of inflammation and cancer, but their heterogeneity in HCC is not fully resolved and might underlie immunotherapy resistance. Here, we present a high-resolution atlas of innate immune cells from patients with HCC that unravels an SLD-associated contexture characterized by influx of inflammatory and immunosuppressive myeloid cells, including a discrete population of THBS1+ regulatory myeloid (Mreg) cells expressing monocyte- and neutrophil-affiliated genes. THBS1+ Mreg cells expand in SLD-associated HCC, populate fibrotic lesions, and are associated with poor prognosis. THBS1+ Mreg cells are CD163+ but distinguished from macrophages by high expression of triggering receptor expressed on myeloid cells 1 (TREM1), which contributes to their immunosuppressive activity and promotes HCC tumor growth in vivo. Our data support myeloid subset-targeted immunotherapies to treat HCC.

Accurate Estimation of Single-Cell Differentiation Potency Based on Network Topology and Gene Ontology Information

One important task in single-cell analysis is to quantify the differentiation potential of single cells. Though various single-cell potency measures have been proposed, they are based on individual biological sources, thus not robust and reliable. It is still a challenge to combine multiple sources to generate a relatively reliable and robust measure to estimate differentiation. In this paper, we propose a New Centrality measure with Gene ontology information (NCG) to estimate single-cell potency. NCG is designed by combining network topology property with edge clustering coefficient, and gene function information using gene ontology function similarity scores. NCG distinguishes pluripotent cells from non-pluripotent cells with high accuracy, correctly ranks different cell types by their differentiation potency, tracks changes during the differentiation process, and constructs the lineage trajectory from human myoblasts into skeletal muscle cells. These indicate that NCG is a reliable and robust measure to estimate single-cell potency. NCG is anticipated to be a useful tool for identifying novel stem or progenitor cell phenotypes from single-cell RNA-Seq data. The source codes and datasets are available at https://github.com/Xinzhe-Ni/NCG.

FitDevo: accurate inference of single-cell developmental potential using sample-specific gene weight

The quantification of developmental potential is critical for determining developmental stages and identifying essential molecular signatures in single-cell studies. Here, we present FitDevo, a novel method for inferring developmental potential using scRNA-seq data. The main idea of FitDevo is first to generate sample-specific gene weight (SSGW) and then infer developmental potential by calculating the correlation between SSGW and gene expression. SSGW is generated using a generalized linear model that combines sample-specific information and gene weight learned from a training dataset covering scRNA-seq data of 17 previously published datasets. We have rigorously validated FitDevo’s effectiveness using a testing dataset with scRNA-seq data from 28 existing datasets and have also demonstrated its superiority over current methods. Furthermore, FitDevo’s broad application scope has been illustrated using three practical scenarios: deconvolution analysis of epidermis, spatial transcriptomic data analysis of hearts and intestines, and developmental potential analysis of breast cancer. The source code and related data are available at https://github.com/jumphone/fitdevo.

Systematic identification of cell-fate regulatory programs using a single-cell atlas of mouse development

Waddington's epigenetic landscape is a metaphor frequently used to illustrate cell differentiation. Recent advances in single-cell genomics are altering our understanding of the Waddington landscape, yet the molecular mechanisms of cell-fate decisions remain poorly understood. We constructed a cell landscape of mouse lineage differentiation during development at the single-cell level and described both lineage-common and lineage-specific regulatory programs during cell-type maturation. We also found lineage-common regulatory programs that are broadly active during the development of invertebrates and vertebrates. In particular, we identified Xbp1 as an evolutionarily conserved regulator of cell-fate determinations across different species. We demonstrated that Xbp1 transcriptional regulation is important for the stabilization of the gene-regulatory networks for a wide range of mouse cell types. Our results offer genetic and molecular insights into cellular gene-regulatory programs and will serve as a basis for further advancing the understanding of cell-fate decisions.

