Single Cell Population Research Articles

Parallelized cytoindentation using convex micropatterned surfaces.

Here we present a high-throughput, parallelized cytoindentor for local compression of live cells. The cytoindentor uses convex lens-induced confinement (CLiC) to indent micrometer-sized areas in single cells and/or populations of cells with submicron precision. This is accomplished using micropatterned poly(dimethylsiloxane) (PDMS) films that are adhered to a convex lens to create arrays of extrusions referred to here as "posts." These posts caused local deformation of subcellular regions without any evidence of cell lysis upon CLiC indentation. Our micropost arrays were also functionalized with glycoproteins, such as fibronectin, to both pull and compress cells under customized confinement geometries. Measurements of Chinese hamster ovary (CHO-K1) cell migration trajectories and oxidative stress showed that the CLiC device did not damage or significantly stress the cells. Our novel tool opens a new area of investigation for visualizing mechanobiology and mechanochemistry within living cells, and the high-throughput nature of the technique will streamline investigations as current tools for mechanically probing material properties and molecular dynamics within cells, such as traditional cytoindentors and atomic force microscopy (AFM), are typically restricted to single-cell manipulation.

In utero methoxychlor exposure increases rat fetal Leydig cell number but inhibits its function.

The objective of the present study is to determine whether in utero exposure to methoxychlor (MXC) affects rat fetal Leydig cell number, cell size, or functions. Pregnant Sprague Dawley dams were gavaged with corn oil (control, 0mg/kg/day MXC) or MXC at doses of 10, 50, or 100mg/kg/day from gestational day (GD) 12 to 21. The results show that MXC increased fetal Leydig cell numbers dose-dependently from 95±8×103 cells/testis (control, mean±SEM) to 101±6, 148±22, and 168±21×103 cells/testis, at the doses of 10, 50, and 100mg/kg, respectively. The increase of Leydig cell number by MXC was contributed by the increase of single cell population of Leydig cells, which increased from 21±2% of the control to 31±4%, 39±3%, or 40±4% at the doses of 10, 50 or 100mg/kg, respectively. Quantitative PCR results show that MXC increased Lhcgr expression at dose of 10mg/kg and Scarb1 and Cyp11a1 mRNA levels at doses of 50 and 100mg/kg. Immunohistochemical staining demonstrated the increase of CYP11A1 protein level from the dose of 10mg/kg. However, at the highest dose (100mg/kg) MXC reduced the testicular testosterone level and MXC (1μM) in vitro treatment also inhibited androgen production from isolated fetal Leydig cells. In conclusion, our findings indicate that at low dose MXC may increase fetal Leydig cell numbers and the expressions of some steroidogenic enzymes, but at high dose it reduces the testicular testosterone level leading to reproductive tract malformations in the male offspring.

Robust Inference of Cell-to-Cell Expression Variations from Single- and K-Cell Profiling.

Quantifying heterogeneity in gene expression among single cells can reveal information inaccessible to cell-population averaged measurements. However, the expression level of many genes in single cells fall below the detection limit of even the most sensitive technologies currently available. One proposed approach to overcome this challenge is to measure random pools of k cells (e.g., 10) to increase sensitivity, followed by computational “deconvolution” of cellular heterogeneity parameters (CHPs), such as the biological variance of single-cell expression levels. Existing approaches infer CHPs using either single-cell or k-cell data alone, and typically within a single population of cells. However, integrating both single- and k-cell data may reap additional benefits, and quantifying differences in CHPs across cell populations or conditions could reveal novel biological information. Here we present a Bayesian approach that can utilize single-cell, k-cell, or both simultaneously to infer CHPs within a single condition or their differences across two conditions. Using simulated as well as experimentally generated single- and k-cell data, we found situations where each data type would offer advantages, but using both together can improve precision and better reconcile CHP information contained in single- and k-cell data. We illustrate the utility of our approach by applying it to jointly generated single- and k-cell data to reveal CHP differences in several key inflammatory genes between resting and inflammatory cytokine-activated human macrophages, delineating differences in the distribution of ‘ON’ versus ‘OFF’ cells and in continuous variation of expression level among cells. Our approach thus offers a practical and robust framework to assess and compare cellular heterogeneity within and across biological conditions using modern multiplexed technologies.

Identification of multi-loci hubs from 4C-seq demonstrates the functional importance of simultaneous interactions.

