Single-cell Variability Research Articles

Prediction of protein subcellular localization in single cells.

The subcellular localization of a protein is important for its function and interaction with other molecules, and its mislocalization is linked to numerous diseases. While atlas-scale efforts have been made to profile protein localization across various cell lines, existing datasets only contain limited pairs of proteins and cell lines which do not cover all human proteins. We present a method that uses both protein sequences and cellular landmark images to perform Predictions of Unseen Proteins' Subcellular localization (PUPS), which can generalize to both proteins and cell lines not used for model training. PUPS combines a protein language model and an image inpainting model to utilize both protein sequence and cellular images for protein localization prediction. The protein sequence input enables generalization to unseen proteins and the cellular image input enables cell type specific prediction that captures single-cell variability. PUPS' ability to generalize to unseen proteins and cell lines enables us to assess the variability in protein localization across cell lines as well as across single cells within a cell line and to identify the biological processes associated with the proteins that have variable localization. Experimental validation shows that PUPS can be used to predict protein localization in newly performed experiments outside of the Human Protein Atlas used for training. Collectively, PUPS utilizes both protein sequences and cellular images to predict protein localization in unseen proteins and cell lines with the ability to capture single-cell variability.

Allele-level visualization of transcription and chromatin by high-throughput imaging

The spatial arrangement of the genome within the nucleus is a pivotal aspect of cellular organization and function with implications for gene expression and regulation. While all genome organization features, such as loops, domains, and radial positioning, are nonrandom, they are characterized by a high degree of single-cell variability. Imaging approaches are ideally suited to visualize, measure, and study single-cell heterogeneity in genome organization. Here, we describe two methods for the detection of DNA and RNA of individual gene alleles by fluorescence in situ hybridization (FISH) in a high-throughput format. We have optimized combined DNA/RNA FISH approaches either using simultaneous or sequential detection of DNA and nascent RNA. These optimized DNA and RNA FISH protocols were implemented in a 384-well plate format alongside automated image and data analysis and enable accurate detection of individual gene alleles and their gene expression status across a large cell population. We successfully visualized MYC and EGFR DNA and nascent RNA with allele-level resolution in multiple cell types, and we determined the radial position of active and inactive MYC and EGFR alleles. These optimized DNA/RNA detection approaches are versatile and sensitive tools for mapping of chromatin features and gene activity at the single-allele level and at high throughput.

A library of reporters of the global regulators of gene expression in Escherichia coli.

The topology of the transcription factor network (TFN) of Escherichia coli is far from uniform, with 22 global regulator (GR) proteins controlling one-third of all genes. So far, their production rates cannot be tracked by comparable fluorescent proteins. We developed a library of fluorescent reporters for 16 GRs for this purpose. Each consists of a single-copy plasmid coding for green fluorescent protein (GFP) fused to the full-length copy of the native promoter. We tracked their activity in exponential and stationary growth, as well as under weak and strong stresses. We show that the reporters have high sensitivity and specificity to all stresses tested and detect single-cell variability in transcription rates. Given the influence of GRs on the TFN, we expect that the new library will contribute to dissecting global transcriptional stress-response programs of E. coli. Moreover, the library can be invaluable in bioindustrial applications that tune those programs to, instead of cell growth, favor productivity while reducing energy consumption.IMPORTANCECells contain thousands of genes. Many genes are involved in the control of cellular activities. Some activities require a few hundred genes to run largely synchronous transcriptional programs. To achieve this, cells have evolved global regulator (GR) proteins that can influence hundreds of genes simultaneously. We have engineered a library of Escherichia coli strains to track the levels over time of these, phenotypically critical, GRs. Each strain has a single-copy plasmid coding for a fast-maturing green fluorescent protein whose transcription is controlled by a copy of the natural GR promoter. By allowing the tracking of GR levels, with sensitivity and specificity, this library should become of wide use in scientific research on bacterial gene expression (from molecular to synthetic biology) and, later, be used in applications in therapeutics and bioindustries.

