Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Mar 5.
Published in final edited form as: Cell Metab. 2018 Dec 20;29(3):539–544. doi: 10.1016/j.cmet.2018.11.016

Single-cell RNA-seq of the Pancreatic Islets – A Promise Not Yet Fulfilled?

Yue J Wang 1, Klaus H Kaestner 1
PMCID: PMC6402960  NIHMSID: NIHMS1515340  PMID: 30581120

SUMMARY

In the past three years, we have seen a flurry of publications on single-cell RNA-seq analyses of the pancreatic islets from mouse and human. This technology holds the promise to refine cell-type signatures and discover cellular heterogeneity among the canonical endocrine cell types such as the glucagon-producing alpha and insulin-producing beta-cells, going as far as suggesting new subtypes. In addition, single-cell RNA-seq has the ability to characterize rare endocrine cell types which are not captured by prior bulk analysis. With transcriptomics data from individual endocrine cells, cellular states can be profiled both along developmental processes and during the emergence of metabolic diseases. However, the promises of this new technology have not been met in full. While the methodology for the first time enabled the transcriptional definition of rare endocrine cell-types such as epsilon cells, some of the conclusions regarding cell-type specific gene expression changes in T2D might need to be revisited once larger sample sizes become available. Data generation and analysis are continuously improving single-cell RNA-seq approaches and are helping us to understand the (mal)adaptations of the islet cells during development, metabolic challenge, and disease.

Keywords: Single-cell RNA-seq, islets, beta-cell, beta-cell heterogeneity

eTOC

Wang and Kaestner review recent advances made in the study of pancreatic islets from mouse and human through the application of single-cell RNA-seq methodologies. The authors summarize novel discoveries regarding the transcriptomes of rare islets cell types and gene expression changes in diabetes, and discuss technological limitations and future developments.

Introduction

Single cell phenotyping of transcriptome changes has been the dream of biologists since the invention of Northern blot analysis for the detection of transcripts more than 40 years ago. We have come a long way since that time, when a million cells were required to obtained sufficient amounts of RNA to detect a single messenger RNA (mRNA). Today, it is possible to determine the mRNA levels of several thousand genes in thousands of cells with ease through single-cell RNA sequencing (scRNA-seq). Multiple methodological technologies are competing in this space, and it will likely take a few more years until a robust standardized approach is adopted by researchers world-wide. Even more development is expected in terms of computational analyses of these complex yet somewhat variable datasets. Nevertheless, even today single cell RNA-seq studies are bringing progress to islet cell biology.

The Promise of Single-cell RNA-seq

Historically, the first attempts to quantify changes in the islet transcriptome in diabetes beyond the determination of single mRNA or protein abundance was the microarray study of whole pancreatic islets (Bensellam et al., 2009; Gunton et al., 2005). Although innovative at the time, the study by Gunton and colleagues is instructive on the limitations of this approach, mostly attributed to the inherent problems with microarray technology, but also from the fact that islet cell composition is highly variable even among non-diabetics controls, with the beta cell percentage varying from 25 to 80%, while alpha and delta cells make up 2–67% and 5–25%, respectively (Brissova et al., 2005; Wang et al., 2016a). Thus, small amplitude changes that occur in beta cells – or any other endocrine cell type – are masked by the variation in the messenger RNA fraction that is contributed by the relevant cell type to total islet RNA.

A major improvement in islet transcriptomics came from the ability to sort live endocrine cell types. This was accomplished in mice through the application of transgenes that express fluorescent proteins or other sortable markers in the cell type of interest (Hara et al., 2003), and in humans by employing cellular stains that enriched for beta cells (Lukowiak et al., 2001), and more recently, the development of a set of surface antibodies that enables the separation of alpha and beta cells from exocrine and other contaminating cell types (Dorrell et al., 2008). The surface marker approach has been useful for the derivation of alpha and beta cell gene signatures and their correlation with epigenomic marks, as well as the analysis of the human endocrine pancreas during postnatal maturation (Arda et al., 2016; Bramswig et al., 2013); (Ackermann et al., 2016). Some limitations remain, for example the fact that the sorting protocol does not allow for the enrichment of the rare endocrine cell types such as delta, epsilon and PP cells. The biggest issue is of course that bulk RNA analysis by definition can deliver only the mean expression value for each gene in the cell ensemble. In comparison, single-cell RNA-seq methodologies can capture a much richer set of gene expression parameters across the cells analyzed. Thus, in addition to the discovery and description of postulated specific biological states such as beta or alpha cell subtypes, changes in gene expression variability driven by metabolic disease can be determined.

Over the past decade, single-cell biology has seen the development of a dizzying array of new methodologies for the analysis of gene expression in biological system in individual cells. Given that cells are the ultimate determinants of tissue biology, including that of the endocrine pancreas, these methods have potential to deliver answers to long-standing questions in islet biology and diabetes: Are the canonical endocrine cell types the final list, or are there hormone-expressing cells that have been missed by prior immunostaining and bulk tissue analyses? Are all beta or alpha cells the same, or are there permanent or at least semi-permanent subtypes with separate functional properties? If these exist, how do these subtypes differentiate during ontogeny, and how do they behave in disease states such as type 1 and type 2 diabetes?