Use of low resolution single cell DNA FISH and population based high resolution chromosome conformation capture techniques have highlighted the importance of pairwise chromatin interactions in gene regulation. However, it is unlikely that associations involving regulatory elements act in isolation of other interacting partners that also influence their impact. Indeed, the influence of multi-loci interactions remains something of an enigma as beyond low-resolution DNA FISH we do not have the appropriate tools to analyze these. Here we present a method that uses standard 4C-seq data to identify multi-loci interactions from the same cell. We demonstrate the feasibility of our method using 4C-seq data sets that identify known pairwise and novel tri-loci interactions involving the Tcrb and Igk antigen receptor enhancers. We further show that the three Igk enhancers, MiEκ, 3′Eκ and Edκ, interact simultaneously in this super-enhancer cluster, which add to our previous findings showing that loss of one element decreases interactions between all three elements as well as reducing their transcriptional output. These findings underscore the functional importance of simultaneous interactions and provide new insight into the relationship between enhancer elements. Our method opens the door for studying multi-loci interactions and their impact on gene regulation in other biological settings.

Development of an Insert Co-culture System of Two Cellular Types in the Absence of Cell-Cell Contact

The role of secreted soluble factors in the modification of cellular responses is a recurrent theme in the study of all tissues and systems. In an attempt to make straightforward the very complex relationships between the several cellular subtypes that compose multicellular organisms, in vitro techniques have been developed to help researchers acquire a detailed understanding of single cell populations. One of these techniques uses inserts with a permeable membrane allowing secreted soluble factors to diffuse. Thus, a population of cells grown in inserts can be co-cultured in a well or dish containing a different cell type for evaluating cellular changes following paracrine signaling in the absence of cell-cell contact. Such insert co-culture systems offer various advantages over other co-culture techniques, namely bidirectional signaling, conserved cell polarity and population-specific detection of cellular changes. In addition to being utilized in the field of inflammation, cancer, angiogenesis and differentiation, these co-culture systems are of prime importance in the study of the intricate relationships that exist between the different cellular subtypes present in the central nervous system, particularly in the context of neuroinflammation. This article offers general methodological guidelines in order to set up an experiment in order to evaluating cellular changes mediated by secreted soluble factors using an insert co-culture system. Moreover, a specific protocol to measure the neuroinflammatory effects of cytokines secreted by lipopolysaccharide-activated N9 microglia on neuronal PC12 cells will be detailed, offering a concrete understanding of insert co-culture methodology.

Development of an Insert Co-culture System of Two Cellular Types in the Absence of Cell-Cell Contact.

The role of secreted soluble factors in the modification of cellular responses is a recurrent theme in the study of all tissues and systems. In an attempt to make straightforward the very complex relationships between the several cellular subtypes that compose multicellular organisms, in vitro techniques have been developed to help researchers acquire a detailed understanding of single cell populations. One of these techniques uses inserts with a permeable membrane allowing secreted soluble factors to diffuse. Thus, a population of cells grown in inserts can be co-cultured in a well or dish containing a different cell type for evaluating cellular changes following paracrine signaling in the absence of cell-cell contact. Such insert co-culture systems offer various advantages over other co-culture techniques, namely bidirectional signaling, conserved cell polarity and population-specific detection of cellular changes. In addition to being utilized in the field of inflammation, cancer, angiogenesis and differentiation, these co-culture systems are of prime importance in the study of the intricate relationships that exist between the different cellular subtypes present in the central nervous system, particularly in the context of neuroinflammation. This article offers general methodological guidelines in order to set up an experiment in order to evaluating cellular changes mediated by secreted soluble factors using an insert co-culture system. Moreover, a specific protocol to measure the neuroinflammatory effects of cytokines secreted by lipopolysaccharide-activated N9 microglia on neuronal PC12 cells will be detailed, offering a concrete understanding of insert co-culture methodology.

Abstract 2412: Reconstructing intratumor heterogeneity: lessons from therapeutic intervention in patient-specific models