Cellograph: a semi-supervised approach to analyzing multi-condition single-cell RNA-sequencing data using graph neural networks

With the growing number of single-cell datasets collected under more complex experimental conditions, there is an opportunity to leverage single-cell variability to reveal deeper insights into how cells respond to perturbations. Many existing approaches rely on discretizing the data into clusters for differential gene expression (DGE), effectively ironing out any information unveiled by the single-cell variability across cell-types. In addition, DGE often assumes a statistical distribution that, if erroneous, can lead to false positive differentially expressed genes. Here, we present Cellograph: a semi-supervised framework that uses graph neural networks to quantify the effects of perturbations at single-cell granularity. Cellograph not only measures how prototypical cells are of each condition but also learns a latent space that is amenable to interpretable data visualization and clustering. The learned gene weight matrix from training reveals pertinent genes driving the differences between conditions. We demonstrate the utility of our approach on publicly-available datasets including cancer drug therapy, stem cell reprogramming, and organoid differentiation. Cellograph outperforms existing methods for quantifying the effects of experimental perturbations and offers a novel framework to analyze single-cell data using deep learning.

Spatiotemporal NF-κB dynamics encodes the position, amplitude, and duration of local immune inputs.

Infected cells communicate through secreted signaling molecules like cytokines, which carry information about pathogens. How differences in cytokine secretion affect inflammatory signaling over space and how responding cells decode information from propagating cytokines are not understood. By computationally and experimentally studying NF-κB dynamics in cocultures of signal-sending cells (macrophages) and signal-receiving cells (fibroblasts), we find that cytokine signals are transmitted by wave-like propagation of NF-κB activity and create well-defined activation zones in responding cells. NF-κB dynamics in responding cells can simultaneously encode information about cytokine dose, duration, and distance to the cytokine source. Spatially resolved transcriptional analysis reveals that responding cells transmit local cytokine information to distance-specific proinflammatory gene expression patterns, creating “gene expression zones.” Despite single-cell variability, the size and duration of the signaling zone are tightly controlled by the macrophage secretion profile. Our results highlight how macrophages tune cytokine secretion to control signal transmission distance and how inflammatory signaling interprets these signals in space and time.

Relationship between fitness and heterogeneity in exponentially growing microbial populations

Despite major environmental and genetic differences, microbial metabolic networks are known to generate consistent physiological outcomes across vastly different organisms. This remarkable robustness suggests that, at least in bacteria, metabolic activity may be guided by universal principles. The constrained optimization of evolutionarily motivated objective functions, such as the growth rate, has emerged as the key theoretical assumption for the study of bacterial metabolism. While conceptually and practically useful in many situations, the idea that certain functions are optimized is hard to validate in data. Moreover, it is not always clear how optimality can be reconciled with the high degree of single-cell variability observed in experiments within microbial populations. To shed light on these issues, we develop an inverse modeling framework that connects the fitness of a population of cells (represented by the mean single-cell growth rate) to the underlying metabolic variability through the maximum entropy inference of the distribution of metabolic phenotypes from data. While no clear objective function emerges, we find that, as the medium gets richer, the fitness and inferred variability for Escherichia coli populations follow and slowly approach the theoretically optimal bound defined by minimal reduction of variability at given fitness. These results suggest that bacterial metabolism may be crucially shaped by a population-level trade-off between growth and heterogeneity.

Machine learning approach for predicting single cell lag time of Salmonella Enteritidis after heat and chlorine treatment

The importance of single-cell variability is increasingly prominent with the developments in foodborne pathogens modeling. Traditional predictive microbiology model cannot accurately describe the growth behavior of small numbers of cells due to individual cell heterogeneity. The objective of the present study was to develop predictive models for single cell lag times of Salmonella Enteritidis after heat and chlorine treatment. A time-lapse microscopy method was employed to evaluate the single cell lag time by monitoring cell divisions. Four supervised machine learning algorithms including gradient boosting regression tree (GBRT), artificial neural network (ANN), random forest (RF), and support vector regression (SVR) were applied and compared. Results show that all four machine learning models have good predictive capabilities without an overfitting of the data. The ANN approach demonstrated superior prediction performance over other machine learning models (RMSE: 0.209, MAE: 0.135 and R2: 0.989). Furthermore, the SHapley Additive exPlanation (SHAP) measures were used to capture the influence of each feature on the model output, and results revealed that population lag times and sublethal injury rate have dominant impacts on the single cell lag time. Consequently, the findings generated from this study may be useful in managing the potential food safety risk caused by single cells of foodborne pathogens.

