Statistically principled feature selection for single cell transcriptomics.

The high dimensionality of data in single cell transcriptomics (scRNAseq) requires investigators to choose subsets of genes (feature selection) for downstream analysis (e.g., unsupervised cell clustering). The evaluation of different approaches to feature selection is hampered by the fact that, as we show here, the performance of feature selection methods varies greatly with the task being performed. For routine cell type identification, even randomly chosen features can perform well, but for cell type differences that are subtle, both number of features and selection strategy can matter strongly. Here we present a simple feature selection method grounded in an analytical model that, without resorting to arbitrary thresholds or user-defined parameters, allows for interpretable delineation of both how many and which features to choose, facilitating identification of biologically meaningful rare cell types. We compare this method to default methods in scanpy and Seurat, as well as SCTransform, showing how greater accuracy can often be achieved with surprisingly few, well-chosen features.

Leveraging gene correlations in single cell transcriptomic data.

BACKGROUND: Many approaches have been developed to overcome technical noise in single cell RNA-sequencing (scRNAseq). As researchers dig deeper into data-looking for rare cell types, subtleties of cell states, and details of gene regulatory networks-there is a growing need for algorithms with controllable accuracy and fewer ad hoc parameters and thresholds. Impeding this goal is the fact that an appropriate null distribution for scRNAseq cannot simply be extracted from data in which ground truth about biological variation is unknown (i.e., usually). RESULTS: We approach this problem analytically, assuming that scRNAseq data reflect only cell heterogeneity (what we seek to characterize), transcriptional noise (temporal fluctuations randomly distributed across cells), and sampling error (i.e., Poisson noise). We analyze scRNAseq data without normalization-a step that skews distributions, particularly for sparse data-and calculate p values associated with key statistics. We develop an improved method for selecting features for cell clustering and identifying gene-gene correlations, both positive and negative. Using simulated data, we show that this method, which we call BigSur (Basic Informatics and Gene Statistics from Unnormalized Reads), captures even weak yet significant correlation structures in scRNAseq data. Applying BigSur to data from a clonal human melanoma cell line, we identify thousands of correlations that, when clustered without supervision into gene communities, align with known cellular components and biological processes, and highlight potentially novel cell biological relationships. CONCLUSIONS: New insights into functionally relevant gene regulatory networks can be obtained using a statistically grounded approach to the identification of gene-gene correlations.

Cohesin prevents cross-domain gene coactivation.

The contrast between the disruption of genome topology after cohesin loss and the lack of downstream gene expression changes instigates intense debates regarding the structure-function relationship between genome and gene regulation. Here, by analyzing transcriptome and chromatin accessibility at the single-cell level, we discover that, instead of dictating population-wide gene expression levels, cohesin supplies a general function to neutralize stochastic coexpression tendencies of cis-linked genes in single cells. Notably, cohesin loss induces widespread gene coactivation and chromatin co-opening tens of million bases apart in cis. Spatial genome and protein imaging reveals that cohesin prevents gene co-bursting along the chromosome and blocks spatial mixing of transcriptional hubs. Single-molecule imaging shows that cohesin confines the exploration of diverse enhancer and core promoter binding transcriptional regulators. Together, these results support that cohesin arranges nuclear topology to control gene coexpression in single cells.

Parsing patterns: Emerging roles of tissue self-organization in health and disease.

Patterned morphologies, such as segments, spirals, stripes, and spots, frequently emerge during embryogenesis through self-organized coordination between cells. Yet, complex patterns also emerge in adults, suggesting that the capacity for spontaneous self-organization is a ubiquitous property of biological tissues. We review current knowledge on the principles and mechanisms of self-organized patterning in embryonic tissues and explore how these principles and mechanisms apply to adult tissues that exhibit features of patterning. We discuss how and why spontaneous pattern generation is integral to homeostasis and healing of tissues, illustrating it with examples from regenerative biology. We examine how aberrant self-organization underlies diverse pathological states, including inflammatory skin disorders and tumors. Lastly, we posit that based on such blueprints, targeted engineering of pattern-driving molecular circuits can be leveraged for synthetic biology and the generation of organoids with intricate patterns.

Gastrulation-stage gene expression in Nipbl+/- mouse embryos foreshadows the development of syndromic birth defects.