Computational identification of preneoplastic cells displaying high stemness and risk of cancer progression

Evidence points towards the differentiation state of cells as a marker of cancer risk and progression. Measuring the differentiation state of single cells in a preneoplastic population could thus enable novel strategies for early detection and risk prediction. Recent maps of somatic mutagenesis in normal tissues from young healthy individuals have revealed cancer driver mutations, indicating that these do not correlate well with differentiation state and that other molecular events also contribute to cancer development. We hypothesized that the differentiation state of single cells can be measured by estimating the regulatory activity of the transcription factors (TFs) that control differentiation within that cell lineage. To this end, we present a novel computational method called CancerStemID that estimates a stemness index of cells from single-cell RNA-Seq data. CancerStemID is validated in two human esophageal squamous cell carcinoma (ESCC) cohorts, demonstrating how it can identify undifferentiated preneoplastic cells whose transcriptomic state is overrepresented in invasive cancer. Spatial transcriptomics and whole-genome bisulfite sequencing demonstrated that differentiation activity of tissue-specific TFs was decreased in cancer cells compared to the basal cell-of-origin layer and established that differentiation state correlated with differential DNA methylation at the promoters of these TFs, independently of underlying NOTCH1 and TP53 mutations. The findings were replicated in a mouse model of ESCC development, and the broad applicability of CancerStemID to other cancer-types was demonstrated. In summary, these data support an epigenetic stem-cell model of oncogenesis and highlight a novel computational strategy to identify stem-like preneoplastic cells that undergo positive selection.

scInTime: A Computational Method Leveraging Single-Cell Trajectory and Gene Regulatory Networks to Identify Master Regulators of Cellular Differentiation

Trajectory inference (TI) or pseudotime analysis has dramatically extended the analytical framework of single-cell RNA-seq data, allowing regulatory genes contributing to cell differentiation and those involved in various dynamic cellular processes to be identified. However, most TI analysis procedures deal with individual genes independently while overlooking the regulatory relations between genes. Integrating information from gene regulatory networks (GRNs) at different pseudotime points may lead to more interpretable TI results. To this end, we introduce scInTime-an unsupervised machine learning framework coupling inferred trajectory with single-cell GRNs (scGRNs) to identify master regulatory genes. We validated the performance of our method by analyzing multiple scRNA-seq data sets. In each of the cases, top-ranking genes predicted by scInTime supported their functional relevance with corresponding signaling pathways, in line with the results of available functional studies. Overall results demonstrated that scInTime is a powerful tool to exploit pseudotime-series scGRNs, allowing for a clear interpretation of TI results toward more significant biological insights.

Single-cell RNA Sequencing Reveals How the Aryl Hydrocarbon Receptor Shapes Cellular Differentiation Potency in the Mouse Colon

Despite recent progress recognizing the importance of aryl hydrocarbon receptor (Ahr)-dependent signaling in suppressing colon tumorigenesis, its role in regulating colonic crypt homeostasis remains unclear. To assess the effects of Ahr on intestinal epithelial cell heterogeneity and functional phenotypes, we utilized single-cell transcriptomics and advanced analytic strategies to generate a high-quality atlas for colonic intestinal crypts from wild-type and intestinal-specific Ahr knockout mice. Here we observed the promotive effects of Ahr deletion on Foxm1-regulated genes in crypt-associated canonical epithelial cell types and subtypes of goblet cells and deep crypt-secretory cells. We also show that intestinal Ahr deletion elevated single-cell entropy (a measure of differentiation potency or cell stemness) and RNA velocity length (a measure of the rate of cell differentiation) in noncycling and cycling Lgr5⁺ stem cells. In general, intercellular signaling cross-talk via soluble and membrane-bound factors was perturbed in Ahr-null colonocytes. Taken together, our single-cell RNA sequencing analyses provide new evidence of the molecular function of Ahr in modulating putative stem cell driver genes, cell potency lineage decisions, and cell-cell communication in vivo. PREVENTION RELEVANCE: Our mouse single-cell RNA sequencing analyses provide new evidence of the molecular function of Ahr in modulating colonic stemness and cell-cell communication in vivo. From a cancer prevention perspective, Ahr should be considered a therapeutic target to recalibrate remodeling of the intestinal stem cell niche.

A review of dynamical systems approaches for the detection of chaotic attractors in cancer networks

Cancers are complex dynamical systems. They remain the leading cause of disease-related pediatric mortality in North America. To overcome this burden, we must decipher the state-space attractor dynamics of gene expression patterns and protein oscillations orchestrated by cancer stemness networks. The review provides an overview of dynamical systems theory to steer cancer research in pattern science. While most of our current tools in network medicine rely on statistical correlation methods, causality inference remains primitively developed. As such, a survey of attractor reconstruction methods and machine algorithms for the detection of causal structures applicable in experimentally derived time series cancer datasets is presented. A toolbox of complex systems approaches are discussed for reconstructing the signaling state space of cancer networks, interpreting causal relationships in their time series gene expression patterns, and assisting clinical decision making in computational oncology. As a proof of concept, the applicability of some algorithms are demonstrated on pediatric brain cancer datasets and the requirement of their time series analysis is highlighted.