SmiFISH and embryo segmentation for single-cell multi-gene RNA quantification in arthropods

Recently, advances in fluorescent in-situ hybridization techniques and in imaging technology have enabled visualization and counting of individual RNA molecules in single cells. This has greatly enhanced the resolution in our understanding of transcriptional processes. Here, we adapt a recently published smiFISH protocol (single-molecule inexpensive fluorescent in-situ hybridization) to whole embryos across a range of arthropod model species, and also to non-embryonic tissues. Using multiple fluorophores with distinct spectra and white light laser confocal imaging, we simultaneously detect and separate single RNAs from up to eight different genes in a whole embryo. We also combine smiFISH with cell membrane immunofluorescence, and present an imaging and analysis pipeline for 3D cell segmentation and single-cell RNA counting in whole blastoderm embryos. Finally, using whole embryo single-cell RNA count data, we propose two alternative single-cell variability measures to the commonly used Fano factor, and compare the capacity of these three measures to address different aspects of single-cell expression variability.

Highlights from Recent Cancer Literature

Highlights from Recent Cancer Literature

Insights on the Control of Yeast Single-Cell Growth Variability by Members of the Trehalose Phosphate Synthase (TPS) Complex.

Single-cell variability of growth is a biological phenomenon that has attracted growing interest in recent years. Important progress has been made in the knowledge of the origin of cell-to-cell heterogeneity of growth, especially in microbial cells. To better understand the origins of such heterogeneity at the single-cell level, we developed a new methodological pipeline that coupled cytometry-based cell sorting with automatized microscopy and image analysis to score the growth rate of thousands of single cells. This allowed investigating the influence of the initial amount of proteins of interest on the subsequent growth of the microcolony. As a preliminary step to validate this experimental setup, we referred to previous findings in yeast where the expression level of Tsl1, a member of the Trehalose Phosphate Synthase (TPS) complex, negatively correlated with cell division rate. We unfortunately could not find any influence of the initial TSL1 expression level on the growth rate of the microcolonies. We also analyzed the effect of the natural variations of trehalose-6-phosphate synthase (TPS1) expression on cell-to-cell growth heterogeneity, but we did not find any correlation. However, due to the already known altered growth of the tps1Δ mutants, we tested this strain at the single-cell level on a permissive carbon source. This mutant showed an outstanding lack of reproducibility of growth rate distributions as compared to the wild-type strain, with variable proportions of non-growing cells between cultivations and more heterogeneous microcolonies in terms of individual growth rates. Interestingly, this variable behavior at the single-cell level was reminiscent to the high variability that is also stochastically suffered at the population level when cultivating this tps1Δ strain, even when using controlled bioreactors.

Genetic screening for single-cell variability modulators driving therapy resistance.

Cellular plasticity describes cells’ ability to transition from one set of phenotypes to another. In melanoma, transient fluctuations in the molecular state of tumor cells mark the formation of rare cells primed to survive BRAF inhibition and reprogram into a stably drug resistant fate. However, the biological processes governing cellular priming remain unknown. We used CRISPR/Cas9 genetic screens to identify genes that affect cell fate decisions by altering cellular plasticity. We found that many factors can independently affect cellular priming and fate decisions. We discovered a novel, plasticity-based mode of increasing resistance to BRAF inhibition that pushes cells towards a more differentiated state. Manipulating cellular plasticity through inhibition of DOT1L before the addition of the BRAF inhibitor resulted in more therapy resistance than concurrent administration. Our results indicate that modulating cellular plasticity can alter cell fate decisions and may prove useful for treating drug resistance in other cancers.

Abstract 625: Selective inhibition of NF1 loss-of-function driven oncogenic signaling in melanomas

Abstract Oncogenic activation of the MAPK signaling pathway occurs across >90% of melanomas, motivating the clinical evaluation of MAPK-targeted therapies in these tumors. However, only BRAFV600E/K-mutated melanomas show consistent clinical responses to MAPK-targeted therapies. Inactivation of the tumor suppressor gene neurofibromin 1 (NF1) is among the mechanisms that drive hyperactivated RAS signaling in >10% of melanomas. These tumors, however, have variable dependency on RAS-mediated pathways, such as MAPK and PI3K/mTOR, and thus exhibit inconsistent sensitivity to targeted inhibitors of these pathways. It is of paramount importance, therefore, to identify new therapeutic strategies to block tumor cell growth or improve MAPK- or PI3K-targeted therapy response in NF1 loss-of-function (NF1LoF) melanomas. To identify actionable vulnerabilities in NF1LoF melanoma cells, we apply a systematic approach, combining multiplexed high-throughput immunofluorescence microscopy, single-cell measurements and computational analysis, to study the responses of differentially mutated human melanoma cell lines to a variety of RAS-related pathway inhibitors. Our analysis led to the identification of a multi-kinase inhibitor, MTX-216, which is efficacious in induction of tumor cell death and inhibition of NF1LoF melanoma growth both in vitro and in vivo, but does not exhibit cytotoxicity in BRAFV600E melanomas and wild-type melanocytes. Single-cell analysis of downstream RAS-mediated signaling markers (e.g. phospho-S6) and proliferation markers (e.g. Ki-67) demonstrated cell-to-cell heterogeneity in NF1LoF melanoma cells treated with commercially available kinase inhibitors, resulting in incomplete drug responses. MTX-216 reduces this single-cell variability, providing evidence that co-suppression of signaling and proliferation pathways is correlated with drug-induced cytotoxicity. Our systematic analysis, together with ongoing studies to uncover the mechanism of action of MTX-216 in NF1LoF melanomas, has the potential to identify new and effective treatment strategies for NF1-mutant melanomas. Citation Format: Cara Abecunas, Mohammad Fallahi-Sichani, Judith Leopold, Christopher Whitehead, Elizabeth Ziemke. Selective inhibition of NF1 loss-of-function driven oncogenic signaling in melanomas [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 625.

