Gene expression programs change over time, differentiation and development, and in response to stimuli. However, nearly all techniques for profiling gene expression in single cells do not directly capture transcriptional dynamics. In the present study, we present a method for combined single-cell combinatorial indexing and messenger RNA labeling (sci-fate), which uses combinatorial cell indexing and 4-thiouridine labeling of newly synthesized mRNA to concurrently profile the whole and newly synthesized transcriptome in each of many single cells. We used sci-fate to study the cortisol response in >6,000 single cultured cells. From these data, we quantified the dynamics of the cell cycle and glucocorticoid receptor activation, and explored their intersection. Finally, we developed software to infer and analyze cell-state transitions. We anticipate that sci-fate will be broadly applicable to quantitatively characterize transcriptional dynamics in diverse systems.
Subscribe to Journal
Get full journal access for 1 year
only $20.83 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
The data generated by this study can be downloaded in raw and processed forms from the National Center for Biotechnology Information’s Gene Expression Omnibus (GSE131351).
Scripts for processing sci-fate sequencing were written in Python and R with code available at https://github.com/JunyueC/sci-fate_analysis.
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Qiu, X. et al. Reversed graph embedding resolves complex single-cell trajectories. Nat. Methods 14, 979–982 (2017).
Wolf, F. A. et al. PAGA: graph abstraction reconciles clustering with trajectory inference through a topology preserving map of single cells. Genome Biol. 20, 59 (2019).
Haghverdi, L., Büttner, M., Wolf, F. A., Buettner, F. & Theis, F. J. Diffusion pseudotime robustly reconstructs lineage branching. Nat. Methods 13, 845–848 (2016).
Setty, M. et al. Wishbone identifies bifurcating developmental trajectories from single-cell data. Nat. Biotechnol. 34, 637–645 (2016).
Street, K. et al. Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genom. 19, 477 (2018).
Moris, N., Pina, C. & Arias, A. M. Transition states and cell fate decisions in epigenetic landscapes. Nat. Rev. Genet. 17, 693–703 (2016).
Herzog, V. A. et al. Thiol-linked alkylation of RNA to assess expression dynamics. Nat. Methods 14, 1198–1204 (2017).
Schofield, J. A., Duffy, E. E., Kiefer, L., Sullivan, M. C. & Simon, M. D. TimeLapse-seq: adding a temporal dimension to RNA sequencing through nucleoside recoding. Nat. Methods 15, 221–225 (2018).
Cao, J. et al. Comprehensive single-cell transcriptional profiling of a multicellular organism. Science 357, 661–667 (2017).
Cleary, M. D., Meiering, C. D., Jan, E., Guymon, R. & Boothroyd, J. C. Biosynthetic labeling of RNA with uracil phosphoribosyltransferase allows cell-specific microarray analysis of mRNA synthesis and decay. Nat. Biotechnol. 23, 232–237 (2005).
Dolken, L. et al. High-resolution gene expression profiling for simultaneous kinetic parameter analysis of RNA synthesis and decay. RNA 14, 1959–1972 (2008).
Miller, C. et al. Dynamic transcriptome analysis measures rates of mRNA synthesis and decay in yeast. Mol. Syst. Biol. 7, 458–458 (2014).
Duffy, E. E. et al. Tracking distinct RNA populations using efficient and reversible covalent chemistry. Mol. Cell 59, 858–866 (2015).
Schwalb, B. et al. TT-seq maps the human transient transcriptome. Science 352, 1225–1228 (2016).
Rabani, M. et al. Metabolic labeling of RNA uncovers principles of RNA production and degradation dynamics in mammalian cells. Nat. Biotechnol. 29, 436–442 (2011).
Miller, M. R., Robinson, K. J., Cleary, M. D. & Doe, C. Q. TU-tagging: cell type-specific RNA isolation from intact complex tissues. Nat. Methods 6, 439–441 (2009).
Cusanovich, D. A. et al. Multiplex single cell profiling of chromatin accessibility by combinatorial cellular indexing. Science 348, 910–914 (2015).
Cao, J. et al. Joint profiling of chromatin accessibility and gene expression in thousands of single cells. Science 361, 1380–1385 (2018).
Ramani, V. et al. Massively multiplex single-cell Hi-C. Preprint at bioRxiv https://doi.org/10.1101/065052 (2016)..
Mulqueen, R. M. et al. Highly scalable generation of DNA methylation profiles in single cells. Nat. Biotechnol. 36, 428–431 (2018).
Vitak, S. A. et al. Sequencing thousands of single-cell genomes with combinatorial indexing. Nat. Methods 14, 302–308 (2017).
Rosenberg, A. B. et al. Single-cell profiling of the developing mouse brain and spinal cord with split-pool barcoding. Science 360, 176–182 (2018).
Yin, Y. et al. High-throughput single-cell sequencing with linear amplification. Mol. Cell. 76, 676–690.e10 (2019).
Cao, J. et al. The single-cell transcriptional landscape of mammalian organogenesis. Nature 566, 496–502 (2019).