Abstract Precision medicine approaches seeking to predict individual susceptibility to single or combinatorial drugs for advanced and/or metastasized solid cancers heavily depend on biological model systems that maintain key morphological, phenotypic and genomic features of the original tumor. Here, we present a case study of a female patient with metastasized colorectal cancer. Using tissue from five distinct areas of the primary tumor and its metastasis of the synchronously removed primary tumor and liver metastasis, we generated cell cultures. To address cell morphology and phenotypic features we performed immunohistochemistry (IHC) and immunofluorescence imaging showing that these cell cultures express colon-specific marker proteins/combinations CDX2 and CK20+/CK7-. Ultra-deep panel sequencing of original tumor tissue samples revealed damaging mutations in KRAS (G12D), PIK3CA (H1047R) and TP53 (C242F). Two cell cultures showed an additional SNP in the tumor suppressor SMAD4 (R361H), which was not detected in the original tissue, thus hinting at the existence of a small sub-population of cancer cells with a divergent mutation pattern. In parallel we established patient-derived xenografts (PDX) models from each region and investigated the response to targeted therapeutics, aiming at key molecules of the MAPK, PI3K/AKT and mTOR pathways in these genetically heterogeneous populations. drug response in single and mixed cell populations to address differential growth characteristics and response to mono- and combinatorial drug treatment. To distinguish the cells-of-origin and to calculate distributions/contribution of each population to the in vivo tumor growth in this model, sub-populations were color-coded using lentiviral constructs, expressing either a green or red tag. FACS analyses after treatment were performed and ratios of SMAD4mut(green) vs. SMAD4wt(red) were determined for each PDX tumor. These analyses, together with confirmatory IHC staining for the respective tag, showed significant effects on both tumor growth and population distributions in a treatment-dependent manner, advocating for systematic evaluation of combinatorial treatment in patients. Citation Format: Dirk Schumacher, Karsten Boehnke, Marlen Keil, Alessandra Silvestri, Maxine Silvestrov, Jens Hoffman, Iduna Fichtner, Reinhold Schaefer, Christian RA Regenbrecht. Reconstructing intratumor heterogeneity: lessons from therapeutic intervention in patient-specific models. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2412.

Comparison of methods to detect differentially expressed genes between single-cell populations

We compared five statistical methods to detect differentially expressed genes between two distinct single-cell populations. Currently, it remains unclear whether differential expression methods developed originally for conventional bulk RNA-seq data can also be applied to single-cell RNA-seq data analysis. Our results in three diverse comparison settings showed marked differences between the different methods in terms of the number of detections as well as their sensitivity and specificity. They, however, did not reveal systematic benefits of the currently available single-cell-specific methods. Instead, our previously introduced reproducibility-optimization method showed good performance in all comparison settings without any single-cell-specific modifications.

Predictive Computer Models for Biofilm Detachment Properties in Pseudomonas aeruginosa.

ABSTRACTMicrobial biofilm communities are protected against environmental extremes or clearance by antimicrobial agents or the host immune response. They also serve as a site from which microbial populations search for new niches by dispersion via single planktonic cells or by detachment by protected biofilm aggregates that, until recently, were thought to become single cells ready for attachment. Mathematically modeling these events has provided investigators with testable hypotheses for further study. Such was the case in the recent article by Kragh et al. (K. N. Kragh, J. B. Hutchison, G. Melaugh, C. Rodesney, A. E. Roberts, Y. Irie, P. Ø. Jensen, S. P. Diggle, R. J. Allen, V. Gordon, and T. Bjarnsholt, mBio 7:e00237-16, 2016, http://dx.doi.org/10.1128/mBio.00237-16), in which investigators were able to identify the differential competitive advantage of biofilm aggregates to directly attach to surfaces compared to the single-celled planktonic populations. Therefore, as we delve deeper into the properties of the biofilm mode of growth, not only do we need to understand the complexity of biofilms, but we must also account for the properties of the dispersed and detached populations and their effect on reseeding.

Numerical rate function determination in partial differential equations modeling cell population dynamics.

This paper introduces a method to solve the inverse problem of determining an unknown rate function in a partial differential equation (PDE) based on discrete measurements of the modeled quantity. The focus is put on a size-structured population balance equation (PBE) predicting the evolution of the number distribution of a single cell population as a function of the size variable. Since the inverse problem at hand is ill-posed, an adequate regularization scheme is required to avoid amplification of measurement errors in the solution method. The technique developed in this work to determine a rate function in a PBE is based on the approximate inverse method, a pointwise regularization scheme, which employs two key ideas. Firstly, the mollification in the directions of time and size variables are separated. Secondly, instable numerical data derivatives are circumvented by shifting the differentiation to an analytically given function. To examine the performance of the introduced scheme, adapted test scenarios have been designed with different levels of data disturbance simulating the model and measurement errors in practice. The success of the method is substantiated by visualizing the results of these numerical experiments.