Each time following a technological breakthrough, whether it is transmission electron microscopy or genome-wide association studies, there is a rush of investigators to apply them to their biological system of interest. Single-cell RNA-seq is no exception, and over the past three years over a dozen publication have reported results from its application to pancreatic islet (Baron et al., 2016; Enge et al., 2017; Lawlor et al., 2017; Li et al., 2016; Muraro et al., 2016; Qiu et al., 2018; Segerstolpe et al., 2016; Wang et al., 2016b; Xin et al., 2016a, 2016b, 2016c, 2018; Zeng et al., 2017). There are several recent reviews that collate the individual findings (Avrahami et al., 2017; Carrano et al., 2017). Here, we will attempt to summarize the development of single-cell RNA methodologies, give an overview of the discoveries made in the islet biology field using these approaches, and discuss their limitations and potentials for future improvements

Single-cell RNA-seq defines the transcriptomes of rare endocrine cell types

Single-cell transcriptomic analysis has undergone dramatic advances over the past five years, with multiple methodologies being employed by islet researchers to interrogate endocrine cell biology. Li and colleagues were the first to profile 70 islet cells from one control donor (Li et al., 2016). Due to the sample size constraint, their analysis was limited to the determination of the major cell types using known marker genes such as INSULIN and GLUCAGON (Li et al., 2016). Over the past three years, multiple studies followed with larger numbers of cells from islets of deceased organ donors, including those with T2D and children (Baron et al., 2016; Enge et al., 2017; Lawlor et al., 2017; Muraro et al., 2016; Qiu et al., 2018; Segerstolpe et al., 2016; Wang et al., 2016b; Xin et al., 2016a, 2016b, 2016c, 2018; Zeng et al., 2017). An important outcome of these studies is the definition of gene signatures for all endocrine cell types, including those with low abundance and/or for which no sorting methodology is available. For instance, Segerstolpe and colleagues defined the transcriptome of ghrelin-producing epsilon cells, and discovered that in addition to ghrelin, these cells are also highly enriched for BMP4 mRNA and a ghrelin-associated long non-coding RNA (Segerstolpe et al., 2016). Thus, single-cell RNA-seq studies delivered on the promise to refine cell-type specific gene signatures in the human pancreas.

Analysis of beta-cell subtypes using single-cell RNA-seq

We have known for decades that beta-cells -- and by extension, the four other endocrine cell types present in islets -- are not functionally equivalent, but analytical tools to identify inherent variations in endocrine cells have been limited. Pertaining to the beta cell, evidence for functional heterogeneity was reported as early as the 1980s by Salomon and Meda, who determined the glucose response of rat beta cells on the single-cell level using the ingenious reverse hemolytic plaque assay (Salomon and Meda, 1986). Several studies confirmed and extended these findings (Bosco and Meda, 1991; Pipeleers, 1992; Pipeleers et al., 1994; Van Schravendijk et al., 1992), including the demonstration that beta cells with high insulin protein synthesis rates responded preferentially to glucose stimulation (Bosco and Meda, 1991). However, these studies did not go beyond in vitro experimentation, and did not establish whether the observed variation represented temporary states or semi-permanent or permanent subtypes of beta cells.

Heterogeneity among beta cells was not identified in the report by Xin et al. (Xin et al., 2016a). In contrast, Segerstolpe et al. uncovered a total of five different beta cell subtypes separated by a combination of RBP4, FFAR4/GPR120, ID1, ID2, and ID3 expression (Segerstolpe et al., 2016). Muraro and colleagues (Muraro et al., 2016) identified a small set of genes that were differentially expressed in beta-cells which have been implicated in the ER and oxidative stress response. Additional evidence for subclusters of beta cells with increased ER stress comes from the study by Baron et al., where the authors found differential expression of ER stress response genes (DDIT3, HERPUD1, and HSPA5) as well as beta cell maturation marker such as Urocortin 3 between different beta-cell groups (Baron et al., 2016).

Subtypes or cellular states might not be restricted only to pancreatic beta cells. Among ductal cells, Baron and colleagues found two expression profiles, which related to either centro-acinor or terminal duct cells (Baron et al., 2016; Rovira et al., 2010). Segerstolpe and colleagues reported subpopulations of alpha, beta and acinar cells (Segerstolpe et al., 2016). Proliferating alpha cells could be identified based on the high transcript count for cell cycle-associated genes such as Ki67. GLUCAGON transcript levels were unchanged in proliferating cells, suggesting that cycling alpha cells do not need to silence transcription of ‘function’ genes. Alternatively, a long half-life of the GLUCAGON mRNA could mask any transient decrease in its promoter activity. Cycling cells in this case likely only represent a temporary change in cell state, and not a true, permanent subpopulation of alpha cells. An important future study will be the integrative analysis of all single-cell transcriptome studies to determine if the subtype-defining genes sets are common to all studies, or if the subtype classification was influenced by data quality and/or sample size.