In animal models, Nipbl deficiency phenocopies gene expression changes and birth defects seen in Cornelia de Lange syndrome, the most common cause of which is Nipbl haploinsufficiency. Previous studies in Nipbl+/- mice suggested that heart development is abnormal as soon as cardiogenic tissue is formed. To investigate this, we performed single-cell RNA sequencing on wild-type and Nipbl+/- mouse embryos at gastrulation and early cardiac crescent stages. Nipbl+/- embryos had fewer mesoderm cells than wild-type and altered proportions of mesodermal cell subpopulations. These findings were associated with underexpression of genes implicated in driving specific mesodermal lineages. In addition, Nanog was found to be overexpressed in all germ layers, and many gene expression changes observed in Nipbl+/- embryos could be attributed to Nanog overexpression. These findings establish a link between Nipbl deficiency, Nanog overexpression, and gene expression dysregulation/lineage misallocation, which ultimately manifest as birth defects in Nipbl+/- animals and Cornelia de Lange syndrome.

The inherent fragility of collective proliferative control.

Tissues achieve and maintain their sizes through active feedback, whereby cells collectively regulate proliferation and differentiation so as to facilitate homeostasis and the ability to respond to disturbances. One of the best understood feedback mechanisms-renewal control-achieves remarkable feats of robustness in determining and maintaining desired sizes. Yet in a variety of biologically relevant situations, we show that stochastic effects should cause rare but catastrophic failures of renewal control. We define the circumstances under which this occurs and raise the possibility such events account for important non-genetic steps in the development of cancer. We further suggest that the spontaneous stochastic reversal of these events could explain cases of cancer normalization or dormancy following treatment. Indeed, we show that the kinetics of post-treatment recurrence for many cancers are often better fit by a model of stochastic re-emergence due to loss of collective proliferative control, than by deterministic models of cancer relapse.

Gastrulation-stage gene expression in Nipbl+/- mouse embryos foreshadows the development of syndromic birth defects.

In animal models, Nipbl -deficiency phenocopies gene expression changes and birth defects seen in Cornelia de Lange Syndrome (CdLS), the most common cause of which is Nipbl -haploinsufficiency. Previous studies in Nipbl +/- mice suggested that heart development is abnormal as soon as cardiogenic tissue is formed. To investigate this, we performed single-cell RNA-sequencing on wildtype (WT) and Nipbl +/- mouse embryos at gastrulation and early cardiac crescent stages. Nipbl +/- embryos had fewer mesoderm cells than WT and altered proportions of mesodermal cell subpopulations. These findings were associated with underexpression of genes implicated in driving specific mesodermal lineages. In addition, Nanog was found to be overexpressed in all germ layers, and many gene expression changes observed in Nipbl +/- embryos could be attributed to Nanog overexpression. These findings establish a link between Nipbl -deficiency, Nanog overexpression, and gene expression dysregulation/lineage misallocation, which ultimately manifest as birth defects in Nipbl +/- animals and CdLS. Teaser: Gene expression changes during gastrulation of Nipbl -deficient mice shed light on early origins of structural birth defects.

Uncovering Minimal Pathways in Melanoma Initiation.

Cutaneous melanomas are clinically and histologically heterogeneous. Most display activating mutations in Braf or Nras and complete loss of function of one or more tumor suppressor genes. Mouse models that replicate such mutations produce fast-growing, pigmented tumors. However, mice that combine Braf activation with only heterozygous loss of Pten also produce tumors and, as we show here, in an Albino background this occurs even with Braf activation alone. Such tumors arise rarely, grow slowly, and express low levels of pigmentation genes. The timing of their appearance was consistent with a single step stochastic event, but no evidence could be found that it required de novo mutation, suggesting instead the involvement of an epigenetic transition. Single-cell transcriptomic analysis revealed such tumors to be heterogeneous, including a minor cell type we term LNM ( L ow-pigment, N eural- and extracellular M atrix-signature) that displays gene expression resembling "neural crest"-like cell subsets detected in the fast-growing tumors of more heavily-mutated mice, as well as in human biopsy and xenograft samples. We provide evidence that LNM cells pre-exist in normal skin, are expanded by Braf activation, can transition into malignant cells, and persist with malignant cells through multiple rounds of transplantation. We discuss the possibility that LNM cells not only serve as a pre-malignant state in the production of some melanomas, but also as an important intermediate in the development of drug resistance.