Development of an epigenetic tetracycline sensor system based on DNA methylation.

Bacterial live cell sensors are potentially powerful tools for the detection of environmental toxins. In this work, we have established and validated a flow cytometry readout for an existing bacterial arabinose sensor system with DNA methylation based memory function (Maier et al., 2017, Nat. Comm., 8:15336). Flow cytometry readout is convenient and enables a multiparameter analysis providing information about single-cell variability, which is beneficial for further development of sensor systems of this type in the future. We then designed a tetracycline sensor system, because of the importance of antibiotics pollution in the light of multi-resistant pathogens. To this end, a tetracycline trigger plasmid was constructed by replacing the araC repressor gene and the ara operator of the arabinose trigger plasmid with the tetR gene coding for the tetracycline repressor and the tet operon. After combination with the memory plasmid, the tetracycline sensor system was shown to be functional in E. coli allowing to detect and memorize the presence of tetracycline. Due to a positive feedback between the trigger and memory systems, the combined whole-cell biosensor showed a very high sensitivity for tetracycline with a detection threshold at 0.1 ng/ml tetracycline, which may be a general property of sensors of this type. Moreover, acute presence of tetracycline and past exposure can be detected by this sensor using the dual readout of two reporter fluorophores.

Gene Networks with Transcriptional Bursting Recapitulate Rare Transient Coordinated High Expression States in Cancer.

Non-genetic transcriptional variability is a potential mechanism for therapy resistance in melanoma. Specifically, rare subpopulations of cells occupy a transient pre-resistant state characterized by coordinated high expression of several genes and survive therapy. How might these rare states arise and disappear within the population? It is unclear whether the canonical models of probabilistic transcriptional pulsing can explain this behavior, or if it requires special, hitherto unidentified mechanisms. We show that a minimal model of transcriptional bursting and gene interactions can give rise to rare coordinated high expression states. These states occur more frequently in networks with low connectivity and depend on three parameters. While entry into these states is initiated by a long transcriptional burst that also triggers entry of other genes, the exit occurs through independent inactivation of individual genes. Together, we demonstrate that established principles of gene regulation are sufficient to describe this behavior and argue for its more general existence. A record of this paper's transparent peer review process is included in the Supplemental Information.

Counting Proteins and Nucleic Acids with Single-molecule Microscopy

Single-molecule localization microscopy (SMLM) has the potential to revolutionize proteomic and genomic analyses by providing information on the number and stoichiometry of proteins or nucleic acids aggregating at spatial scales below the diffraction limit of light. Here, we present our progress in developing a molecular counting technique for SMLM data built upon the exponentially distributed blinking statistics of photoswitchable fluorophores. The approach was motivated by a desire to quantify the single-cell variability in plasmid copy number, but can be applied to a wide range of applications. We will present our work benchmarking the method on fluorescently labeled, surface mounted DNA origami grids and our progress in counting plasmids. The accuracy of our results illustrates SMLM's utility for optical -omics analysis.

From single-cell variability to population growth.

Single-cell experiments have revealed cell-to-cell variability in generation times and growth rates for genetically identical cells. Theoretical models relating the fluctuating generation times of single cells to the population growth rate are usually based on the assumption that the generation times of mother and daughter cells are uncorrelated. This assumption, however, is inconsistent with the exponential growth of cell volume in time observed for many cell types. Here we develop a more general and biologically relevant model in which cells grow exponentially and generation times are correlated in a manner which controls cell size. In addition to the fluctuating generation times, we also allow the single-cell growth rates to fluctuate and account for their correlations across the lineage tree. Surprisingly, we find that the population growth rate only depends on the distribution of single-cell growth rates and their correlations.