New insights into beta-cell pathology in T2D – a work in progress

With regard to the islet cell transcriptome in T2D, several groups identified differentially expressed genes in T2D compared to control islets spreading across all endocrine cell types (Lawlor et al., 2017; Segerstolpe et al., 2016; Xin et al., 2016a). In Figure 1, we intersected the gene lists from the aforementioned three publications (Lawlor et al., 2017; Segerstolpe et al., 2016; Xin et al., 2016a), focusing on genes up or down regulated in T2D beta cells compared with controls. We observed that the lists of differentially expressed genes described by the different groups are largely non-overlapping (Figure 1). This discrepancy reflects in large part the complex etiology of T2D, and the fact thus far only a limited number of donor samples have been available for analysis. It is likely that once the scope of these studies can be expanded to a large number of islet donors, or to specific, preselected patient cohorts, common disease-associated transcriptome changes will be discovered. In addition, as none of these T2D-altered genes were functionally validated, it remains to be determined if they play any role in the disease process (Lawlor et al., 2017; Segerstolpe et al., 2016; Xin et al., 2016a). By including samples from both T2D donors and children, Wang and colleagues discovered that both alpha and beta cells from T2D donors have gene expression signatures resembling pediatric donors, suggestive of a dedifferentiation process (Wang et al., 2016b). However, Wang et al. did not directly perform differential expression analysis comparing T2D samples with control donors (Wang et al., 2016b).

Figure 1:

Figure 1:

Open in a new tab

Venn diagram displaying the overlap among the sets of differentially expressed genes (comparing beta cells from control and T2D patients) from three single cell RNA-seq studies (Lawlor et al., 2017; Segerstolpe et al., 2016; Xin et al., 2016a), separated for upregulated and downregulated genes. Note the limited overlap among the genes identified as differentially regulated by the three studies.

The time dimension – analyzing development of endocrine cells at the single cell level

Another informative application of single-cell RNA-seq is to infer temporal gene expression kinetics. This is achieved by first ordering individual cells by their ‘pseudotimes’ based on their transcriptomic profiles, followed by identifying genes that display differential expression along the progression of the pseudotime (Saelens et al., 2018; Trapnell et al., 2014). The name “pseudo”time originates from the fact that it is related to but not necessarily identical with the chronological time of sampling, due to the asynchronous and heterogeneous nature of biological samples. Compared with data from bulk mRNA abundance measurements, single-cell RNA-seq offers a significantly increased resolution to study longitudinal gene expression changes because each cell can be assigned with an unambiguous cell type and cell stage label. The pseudotime reconstruction framework has been applied to study the maturation process of beta and alpha cells in the mouse pancreas (Qiu et al., 2018; Zeng et al., 2017). Both groups analyzed beta and alpha cells isolated from mouse pancreata at different developmental stages and performed pseudotime ordering to map these beta or alpha cells along their maturation trajectory. Both groups subsequently identified signatures of immature beta cells (Qiu et al., 2018; Zeng et al., 2017). Zeng et al. carried the study further by validating the pathways that were associated with beta cell proliferation using in vitro cell culture and in vivo mouse model (Zeng et al., 2017).

The application of pseudotime analysis to derive signatures of different states of the human pancreatic endocrine cells has lagged behind somewhat, due to the limited availability of human islets compared to rodents, and the high degree of variability in islet preparation among human organ donors. In a recent report, Xin and colleagues profiled a large number of islet cells from 12 non-diabetic deceased organ donors and employed the aforementioned pseudotime analysis to order beta cells (Xin et al., 2018). Their data suggest that human beta cells are separated by states of high insulin biosynthesis followed by recovery from ER stress by activation of the unfolded protein response (Xin et al., 2018). However, the pseudotime trajectory assumed a branching pattern rather than a linear transition between different states. Thus, this analysis revealed beta cell subpopulations and did not provide information on how beta cells transverse between different states. In addition, it is unclear how these beta cells change in the context of metabolic disorders such as diabetes. Nevertheless, the common theme between this study and several other single-cell RNA-seq studies points to a potential subgroup of beta cells with high ER stress and reduced insulin production (Baron et al., 2016; Muraro et al., 2016; Xin et al., 2018).

Single-cell RNA-seq can not only reveal transcriptomic profiles at the cellular level, the data can further be mined for mutation detection, alternative splicing analysis, eQTL imputing and other downstream studies (Arzalluz-Luque and Conesa, 2018; Piskol et al., 2013; Wills et al., 2013). This concept is nicely illustrated in the recent paper by Enge et al. (Enge et al., 2017), where the authors derived mutational history based on single-cell data from endocrine cell isolated from donors with different ages.