Quantifying the relationship between cell proliferation and morphology during development of the face.

Morphogenesis requires highly coordinated, complex interactions between cellular processes: proliferation, migration, and apoptosis, along with physical tissue interactions. How these cellular and tissue dynamics drive morphogenesis remains elusive. Three dimensional (3D) microscopic imaging poses great promise, and generates elegant images. However, generating even moderate through-put quantified images is challenging for many reasons. As a result, the association between morphogenesis and cellular processes in 3D developing tissues has not been fully explored. To address this critical gap, we have developed an imaging and image analysis pipeline to enable 3D quantification of cellular dynamics along with 3D morphology for the same individual embryo. Specifically, we focus on how 3D distribution of proliferation relates to morphogenesis during mouse facial development. Our method involves imaging with light-sheet microscopy, automated segmentation of cells and tissues using machine learning-based tools, and quantification of external morphology via geometric morphometrics. Applying this framework, we show that changes in proliferation are tightly correlated to changes in morphology over the course of facial morphogenesis. These analyses illustrate the potential of this pipeline to investigate mechanistic relationships between cellular dynamics and morphogenesis during embryonic development.

Predictive nonlinear modeling of malignant myelopoiesis and tyrosine kinase inhibitor therapy.

Chronic myeloid leukemia (CML) is a blood cancer characterized by dysregulated production of maturing myeloid cells driven by the product of the Philadelphia chromosome, the BCR-ABL1 tyrosine kinase. Tyrosine kinase inhibitors (TKIs) have proved effective in treating CML, but there is still a cohort of patients who do not respond to TKI therapy even in the absence of mutations in the BCR-ABL1 kinase domain that mediate drug resistance. To discover novel strategies to improve TKI therapy in CML, we developed a nonlinear mathematical model of CML hematopoiesis that incorporates feedback control and lineage branching. Cell-cell interactions were constrained using an automated model selection method together with previous observations and new in vivo data from a chimeric BCR-ABL1 transgenic mouse model of CML. The resulting quantitative model captures the dynamics of normal and CML cells at various stages of the disease and exhibits variable responses to TKI treatment, consistent with those of CML patients. The model predicts that an increase in the proportion of CML stem cells in the bone marrow would decrease the tendency of the disease to respond to TKI therapy, in concordance with clinical data and confirmed experimentally in mice. The model further suggests that, under our assumed similarities between normal and leukemic cells, a key predictor of refractory response to TKI treatment is an increased maximum probability of self-renewal of normal hematopoietic stem cells. We use these insights to develop a clinical prognostic criterion to predict the efficacy of TKI treatment and design strategies to improve treatment response. The model predicts that stimulating the differentiation of leukemic stem cells while applying TKI therapy can significantly improve treatment outcomes.

Modeling collective cell behavior in cancer: Perspectives from an interdisciplinary conversation.

Collective cell behavior contributes to all stages of cancer progression. Understanding how collective behavior emerges through cell-cell interactions and decision-making will advance our understanding of cancer biology and provide new therapeutic approaches. Here, we summarize an interdisciplinary discussion on multicellular behavior in cancer, draw lessons from other scientific disciplines, and identify future directions.

Leveraging gene correlations in single cell transcriptomic data.

BACKGROUND: Many approaches have been developed to overcome technical noise in single cell RNA-sequencing (scRNAseq). As researchers dig deeper into data-looking for rare cell types, subtleties of cell states, and details of gene regulatory networks-there is a growing need for algorithms with controllable accuracy and fewer ad hoc parameters and thresholds. Impeding this goal is the fact that an appropriate null distribution for scRNAseq cannot simply be extracted from data when ground truth about biological variation is unknown (i.e., usually). RESULTS: We approach this problem analytically, assuming that scRNAseq data reflect only cell heterogeneity (what we seek to characterize), transcriptional noise (temporal fluctuations randomly distributed across cells), and sampling error (i.e., Poisson noise). We analyze scRNAseq data without normalization-a step that skews distributions, particularly for sparse data-and calculate p-values associated with key statistics. We develop an improved method for selecting features for cell clustering and identifying gene-gene correlations, both positive and negative. Using simulated data, we show that this method, which we call BigSur (Basic Informatics and Gene Statistics from Unnormalized Reads), captures even weak yet significant correlation structures in scRNAseq data. Applying BigSur to data from a clonal human melanoma cell line, we identify thousands of correlations that, when clustered without supervision into gene communities, align with known cellular components and biological processes, and highlight potentially novel cell biological relationships. CONCLUSIONS: New insights into functionally relevant gene regulatory networks can be obtained using a statistically grounded approach to the identification of gene-gene correlations.