Localization and Possible Function of P2X Receptors in Normal and Diseased Retinae

Purines, when present in the extracellular space, can mediate fast neurotransmission in the retina and central nervous system. Over the last decade there has been emerging evidence for the expression of P2X and P2Y receptors in a range of retinal neuronal subtypes. These results have highlighted important roles for purines in modulating specific retinal circuits, including the rod pathway and amacrine cell circuits. Traditionally, synaptic release of adenosine triphosphate (ATP) involves the novel anion vesicular nucleotide transporter, VNUT, which has recently been identified in a single wide-field amacrine cell population. In addition, nontraditional, conductive mechanisms of release have also been described in the retina. In the synapse, the enzymes involved in rapid degradation of purines are present in both plexiform layers of the retina. A role for P2X receptors in retinal diseases has also emerged recently. High concentrations of ATP lead to photoreceptor loss, through mechanisms involving P2X7 receptors. In addition, activation of P2X7 receptors is associated with activation of the inflammasome, a protein complex important for the release of proinflammatory cytokines. P2X receptors, especially P2X7, are emerging as targets to combat retinal disease.

Label-free analysis of GPCR-stimulation: The critical impact of cell adhesion

Label-free cell-based assays have been attracting growing attention in drug research. Optical approaches based on evanescent electric fields (e.g. EPIC, RWG/DMR, SPR) and electrochemical impedance analysis (ECIS, xCELLigence) are by far the most widespread techniques for such purposes. We compared three label-free approaches (ECIS, RWG/DMR and SPR) with respect to the activation of the human histamine H1 receptor (H1R) expressed by U-373 MG glioblastoma and genetically engineered HEK 293T cells. HEK 293T cells were either expressing the hH1R alone or in combination with the adhesion protein hMSR1. The β2-adrenergic receptor (β2-AR) expressed by bovine aortic endothelial cells (BAEC) served as a second cell model. Reduced cell adhesion to the surface of the sensing devices affected both, the optical and the impedance-based readout, but became much more obvious in case of RWG- or SPR-based assays. By contrast, the co-expression of hH1R and hMSR1 in HEK 293T cells strongly enhanced the signal compared to hH1R expression alone. As the sensitivity of the optical readouts is confined to a distance of 100–200nm from the surface, depending on the wavelength of the incident light, this observation is in accordance with tighter adhesion of the co-transfectants, inducing a shorter distance between the cell membrane and the substrate. Combining ECIS and SPR, allowing for simultaneous registration of both signals for a single cell population, provided a direct correlation of both readouts, when H1R or β2-AR stimulation was investigated for the same cell populations. Cell adhesion was found to have a critical impact on the results of label-free cell monitoring, in particular when techniques based on evanescent electric fields are applied.

Lymphocyte Isolation from Human Skin for Phenotypic Analysis and Ex Vivo Cell Culture.

Human skin has an important barrier function and contains various immune cells that contribute to tissue homeostasis and protection from pathogens. As the skin is relatively easy to access, it provides an ideal platform to study peripheral immune regulatory mechanisms. Immune resident cells in healthy skin conduct immunosurveillance, but also play an important role in the development of inflammatory skin disorders, such as psoriasis. Despite emerging insights, our understanding of the biology underlying various inflammatory skin diseases is still limited. There is a need for good quality (single) cell populations isolated from biopsied skin samples. So far, isolation procedures have been seriously hampered by a lack of obtaining a sufficient number of viable cells. Isolation and subsequent analysis have also been affected by the loss of immune cell lineage markers, due to the mechanical and chemical stress caused by the current dissociation procedures to obtain single cell suspension. Here, we describe a modified method to isolate T cells from both healthy and involved psoriatic human skin by combining mechanical skin dissociation using an automated tissue dissociator and collagenase treatment. This methodology preserves expression of most immune lineage markers such as CD4, CD8, Foxp3 and CD11c upon the preparation of single cell suspensions. Examples of successful CD4(+) T cell isolation and subsequent phenotypic and functional analysis are shown.

P53 Pulses Diversify Target Gene Expression Dynamics in an mRNA Half-Life-Dependent Manner and Delineate Co-regulated Target Gene Subnetworks

The transcription factor p53 responds to DNA double-strand breaks by increasing in concentration in a series of pulses of fixed amplitude, duration, and period. How p53 pulses influence the dynamics ofp53 target gene expression is not understood. Here, we show that, in bulk cell populations, patterns of p53 target gene expression cluster into groups with stereotyped temporal behaviors, including pulsing and rising dynamics. These behaviors correlate statistically with the mRNA decay rates of target genes: short mRNA half-lives produce pulses of gene expression. This relationship can be recapitulated by mathematical models of p53-dependent gene expression in single cells and cell populations. Single-cell transcriptional profiling demonstrates that expression of a subset of p53 target genes is coordinated across time within single cells; p53 pulsing attenuates this coordination. These results help delineate how p53 orchestrates the complex DNA damage response and give insight into the function of pulsatile signaling pathways.