Limitations of Current Single-cell RNA-seq Islet Studies

A limitation common to all single-cell RNA-seq protocols in use to date is the partial capture of the cellular transcriptome. While in bulk RNA-seq, generally transcripts from more than 10,000 genes are easily discernable, in the case of single-cell RNA-seq of the pancreatic cells, the number of genes detected is typically in the scale of a few hundreds to a few thousands. Because the initial capture of mRNA molecules and their conversion to cDNA is of limited efficiency, very low abundance transcripts cannot yet be determined reliably. An example illustrating this issue is given in Figure 2, where among human beta cells, many marker genes are detected in only subsets of the cells, even though it is very likely that their mRNAs were in fact present in all cells analyzed. The issue of gene “dropouts” leads to loss of valuable information and increased sampling errors, which affects in particular genes expressed at low levels. Because these rare transcripts will – by chance – be captured in some cells but not in others, this has the potential to lead to artefactual overestimation of heterogeneity among cells in a given population (Kolodziejczyk et al., 2015; Wagner et al., 2016). It is expected that the efficacy of reverse transcription and initial amplification will be improved in the near future to minimize the impact of this issue. In addition, several computational methods have been published to specifically deal with the dropout issue by imputing undetected gene expression from underlying population gene-expression distribution (van Dijk et al., 2018; Huang et al., 2018; Li and Li, 2018; Pierson and Yau, 2015). However, these in silico methods are based on multiple assumptions that may or may not hold true for the biological samples under study.

Figure 2:

Figure 2:

Open in a new tab

Heatmap of beta cell markers genes in beta cells from Fluidigm 800HT platforms. “Dropouts”, i.e. cells with zero reads for the genes in question, are indicated in dark blue. Note that even key marker genes such as PCSK1 and MAFB, which are likely expressed in all human beta cells are not detected in a significant fraction of the cells analyzed, illustrating the limitations of the current technology.

Another issue pertaining to single-cell RNA-seq studies on the pancreas is that there is a wide range of gene expression levels within the same cell. The typical mammalian cells contains about 10 pg RNA, of which only ~2% are messenger RNAs. Assuming an average length of 1,000 nt (320,659 g/mol in molar mass), this corresponds to about 400,000 mRNA molecules. In alpha cells, for instance, the GLUCAGON mRNA alone makes up about 30% of the total mRNA molecules, while the mRNAs encoding DNA-binding transcription factors might exist only in a few copies per cell. Thus, the RNA isolation, reverse transcription, cDNA synthesis and amplification have to maintain a dynamic range of ~105 fold while allowing for the detection of single mRNA molecules. Fortunately, improvements to the methodology are constantly being made, including the recent development of ‘molecular crowding SCRB-seq’ (mcSCRB-seq), which includes polyethylene glycol in the cDNA synthesis reaction to decrease the effective reaction volume and thus increase efficacy by several fold (Bagnoli et al., 2018).

Biological variation is not only contributed by true subpopulations with distinct biological functions but also different cell cycle phases at time of isolation, the cell size, transcriptional bursting, and potential stress introduced during cell isolation (van den Brink et al., 2017; Buettner et al., 2015; McDavid et al., 2016; Raj et al., 2006). With regards to endocrine cells, the cell cycle does not contribute greatly to heterogeneity since the majority of adult islet cells are quiescent. However, one could envision that the gene expression profile of a beta cells that has just completed a burst of insulin secretion is different from one that has been functionally quiescent for hours during a prolonged fast. The transcriptomes will be different because of the physiological state, but one would not be able to determine whether the differences are only transient or represent a permanent division of labor. Regarding isolation stress, several groups developed single-nucleus RNA-seq, for which input can be obtained from fixed or frozen tissue archives (Habib et al., 2017; Lake et al., 2016). Compared with single-cell RNA-seq, the singlenucleus method not only can considerably increase the number of available biological sources, but can also preserve the cellular state at the time when tissues are fixed, thus reducing transcriptomic changes introduced during cell dissociation.

Methodology development to increase the power of single cell analysis

Methodology development can help to close the gap between cellular function and transcriptomics. Recently, two groups of neuroscientists combined whole-cell patch-clamp with single-cell RNA-seq (Cadwell et al., 2016; Fuzik et al., 2016). It is foreseeable that such type of studies on the pancreatic endocrine cells will provide additional insights on cellular functional heterogeneity. From the data analysis stand point, while tSNA, PCA and MDS plots reduce the data dimensions and help to reveal major cell clusters, the relationship information among cells is limited, since the relative position of the clusters to each other shown in the 2D graph may not accurately reflect the true relationship between neighboring cells. Thus, future topological methods that retain and display similarity values, which should represent lineage relationships, will greatly increase the value of single cell RNA-Seq datasets.

It is unlikely that single-cell transcriptome analyses has reached its limit; rather, the next five years will see many refinements and extensions to the technology. One example is the application of spatial resolution to medium and high throughput single-cell RNA analysis, resulting in “spatial genomics”. Already, the multiplexing of single molecule RNA fluorescence in situ hybridization promises the simultaneous detection and quantification of dozens or even hundreds of mRNA species in tissue sections (Shah et al., 2016a, 2016b). These novel technologies will most certainly uncover key novel aspects of islet cell biology. Another recent advance is the simultaneous analysis of mRNA expression levels and chromatin state (Cao et al., 2018), and multiplexed limited proteomics in tissue sections (Wang et al, Bodenmiller; Cell Metabolism, in review). These technological advances will allow diabetes researchers worldwide to analyze islet biology in development, function, and disease at unprecedented detail.