Learning unsupervised feature representations for single cell microscopy images with paired cell inpainting.

Cellular microscopy images contain rich insights about biology. To extract this information, researchers use features, or measurements of the patterns of interest in the images. Here, we introduce a convolutional neural network (CNN) to automatically design features for fluorescence microscopy. We use a self-supervised method to learn feature representations of single cells in microscopy images without labelled training data. We train CNNs on a simple task that leverages the inherent structure of microscopy images and controls for variation in cell morphology and imaging: given one cell from an image, the CNN is asked to predict the fluorescence pattern in a second different cell from the same image. We show that our method learns high-quality features that describe protein expression patterns in single cells both yeast and human microscopy datasets. Moreover, we demonstrate that our features are useful for exploratory biological analysis, by capturing high-resolution cellular components in a proteome-wide cluster analysis of human proteins, and by quantifying multi-localized proteins and single-cell variability. We believe paired cell inpainting is a generalizable method to obtain feature representations of single cells in multichannel microscopy images.

Hyperosmotic Stress Response Memory is Modulated by Gene Positioning in Yeast.

Cellular memory is a critical ability that allows microorganisms to adapt to potentially detrimental environmental fluctuations. In the unicellular eukaryote Saccharomyces cerevisiae, cellular memory can take the form of faster or slower responses within the cell population to repeated stresses. Using microfluidics and fluorescence time-lapse microscopy, we studied how yeast responds to short, pulsed hyperosmotic stresses at the single-cell level by analyzing the dynamic behavior of the stress-responsive STL1 promoter (pSTL1) fused to a fluorescent reporter. We established that pSTL1 exhibits variable successive activation patterns following two repeated short stresses. Despite this variability, most cells exhibited a memory of the first stress as decreased pSTL1 activity in response to the second stress. Notably, we showed that genomic location is important for the memory effect, since displacement of the promoter to a pericentromeric chromatin domain decreased the transcriptional strength of pSTL1 and led to a loss of memory. This study provides a quantitative description of a cellular memory that includes single-cell variability and highlights the contribution of chromatin structure to stress memory.

Single-cell variability of growth interactions within a two-species bacterial community

Cell–cell interaction networks have been examined in many high diversity microbial communities using macroscale approaches. Microscale studies of multispecies communities are lacking and it remains unclear how macroscale trends scale down to small groups of cells. Experimental approaches using microfluidic devices have revealed heterogeneity in the behavior of single cells, however, this analysis has not been extended towards the variability of cell–cell interactions. Using a microwell device, we analyzed cell growth within hundreds of replicate microbial communities consisting of two species and small population sizes. The wells of the devices were inoculated with a coculture of Escherichia coli and Enterobacter cloacae. Each species expressed a unique fluorescent protein enabling simultaneously tracking of cell number for each species over time. Growth dynamics within the device were consistent with bulk measurements. The device enabled monitoring of replicate, isolated coculture populations at high magnification, revealing both the growth interaction between the two species and the variability of such cell–cell interactions within small groups of cells. The device enables new experimental measurements of the heterogeneity of interactions within small, multispecies populations of bacteria.

Glycosphingolipid metabolism in cell fate specification.

Glycosphingolipids (GSLs) are ubiquitous components of eukaryotic plasma membranes that consist of a ceramide backbone linked to a glycan moiety. Both the ceramide and the glycan parts of GSLs display structural variations that result in a remarkable repertoire of diverse compounds. This diversity of GSLs is exploited during embryogenesis, when different GSLs are produced at specific developmental stages and along several differentiation trajectories. Importantly, plasma membrane receptors interact with GSLs to modify their activities. Consequently, two otherwise identical cells can respond differently to the same stimulus owing to their different GSL composition. The metabolic reprograming of GSLs is in fact a necessary part of developmental programs, as its impairment results in developmental failure or tissue-specific defects. Moreover, single-cell variability is emerging as a fundamental player in development: GSL composition displays cell-to-cell variability in syngeneic cell populations owing to the regulatory gene expression circuits involved in microenvironment adaptation and in differentiation. Here, we discuss how GSLs are synthesized and classified and review the role of GSLs in the establishment and maintenance of cell identity. We further highlight the existence of the regulatory circuits that modify GSL pathways and speculate how GSL heterogeneity might contribute to developmental patterning.