Multi‐chambers Microsystem for Simultaneous and Direct Electrochemical Detection of Reactive Oxygen and Nitrogen Species Released by Cell Populations

AbstractElectrochemical quantification of four reactive oxygen and nitrogen species (e.g., H2O2, ONOO−, NO and NO2−) emitted from macrophages populations was completed using a four‐chambers microchip. This microchip was fabricated from a glass substrate equipped with four sets of microband electrodes. Each set consisted of a platinum‐black coated working electrode combined to an Ag/AgCl reference and a Pt counter electrodes. A cover made of polydimethylsiloxane was designed with four wells whose open bottoms fitted over each set. Loading, culture and stimulation of cells were conducted within each well. After stimulation of macrophages by a calcium ionophore, amperometric responses were then monitored using a multipotentiostat to detect one species per well simultaneously. The results showed good reproducibility and were quite consistent with averaged amperometric detections obtained from single cell measurements. This methodology allowed drastically diminishing the number of experiments necessary to reach a statistical significance from n=100–160 experiments required for single cell detection compared to n=5 experiments to obtain the mean behavior of the cell population. This work demonstrated that this simple and versatile microchip could be used for quantification of species in oxidative stress investigations on cell populations as well as a complementary tool to compare behaviors between single cell and cell population for investigations involving many experimental variables.

Effects of oxygen tension and dextran-shelled/2H,3H-decafluoropentane-cored oxygen-loaded nanodroplets on secretion of gelatinases and their inhibitors in term human placenta.

Matrix metalloproteinases (MMPs) and their endogenous inhibitors (TIMPs) need to be finely modulated in physiological processes. However, oxygen tension influences MMP/TIMP balances, potentially leading to pathology. Intriguingly, new 2H,3H-decafluoropentane-based oxygen-loaded nanodroplets (OLNDs) have proven effective in abrogating hypoxia-dependent dysregulation of MMP and TIMP secretion by single cell populations. This work explored the effects of different oxygen tensions and dextran-shelled OLNDs on MMP/TIMP production in an organized and multicellular tissue (term human placenta). Chorionic villous explants from normal third-trimester pregnancies were incubated with/without OLNDs in 3 or 20% O2. Explants cultured at higher oxygen tension released constitutive proMMP-2, proMMP-9, TIMP-1, and TIMP-2. Hypoxia significantly altered MMP-2/TIMP-2 and MMP-9/TIMP-1 ratios enhancing TIMP-2 and reducing proMMP-2, proMMP-9, and TIMP-1 levels. Intriguingly, OLNDs effectively counteracted the effects of low oxygen tension. Collectively, these data support OLND potential as innovative, nonconventional, and cost-effective tools to counteract hypoxia-dependent dysregulation of MMP/TIMP balances in human tissues.

Selective and reversible suppression of intestinal stem cell differentiation by pharmacological inhibition of BET bromodomains.

Absorptive and secretory cells of the small intestine are derived from a single population of Lgr5-expressing stem cells. While key genetic pathways required for differentiation into specific lineages have been defined, epigenetic programs contributing to this process remain poorly characterized. Members of the BET family of chromatin adaptors contain tandem bromodomains that mediate binding to acetylated lysines on target proteins to regulate gene expression. In this study, we demonstrate that mice treated with a small molecule inhibitor of BET bromodomains, CPI203, exhibit greater than 90% decrease in tuft and enteroendocrine cells in both crypts and villi of the small intestine, with no changes observed in goblet or Paneth cells. BET bromodomain inhibition did not alter the abundance of Lgr5-expressing stem cells in crypts, but rather exerted its effects on intermediate progenitors, in part through regulation of Ngn3 expression. When BET bromodomain inhibition was combined with the chemotherapeutic gemcitabine, pervasive apoptosis was observed in intestinal crypts, revealing an important role for BET bromodomain activity in intestinal homeostasis. Pharmacological targeting of BET bromodomains defines a novel pathway required for tuft and enteroendocrine differentiation and provides an important tool to further dissect the progression from stem cell to terminally differentiated secretory cell.