Conclusions

These are exciting times for islet biologists. Novel technologies are greatly increasing the resolution and thus the power of transcriptome analysis. While conventional determinations of mRNA levels using whole islets or even sorted islet cell populations captured only the gene expression changes with the largest amplitude averaged across the entire cell fraction, single cell technologies such as RNA-seq can capture changes in gene activity that occur only in a subset of cells but might be even more important in understanding disease progression. With further improvements to mRNA capture and cDNA synthesis, to extending these studies to much larger sample sizes, to advances in computational analysis methods, single cell RNA-seq and related approaches will greatly improve our understanding of the role of islets cells in the pathogenesis of both type 1 and type 2 diabetes.

Acknowledgements

Related work in the Kaestner lab is supported through NIH grants UC4DK104119 and UC4DK112217. We thank Golnaz Vahedi and Jonathan Schug for critical reading and discussion of the manuscript and apologize to all our colleagues whose related work we were not able to cite due to space limitations.

Footnotes

The authors have no conflicts of interest to declare.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Ackermann AM, Wang Z, Schug J, Naji A, and Kaestner KH (2016). Integration of ATAC-seq and RNA-seq identifies human alpha cell and beta cell signature genes. Mol Metab 5, 233–244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Andrews TS, and Hemberg M (2016). Dropout-based feature selection for scRNASeq. [DOI] [PMC free article] [PubMed]
  3. Arda HE, Efsun Arda H, Li L, Tsai J, Torre EA, Rosli Y, Peiris H, Spitale RC, Dai C, Gu X, et al. (2016). Age-Dependent Pancreatic Gene Regulation Reveals Mechanisms Governing Human β Cell Function. Cell Metab. 23, 909–920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Arzalluz-Luque Á, and Conesa A (2018). Single-cell RNAseq for the study of isoforms—how is that possible? Genome Biol. 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Avrahami D, Wang YJ, Klochendler A, Dor Y, Glaser B, and Kaestner KH (2017). Β-Cells are not uniform after all-Novel insights into molecular heterogeneity of insulin-secreting cells. Diabetes Obes. Metab. 19, 147–152. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bagnoli JW, Ziegenhain C, Janjic A, Wange LE, Vieth B, Parekh S, Geuder J, Hellmann I, and Enard W (2018). Sensitive and powerful single-cell RNA sequencing using mcSCRB-seq. Nat. Commun 9, 2937. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Baron M, Veres A, Wolock SL, Faust AL, Gaujoux R, Vetere A, Ryu JH, Wagner BK, Shen-Orr SS, Klein AM, et al. (2016). A Single-Cell Transcriptomic Map of the Human and Mouse Pancreas Reveals Inter- and Intra-cell Population Structure. Cell Systems 3, 346–360.e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bensellam M, Van Lommel L, Overbergh L, Schuit FC, and Jonas JC (2009). Cluster analysis of rat pancreatic islet gene mRNA levels after culture in low-, intermediate- and high-glucose concentrations. Diabetologia 52, 463–476. [DOI] [PubMed] [Google Scholar]
  9. Bosco D, and Meda P (1991). Actively synthesizing beta-cells secrete preferentially after glucose stimulation. Endocrinology 129, 3157–3166. [DOI] [PubMed] [Google Scholar]
  10. Bramswig NC, Everett LJ, Schug J, Dorrell C, Liu C, Luo Y, Streeter PR, Naji A, Grompe M, and Kaestner KH (2013). Epigenomic plasticity enables human pancreatic α to β cell reprogramming. J. Clin. Invest 123, 1275–1284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Brennecke P, Anders S, Kim JK, Kołodziejczyk AA, Zhang X, Proserpio V, Baying B, Benes V, Teichmann SA, Marioni JC, et al. (2013). Accounting for technical noise in singlecell RNA-seq experiments. Nat. Methods 10, 1093–1095. [DOI] [PubMed] [Google Scholar]
  12. van den Brink SC, Sage F, Vértesy Á, Spanjaard B, Peterson-Maduro J, Baron CS, Robin C, and van Oudenaarden A (2017). Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations. Nat. Methods 14, 935–936. [DOI] [PubMed] [Google Scholar]
  13. Brissova M, Fowler MJ, Nicholson WE, Chu A, Hirshberg B, Harlan DM, and Powers AC (2005). Assessment of human pancreatic islet architecture and composition by laser scanning confocal microscopy. J. Histochem. Cytochem 53, 1087–1097. [DOI] [PubMed] [Google Scholar]