Buckingham, J. C. Glucocorticoids: exemplars of multi-tasking. Br. J. Pharmacol. 147, S258 (2006).
Reddy, T. E. et al. Genomic determination of the glucocorticoid response reveals unexpected mechanisms of gene regulation. Genome Res. 19, 2163–2171 (2009).
John, S. et al. Chromatin accessibility pre-determines glucocorticoid receptor binding patterns. Nat. Genet. 43, 264–268 (2011).
Reddy, T. E., Gertz, J., Crawford, G. E., Garabedian, M. J. & Myers, R. M. The hypersensitive glucocorticoid response specifically regulates period 1 and expression of circadian genes. Mol. Cell. Biol. 32, 3756–3767 (2012).
Vockley, C. M. et al. Direct GR binding sites potentiate clusters of TF binding across the human genome. Cell 166, 1269–1281.e19 (2016).
La Manno, G. et al. RNA velocity of single cells. Nature 560, 494 (2018).
McInnes, L., Healy, J., Saul, N. & Großberger, L. UMAP: Uniform Manifold Approximation and Projection. J. Open Source Software 3, 861 (2018).
Binder, E. B. The role of FKBP5, a co-chaperone of the glucocorticoid receptor in the pathogenesis and therapy of affective and anxiety disorders. Psychoneuroendocrinology 34, S186–S195 (2009).
Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018).
ENCODE Project Consortium et al. A user’s guide to the Encyclopedia of DNA Elements (ENCODE). PLoS Biol. 9, e1001046 (2011).
The ENCODE Project Consortium. The ENCODE (ENCyclopedia Of DNA Elements) project. Science 306, 636–640 (2004).
Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017).
Han, H. et al. TRRUST v2: an expanded reference database of human and mouse transcriptional regulatory interactions. Nucleic Acids Res. 46, D380–D386 (2018).
Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 44, W90–W97 (2016).
Boruk, M., Savory, J. G. A. & Haché, R. J. G. AF-2-dependent potentiation of CCAAT enhancer binding proteinβ -mediated transcriptional activation by glucocorticoid receptor. Mol. Endocrinol. 12, 1749–1763 (1998).
Qin, W. et al. Identification of functional glucocorticoid response elements in the mouse FoxO1 promoter. Biochem. Biophys. Res. Commun. 450, 979–983 (2014).
Sheela Rani, C. S., Elango, N., Wang, S.-S., Kobayashi, K. & Strong, R. Identification of an activator protein-1-like sequence as the glucocorticoid response element in the rat tyrosine hydroxylase gene. Mol. Pharmacol. 75, 589 (2009).
Fischer, M. & Müller, G. A. Cell cycle transcription control: DREAM/MuvB and RB-E2F complexes. Crit. Rev. Biochem. Mol. Biol. 52, 638–662 (2017).
Chou, J., Provot, S. & Werb, Z. GATA3 in development and cancer differentiation: cells GATA have it! J. Cell. Physiol. 222, 42–49 (2010).
Biswas, M. & Chan, J. Y. Role of Nrf1 in antioxidant response element-mediated gene expression and beyond. Toxicol. Appl. Pharmacol. 244, 16 (2010).
Ryoo, I.-G. & Kwak, M.-K. Regulatory crosstalk between the oxidative stress-related transcription factor Nfe2l2/Nrf2 and mitochondria. Toxicol. Appl. Pharmacol. 359, 24–33 (2018).
Heer, R., Robson, C. N., Shenton, B. K. & Leung, H. Y. The role of androgen in determining differentiation and regulation of androgen receptor expression in the human prostatic epithelium transient amplifying population. J. Cell. Physiol. 212, 572–578 (2007).
Meixner, A., Karreth, F., Kenner, L., Penninger, J. M. & Wagner, E. F. Jun and JunD-dependent functions in cell proliferation and stress response. Cell Death Differ. 17, 1409–1419 (2010).
Li, M. et al. Krüppel-like factor 5 promotes epithelial proliferation and DNA damage repair in the intestine of irradiated mice. Int. J. Biol. Sci. 11, 1458–1468 (2015).
Eberlé, D., Hegarty, B., Bossard, P., Ferré, P. & Foufelle, F. SREBP transcription factors: master regulators of lipid homeostasis. Biochimie 86, 839–848 (2004).
Shermoen, A. W. & O’Farrell, P. H. Progression of the cell cycle through mitosis leads to abortion of nascent transcripts. Cell 67, 303–310 (1991).
Palozola, K. C. et al. Mitotic transcription and waves of gene reactivation during mitotic exit. Science 358, 119–122 (2017).
Parsons, G. G. & Spencer, C. A. Mitotic repression of RNA polymerase II transcription is accompanied by release of transcription elongation complexes. Mol. Cell. Biol. 17, 5791–5802 (1997).
Sanchez-Alvarez, M., Zhang, Q., Finger, F., Wakelam, M. J. O. & Bakal, C. Cell cycle progression is an essential regulatory component of phospholipid metabolism and membrane homeostasis. Open Biol. 5, 150093 (2015).