Statistical Mechanical Framework for Predicting Cellular Responses from Single-Cell Data

Developments in single-cell analysis techniques have provided detailed resolution on cellular component copy number and variation within a cell population. These data provide a probability distribution for all possible states of the cell, as determined by the measured component copy number per cell. We have developed a statistical mechanical framework that uses single-cell protein population data to model the effective interactions of the measured proteins. This method enables us to predict ensemble changes in cellular components under a variety of perturbations. This framework uses Bayesian inference to compare given models or descriptions of the system, as well as determine the uncertainty in the parameter estimations given the data. We have applied this methodology to study the impact of changes in oxygen partial pressure on the behavior of glioblastoma multiforme (GBM) cancer cells. Our framework shows how we can determine the coupling of oxygen concentration changes to the measured proteins and their underlying effective interactions.

Isolation and gene expression analysis of single potential human spermatogonial stem cells.

It is possible to isolate pure populations of single potential human spermatogonial stem cells without somatic contamination for down-stream applications, for example cell culture and gene expression analysis. We isolated pure populations of single potential human spermatogonial stem cells (hSSC) without contaminating somatic cells and analyzed gene expression of these cells via single-cell real-time RT-PCR. The isolation of a pure hSSC fraction could enable clinical applications such as fertility preservation for prepubertal boys and in vitro-spermatogenesis. By utilizing largely nonspecific markers for the isolation of spermatogonia (SPG) and hSSC, previously published cell selection methods are not able to deliver pure target cell populations without contamination by testicular somatic cells. However, uniform cell populations free of somatic cells are necessary to guarantee defined growth conditions in cell culture experiments and to prevent unintended stem cell differentiation. Fibroblast growth factor receptor 3 (FGFR3) is a cell surface protein of human undifferentiated A-type SPG and a promising candidate marker for hSSC. It is exclusively expressed in small, non-proliferating subgroups of this spermatogonial cell type together with the pluripotency-associated protein and spermatogonial nuclear marker undifferentiated embryonic cell transcription factor 1 (UTF1). We specifically selected the FGFR3-positive spermatogonial subpopulation from two 30 mg biopsies per patient from a total of 37 patients with full spermatogenesis and three patients with meiotic arrest. We then employed cell selection with magnetic beads in combination with a fluorescence-activated cell sorter antibody directed against human FGFR3 to tag and visually identify human FGFR3-positive spermatogonia. Positively selected and bead-labeled cells were subsequently picked with a micromanipulator. Analysis of the isolated cells was carried out by single-cell real-time RT-PCR, real-time RT-PCR, immunocytochemistry and live/dead staining. Single-cell real-time RT-PCR and real-time RT-PCR of pooled cells indicate that bead-labeled single cells express FGFR3 with high heterogeneity at the mRNA level, while bead-unlabeled cells lack FGFR3 mRNA. Furthermore, isolated cells exhibit strong immunocytochemical staining for the stem cell factor UTF1 and are viable. The cell population isolated in this study has to be tested for their potential stem cell characteristics via xenotransplantation. Due to the small amount of the isolated cells, propagation by cell culture will be essential. Other potential hSSC without FGFR3 surface expression will not be captured with the provided experimental design. The technical approach as developed in this work could encourage the scientific community to test other established or novel hSSC markers on single SPG that present with potential stem cell-like features. The project was funded by the DFG Research Unit FOR1041 Germ cell potential (SCH 587/3-2) and DFG grants to K.v.K. (KO 4769/2-1) and A.-N.S. (SP 721/4-1). The authors declare no competing interests.

Robust detection of alternative splicing in a population of single cells.

Single cell RNA-seq experiments provide valuable insight into cellular heterogeneity but suffer from low coverage, 3′ bias and technical noise. These unique properties of single cell RNA-seq data make study of alternative splicing difficult, and thus most single cell studies have restricted analysis of transcriptome variation to the gene level. To address these limitations, we developed SingleSplice, which uses a statistical model to detect genes whose isoform usage shows biological variation significantly exceeding technical noise in a population of single cells. Importantly, SingleSplice is tailored to the unique demands of single cell analysis, detecting isoform usage differences without attempting to infer expression levels for full-length transcripts. Using data from spike-in transcripts, we found that our approach detects variation in isoform usage among single cells with high sensitivity and specificity. We also applied SingleSplice to data from mouse embryonic stem cells and discovered a set of genes that show significant biological variation in isoform usage across the set of cells. A subset of these isoform differences are linked to cell cycle stage, suggesting a novel connection between alternative splicing and the cell cycle.