  14. Buettner F, Natarajan KN, Casale FP, Proserpio V, Scialdone A, Theis FJ, Teichmann SA, Marioni JC, and Stegle O (2015). Computational analysis of cell-to-cell heterogeneity in single-cell RNA-sequencing data reveals hidden subpopulations of cells. Nat. Biotechnol 33, 155–160. [DOI] [PubMed] [Google Scholar]
  15. Butler A, Hoffman P, Smibert P, Papalexi E, and Satija R (2018). Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol 36, 411–420. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Cadwell CR, Palasantza A, Jiang X, Berens P, Deng Q, Yilmaz M, Reimer J, Shen S, Bethge M, Tolias KF, et al. (2016). Electrophysiological, transcriptomic and morphologic profiling of single neurons using Patch-seq. Nat. Biotechnol 34, 199–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Cao J, Cusanovich DA, Ramani V, Aghamirzaie D, Pliner HA, Hill AJ, Daza RM, McFaline-Figueroa JL, Packer JS, Christiansen L, et al. (2018). Joint profiling of chromatin accessibility and gene expression in thousands of single cells. Science. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Carrano AC, Mulas F, Zeng C, and Sander M (2017). Interrogating islets in health and disease with single-cell technologies. Mol Metab 6, 991–1001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Crow M, Paul A, Ballouz S, Huang ZJ, and Gillis J (2018). Characterizing the replicability of cell types defined by single cell RNA-sequencing data using MetaNeighbor. Nat. Commun 9, 884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. van Dijk D, Sharma R, Nainys J, Yim K, Kathail P, Carr AJ, Burdziak C, Moon KR, Chaffer CL, Pattabiraman D, et al. (2018). Recovering Gene Interactions from Single-Cell Data Using Data Diffusion. Cell 174, 716–729.e27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dorrell C, Abraham SL, Lanxon-Cookson KM, Canaday PS, Streeter PR, and Grompe M (2008). Isolation of major pancreatic cell types and long-term culture-initiating cells using novel human surface markers. Stem Cell Res. 1, 183–194. [DOI] [PubMed] [Google Scholar]
  22. Enge M, Efsun Arda H, Mignardi M, Beausang J, Bottino R, Kim SK, and Quake SR (2017). Single-Cell Analysis of Human Pancreas Reveals Transcriptional Signatures of Aging and Somatic Mutation Patterns. Cell 171, 321–330.e14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Fuzik J, Zeisel A, Máté Z, Calvigioni D, Yanagawa Y, Szabó G, Linnarsson S, and Harkany T (2016). Integration of electrophysiological recordings with single-cell RNA-seq data identifies neuronal subtypes. Nat. Biotechnol 34, 175–183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gunton JE, Kulkarni RN, Yim S, Okada T, Hawthorne WJ, Tseng Y-H, Roberson RS, Ricordi C, O’Connell PJ, Gonzalez FJ, et al. (2005). Loss of ARNT/HIF1β Mediates Altered Gene Expression and Pancreatic-Islet Dysfunction in Human Type 2 Diabetes. Cell 122, 337–349. [DOI] [PubMed] [Google Scholar]
  25. Habib N, Avraham-Davidi I, Basu A, Burks T, Shekhar K, Hofree M, Choudhury SR, Aguet F, Gelfand E, Ardlie K, et al. (2017). Massively parallel single-nucleus RNA-seq with DroNc-seq. Nat. Methods 14, 955–958. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Haghverdi L, Lun ATL, Morgan MD, and Marioni JC (2018). Batch effects in single-cell RNA-sequencing data are corrected by matching mutual nearest neighbors. Nat. Biotechnol 36, 421–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hara M, Wang X, Kawamura T, Bindokas VP, Dizon RF, Alcoser SY, Magnuson MA, and Bell GI (2003). Transgenic mice with green fluorescent protein-labeled pancreatic β-cells. American Journal of Physiology-Endocrinology and Metabolism 284, E177–E183. [DOI] [PubMed] [Google Scholar]
  28. Hashimshony T, Senderovich N, Avital G, Klochendler A, de Leeuw Y, Anavy L, Gennert D, Li S, Livak KJ, Rozenblatt-Rosen O, et al. (2016). CEL-Seq2: sensitive highly-multiplexed single-cell RNA-Seq. Genome Biol. 17, 77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Huang M, Wang J, Torre E, Dueck H, Shaffer S, Bonasio R, Murray JI, Raj A, Li M, and Zhang NR (2018). SAVER: gene expression recovery for single-cell RNA sequencing. Nat. Methods 15, 539–542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Jaitin DA, Kenigsberg E, Keren-Shaul H, Elefant N, Paul F, Zaretsky I, Mildner A, Cohen N, Jung S, Tanay A, et al. (2014). Massively Parallel Single-Cell RNA-Seq for Marker-Free Decomposition of Tissues into Cell Types. Science 343, 776–779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kiselev VY, Yiu A, and Hemberg M (2018). scmap: projection of single-cell RNA-seq data across data sets. Nat. Methods 15, 359–362. [DOI] [PubMed] [Google Scholar]