Decreasing Wapl dosage partially corrects embryonic growth and brain transcriptome phenotypes in Nipbl+/- embryos.

Cohesin rings interact with DNA and modulate the expression of thousands of genes. NIPBL loads cohesin onto chromosomes, and WAPL takes it off. Haploinsufficiency for NIPBL causes a developmental disorder, Cornelia de Lange syndrome (CdLS), that is modeled by Nipbl+/- mice. Mutations in WAPL have not been shown to cause disease or gene expression changes in mammals. Here, we show dysregulation of >1000 genes in WaplΔ/+ embryonic mouse brain. The patterns of dysregulation are highly similar in Wapl and Nipbl heterozygotes, suggesting that Wapl mutations may also cause human disease. Since WAPL and NIPBL have opposite effects on cohesin's association with DNA, we asked whether decreasing Wapl dosage could correct phenotypes seen in Nipbl+/- mice. Gene expression and embryonic growth are partially corrected, but perinatal lethality is not. Our data are consistent with the view that cohesin dynamics play a key role in regulating gene expression.

MusMorph, a database of standardized mouse morphology data for morphometric meta-analyses.

Complex morphological traits are the product of many genes with transient or lasting developmental effects that interact in anatomical context. Mouse models are a key resource for disentangling such effects, because they offer myriad tools for manipulating the genome in a controlled environment. Unfortunately, phenotypic data are often obtained using laboratory-specific protocols, resulting in self-contained datasets that are difficult to relate to one another for larger scale analyses. To enable meta-analyses of morphological variation, particularly in the craniofacial complex and brain, we created MusMorph, a database of standardized mouse morphology data spanning numerous genotypes and developmental stages, including E10.5, E11.5, E14.5, E15.5, E18.5, and adulthood. To standardize data collection, we implemented an atlas-based phenotyping pipeline that combines techniques from image registration, deep learning, and morphometrics. Alongside stage-specific atlases, we provide aligned micro-computed tomography images, dense anatomical landmarks, and segmentations (if available) for each specimen (N = 10,056). Our workflow is open-source to encourage transparency and reproducible data collection. The MusMorph data and scripts are available on FaceBase ( www.facebase.org , https://doi.org/10.25550/3-HXMC ) and GitHub ( https://github.com/jaydevine/MusMorph ).

NTR 2.0: a rationally engineered prodrug-converting enzyme with substantially enhanced efficacy for targeted cell ablation.

Transgenic expression of bacterial nitroreductase (NTR) enzymes sensitizes eukaryotic cells to prodrugs such as metronidazole (MTZ), enabling selective cell-ablation paradigms that have expanded studies of cell function and regeneration in vertebrates. However, first-generation NTRs required confoundingly toxic prodrug treatments to achieve effective cell ablation, and some cell types have proven resistant. Here we used rational engineering and cross-species screening to develop an NTR variant, NTR 2.0, which exhibits ~100-fold improvement in MTZ-mediated cell-specific ablation efficacy, eliminating the need for near-toxic prodrug treatment regimens. NTR 2.0 therefore enables sustained cell-loss paradigms and ablation of previously resistant cell types. These properties permit enhanced interrogations of cell function, extended challenges to the regenerative capacities of discrete stem cell niches, and novel modeling of chronic degenerative diseases. Accordingly, we have created a series of bipartite transgenic reporter/effector resources to facilitate dissemination of NTR 2.0 to the research community.

When oncogenes do not cause cancer.

Environmental cues, not oncogene-induced senescence, may stop melanocytes with an activating mutation in the BRAF gene from turning into melanoma.

Optimal experimental design for mathematical models of haematopoiesis.