  32. Kolodziejczyk AA, Kim JK, Svensson V, Marioni JC, and Teichmann SA (2015). The Technology and Biology of Single-Cell RNA Sequencing. Mol. Cell 58, 610–620. [DOI] [PubMed] [Google Scholar]
  33. Lake BB, Ai R, Kaeser GE, Salathia NS, Yung YC, Liu R, Wildberg A, Gao D,-L. Fung H, Chen S, et al. (2016). Neuronal subtypes and diversity revealed by single-nucleus RNA sequencing of the human brain. Science 352, 1586–1590. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lawlor N, George J, Bolisetty M, Kursawe R, Sun L, Sivakamasundari V, Kycia I, Robson P, and Stitzel ML (2017). Single-cell transcriptomes identify human islet cell signatures and reveal cell-type-specific expression changes in type 2 diabetes. Genome Res. 27, 208–222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Li WV, and Li JJ (2018). An accurate and robust imputation method scImpute for single-cell RNA-seq data. Nat. Commun 9, 997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Li J, Klughammer J, Farlik M, Penz T, Spittler A, Barbieux C, Berishvili E, Bock C, and Kubicek S (2016). Single-cell transcriptomes reveal characteristic features of human pancreatic islet cell types. EMBO Rep. 17, 178–187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Lukowiak B, Vandewalle B, Riachy R, Kerr-Conte J, Gmyr V, Belaich S, Lefebvre J, and Pattou F (2001). Identification and purification of functional human beta-cells by a new specific zinc-fluorescent probe. J. Histochem. Cytochem 49, 519–528. [DOI] [PubMed] [Google Scholar]
  38. Macosko EZ, Basu A, Satija R, Nemesh J, Shekhar K, Goldman M, Tirosh I, Bialas AR, Kamitaki N, Martersteck EM, et al. (2015). Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets. Cell 161, 1202–1214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. McDavid A, Finak G, and Gottardo R (2016). The contribution of cell cycle to heterogeneity in single-cell RNA-seq data. Nat. Biotechnol 34, 591–593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Muraro MJ, Dharmadhikari G, Grün D, Groen N, Dielen T, Jansen E, van Gurp L, Engelse MA, Carlotti F, de Koning EJP, et al. (2016). A Single-Cell Transcriptome Atlas of the Human Pancreas. Cell Systems 3, 385–394.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Picelli S, Faridani OR, Björklund AK, Winberg G, Sagasser S, and Sandberg R (2014). Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc 9, 171–181. [DOI] [PubMed] [Google Scholar]
  42. Pierson E, and Yau C (2015). ZIFA: Dimensionality reduction for zero-inflated single-cell gene expression analysis. Genome Biol. 16, 241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Pipeleers DG (1992). Heterogeneity in pancreatic beta-cell population. Diabetes 41, 777–781. [DOI] [PubMed] [Google Scholar]
  44. Pipeleers D, Kiekens R, Ling Z, Wilikens A, and Schuit F (1994). Physiologic relevance of heterogeneity in the pancreatic beta-cell population. Diabetologia 37 Suppl 2, S57–S64. [DOI] [PubMed] [Google Scholar]
  45. Piskol R, Ramaswami G, and Li JB (2013). Reliable identification of genomic variants from RNA-seq data. Am. J. Hum. Genet 93, 641–651. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Qiu W-L, Zhang Y-W, Feng Y, Li L-C, Yang L, and Xu C-R (2018). Deciphering Pancreatic Islet β Cell and α Cell Maturation Pathways and Characteristic Features at the Single-Cell Level. Cell Metab. 27, 702. [DOI] [PubMed] [Google Scholar]
  47. Raj A, Peskin CS, Tranchina D, Vargas DY, and Tyagi S (2006). Stochastic mRNA Synthesis in Mammalian Cells. PLoS Biol. 4, e309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Rovira M, Scott S-G, Liss AS, Jensen J, Thayer SP, and Leach SD (2010). Isolation and characterization of centroacinar/terminal ductal progenitor cells in adult mouse pancreas. Proc. Natl. Acad. Sci. U. S. A 107, 75–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Saelens W, Cannoodt R, Todorov H, and Saeys Y (2018). A comparison of single-cell trajectory inference methods: towards more accurate and robust tools. [DOI] [PubMed]
  50. Salomon D, and Meda P (1986). Heterogeneity and contact-dependent regulation of hormone secretion by individual B cells. Exp. Cell Res 162, 507–520. [DOI] [PubMed] [Google Scholar]