The haematopoietic system has a highly regulated and complex structure in which cells are organized to successfully create and maintain new blood cells. It is known that feedback regulation is crucial to tightly control this system, but the specific mechanisms by which control is exerted are not completely understood. In this work, we aim to uncover the underlying mechanisms in haematopoiesis by conducting perturbation experiments, where animal subjects are exposed to an external agent in order to observe the system response and evolution. We have developed a novel Bayesian hierarchical framework for optimal design of perturbation experiments and proper analysis of the data collected. We use a deterministic model that accounts for feedback and feedforward regulation on cell division rates and self-renewal probabilities. A significant obstacle is that the experimental data are not longitudinal, rather each data point corresponds to a different animal. We overcome this difficulty by modelling the unobserved cellular levels as latent variables. We then use principles of Bayesian experimental design to optimally distribute time points at which the haematopoietic cells are quantified. We evaluate our approach using synthetic and real experimental data and show that an optimal design can lead to better estimates of model parameters.

Dynamics of nevus development implicate cell cooperation in the growth arrest of transformed melanocytes.

Mutational activation of the BRAF proto-oncogene in melanocytes reliably produces benign nevi (pigmented 'moles'), yet the same change is the most common driver mutation in melanoma. The reason nevi stop growing, and do not progress to melanoma, is widely attributed to a cell-autonomous process of 'oncogene-induced senescence'. Using a mouse model of Braf-driven nevus formation, analyzing both proliferative dynamics and single-cell gene expression, we found no evidence that nevus cells are senescent, either compared with other skin cells, or other melanocytes. We also found that nevus size distributions could not be fit by any simple cell-autonomous model of growth arrest, yet were easily fit by models based on collective cell behavior, for example in which arresting cells release an arrest-promoting factor. We suggest that nevus growth arrest is more likely related to the cell interactions that mediate size control in normal tissues, than to any cell-autonomous, 'oncogene-induced' program of senescence.

Melanocytes are pigment-producing cells found throughout the skin. Mutations that activate a gene called BRAF cause these cells to divide and produce melanocytic nevi, also known as "moles". These mutations are oncogenic, meaning they can cause cancer. Indeed, BRAF is the most commonly mutated gene in melanoma, a deadly skin cancer that arises from melanocytes. Yet, moles hardly ever progress to melanoma. A proposed explanation for this behavior is that, once activated, BRAF initiates a process called "oncogene-induced senescence" in each melanocyte. This process, likened to premature aging, is thought to be what causes cells in a mole to quit dividing. Although this hypothesis is widely accepted, it has proved difficult to test directly. To investigate this notion, Ruiz-Vega et al. studied mice with hundreds of moles created by the same BRAF mutation found in human moles. Analyzing the activity of genes in individual cells revealed that nevus melanocytes that have stopped growing are no more senescent than other skin cells, including non-mole melanocytes. Ruiz-Vega et al. then analyzed the sizes at which moles stopped growing, estimating the number of cells in each mole. The data were then compared with the results of a simulation and mathematical modeling. This revealed that any model based on the idea of cells independently shutting down after a number of random events could not reproduce the distribution of mole sizes that had been experimentally observed. On the other hand, models based on melanocytes acting collectively to shut down each other's growth fit the observed data much better. These findings suggest that moles do not stop growing as a direct result of the activation of BRAF, but because they sense and respond to their own overgrowth. The same kind of collective sensing is observed in normal tissues that maintain a constant size. Discovering that melanocytes do this not only sheds light on why moles stop growing, it could also help researchers devise new ways to prevent melanomas from forming.

Scaling a Dpp Morphogen Gradient through Feedback Control of Receptors and Co-receptors.

Gradients of decapentaplegic (Dpp) pattern Drosophila wing imaginal discs, establishing gene expression boundaries at specific locations. As discs grow, Dpp gradients expand, keeping relative boundary positions approximately stationary. Such scaling fails in mutants for Pentagone (pent), a gene repressed by Dpp that encodes a diffusible protein that expands Dpp gradients. Although these properties fit a recent mathematical model of automatic gradient scaling, that model requires an expander that spreads with minimal loss throughout a morphogen field. Here, we show that Pent's actions are confined to within just a few cell diameters of its site of synthesis and can be phenocopied by manipulating non-diffusible Pent targets strictly within the Pent expression domain. Using genetics and mathematical modeling, we develop an alternative model of scaling driven by feedback downregulation of Dpp receptors and co-receptors. Among the model's predictions is a size beyond which scaling fails-something we observe directly in wing discs.