  51. Segerstolpe Å, Palasantza A, Eliasson P, Andersson E-M, Andréasson A-C, Sun X, Picelli S, Sabirsh A, Clausen M, Bjursell MK, et al. (2016). Single-Cell Transcriptome Profiling of Human Pancreatic Islets in Health and Type 2 Diabetes. Cell Metab. 24, 593–607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Shah S, Lubeck E, Schwarzkopf M, He T-F, Greenbaum A, Sohn CH, Lignell A, Choi HMT, Gradinaru V, Pierce NA, et al. (2016a). Single-molecule RNA detection at depth by hybridization chain reaction and tissue hydrogel embedding and clearing. Development 143, 2862–2867. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Shah S, Lubeck E, Zhou W, and Cai L (2016b). In Situ Transcription Profiling of Single Cells Reveals Spatial Organization of Cells in the Mouse Hippocampus. Neuron 92, 342–357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Svensson V, Natarajan KN, Ly L-H, Miragaia RJ, Labalette C, Macaulay IC, Cvejic A, and Teichmann SA (2017). Power analysis of single-cell RNA-sequencing experiments. Nat. Methods 14, 381–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Trapnell C, Cacchiarelli D, Grimsby J, Pokharel P, Li S, Morse M, Lennon NJ, Livak KJ, Mikkelsen TS, and Rinn JL (2014). The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol 32, 381–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Trombetta JJ, Gennert D, Lu D, Satija R, Shalek AK, and Regev A (2014). Preparation of Single-Cell RNA-Seq Libraries for Next Generation Sequencing. Curr. Protoc. Mol. Biol 107, 4.221–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Van Schravendijk CF, Kiekens R, and Pipeleers DG (1992). Pancreatic beta cell heterogeneity in glucose-induced insulin secretion. J. Biol. Chem 267, 21344–21348. [PubMed] [Google Scholar]
  58. Volden R, Palmer T, Byrne A, Cole C, Schmitz RJ, Green RE, and Vollmers C (2018). Improving nanopore read accuracy with the R2C2 method enables the sequencing of highly multiplexed full-length single-cell cDNA. Proc. Natl. Acad. Sci. U. S. A [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wagner A, Regev A, and Yosef N (2016). Revealing the vectors of cellular identity with single-cell genomics. Nat. Biotechnol 34, 1145–1160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Wang YJ, Golson ML, Schug J, Traum D, Liu C, Vivek K, Dorrell C, Naji A, Powers AC, Chang K-M, et al. (2016a). Single-Cell Mass Cytometry Analysis of the Human Endocrine Pancreas. Cell Metab. 24, 616–626. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Wang YJ, Schug J, Won K-J, Liu C, Naji A, Avrahami D, Golson ML, and Kaestner KH (2016b). Single-Cell Transcriptomics of the Human Endocrine Pancreas. Diabetes 65, 3028–3038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Wills QF, Livak KJ, Tipping AJ, Enver T, Goldson AJ, Sexton DW, and Holmes C (2013). Single-cell gene expression analysis reveals genetic associations masked in whole-tissue experiments. Nat. Biotechnol 31, 748–752. [DOI] [PubMed] [Google Scholar]
  63. Wu AR, Neff NF, Kalisky T, Dalerba P, Treutlein B, Rothenberg ME, Mburu FM, Mantalas GL, Sim S, Clarke MF, et al. (2014). Quantitative assessment of single-cell RNA-sequencing methods. Nat. Methods 11, 41–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Xin Y, Kim J, Okamoto H, Ni M, Wei Y, Adler C, Murphy AJ, Yancopoulos GD, Lin C, and Gromada J (2016a). RNA Sequencing of Single Human Islet Cells Reveals Type 2 Diabetes Genes. Cell Metab. 24, 608–615. [DOI] [PubMed] [Google Scholar]
  65. Xin Y, Kim J, Ni M, Wei Y, Okamoto H, Lee J, Adler C, Cavino K, Murphy AJ, Yancopoulos GD, et al. (2016b). Use of the Fluidigm C1 platform for RNA sequencing of single mouse pancreatic islet cells. Proc. Natl. Acad. Sci. U. S. A 113, 3293–3298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Xin Y, Okamoto H, Kim J, Ni M, Adler C, Cavino K, Na E, Murphy AJ, Yancopoulos GD, Lin C, et al. (2016c). Single-Cell RNAseq Reveals That Pancreatic β-Cells From Very Old Male Mice Have a Young Gene Signature. Endocrinology 157, 3431–3438. [DOI] [PubMed] [Google Scholar]
  67. Xin Y, Gutierrez GD, Okamoto H, Kim J, Lee A-H, Adler C, Ni M, Yancopoulos GD, Murphy AJ, and Gromada J (2018). Pseudotime Ordering of Single Human β-Cells Reveals States of Insulin Production and Unfolded Protein Response. Diabetes 67, 1783–1794. [DOI] [PubMed] [Google Scholar]
  68. Zeng C, Mulas F, Sui Y, Guan T, Miller N, Tan Y, Liu F, Jin W, Carrano AC, Huising MO, et al. (2017). Pseudotemporal Ordering of Single Cells Reveals Metabolic Control of Postnatal β Cell Proliferation. Cell Metab. 25, 1160–1175.e11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Ziegenhain C, Vieth B, Parekh S, Reinius B, Guillaumet-Adkins A, Smets M, Leonhardt H, Heyn H, Hellmann I, and Enard W (2017). Comparative Analysis of Single-Cell RNA Sequencing Methods. Mol. Cell 65, 631–643.e4. [DOI] [PubMed] [Google Scholar]

RESOURCES