Harmon, J. M., Norman, M. R., Fowlkes, B. J. & Thompson, E. B. Dexamethasone induces irreversible G1 arrest and death of a human lymphoid cell line. J. Cell. Physiol. 98, 267–278 (1979).
Greenberg, A. K. et al. Glucocorticoids inhibit lung cancer cell growth through both the extracellular signal-related kinase pathway and cell cycle regulators. Am. J. Respir. Cell Mol. Biol. 27, 320–328 (2002).
Erhard, F. et al. scSLAM-seq reveals core features of transcription dynamics in single cells. Nature 571, 419–423 (2019).
Hendriks, G.-J. et al. NASC-seq monitors RNA synthesis in single cells. Nat. Commun. 10, 3138 (2019).
Baran-Gale, J., Chandra, T. & Kirschner, K. Experimental design for single-cell RNA sequencing. Briefings in functional genomics 17, 233–239 (2018).
Chen, W. et al. UMI-count modeling and differential expression analysis for single-cell RNA sequencing. Genome Biol. 19, 70 (2018).
Matsushima, W. et al. SLAM-ITseq: sequencing cell type-specific transcriptomes without cell sorting. Development 145, dev164640 (2018).
Sharma, U. et al. Small RNAs are trafficked from the epididymis to developing mammalian sperm. Dev. Cell 46, 481–494 (2018).
Gay, L. et al. Mouse TU tagging: a chemical/genetic intersectional method for purifying cell type-specific nascent RNA. Genes Dev. 27, 98–115 (2013).
Hastie, T. & Stuetzle, W. Principal curves. J. Am. Stat. Assoc. 84, 502 (1989).
Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer, 2016).
Muhar, M. et al. SLAM-seq defines direct gene-regulatory functions of the BRD4-MYC axis. Science 360, 800–805 (2018).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Lindenbaum, P. JVarkit: java-based utilities for Bioinformatics. figshare (2015).
Krueger, F.. Trim Galore. GitHub https://github.com/FelixKrueger/TrimGalore (2019).
Li, H. et al. The sequence alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Broad Institute. Picard Tools. GitHub http://broadinstitute.github.io/picard/ (2019).
Koboldt, D. C. et al. VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res. 22, 568–576 (2012).
Wolock, S. L., Lopez, R. & Klein, A. M. Scrublet: computational identification of cell doublets in single-cell transcriptomic data. Preprint at bioRxiv https://doi.org/10.1101/357368 (2018).
Cole Trapnell Lab. Monocle release. GitHub https://github.com/cole-trapnell-lab/monocle-release (2019).
Friedman, J., Hastie, T. & Tibshirani, R. Regularization paths for generalized linear models via coordinate descent. J. Stat. Softw. 33, 1–22 (2010).
Kolde, R. pheatmap. GitHub https://github.com/raivokolde/pheatmap (2018).
Rodriguez, A. & Laio, A. Clustering by fast search and find of density peaks. Science 344, 1492–1496 (2014).
Wolf, F. A., Angerer, P. & Theis, F. J. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15 (2018).
We thank the members of the Shendure labs for helpful discussions, particularly X. Huang, R. Blecher, B. Martin, F. Chardon and R. Qiu; J. Rose, D. Maly, L. VandenBosch, and T. Reh’s lab for sharing the NIH/3T3 cell line; and J. McFaline-Figueroa for sharing the A549 cell line. This work was funded by the Paul G. Allen Frontiers Foundation (Allen Discovery Center grant to J.S. and C.T.), grants from the National Institutes of Health (grant nos. DP1HG007811 and R01HG006283 to J.S.; grant no. DP2 HD088158 to C.T.), the W. M. Keck Foundation (to C.T. and J.S.), the Dale. F. Frey Award for Breakthrough Scientists (to C.T.), the Alfred P. Sloan Foundation Research Fellowship (to C.T.) and the Brotman Baty Institute for Precision Medicine. J.S. is an investigator of the Howard Hughes Medical Institute.
F.J.S. declares competing financial interests in the form of stock ownership and paid employment by Illumina, Inc. One or more embodiments of one or more patents and patent applications filed by Illumina may encompass the methods, reagents and data disclosed in this article.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Integrated supplementary information
(a) Scatter plot of mouse (NIH/3T3) vs. human (HEK293T) UMI counts per cell in sci-fate. (b-d) Boxplot showing the proportion of reads mapping to the expected species (b), number of UMIs (c) and ratio of 4sU labeled reads (d) per cell from HEK293T (n = 932) and NIH/3T3 (n = 438) cells. For all box plots: thick middle lines, medians; upper and lower box edges, first and third quartiles, respectively; whiskers, 1.5 times the interquartile range; circles, outliers. (e-f) Spearman’s correlation between gene expression measurements in aggregated profiles of HEK293T (Gene number n = 22,687) (e) and NIH/3T3 cells (Gene number n = 21,973) (f) from sci-fate (y-axis) vs. sci-RNA-seq (x-axis) cells.
(a, b) Violin plot showing the number of UMIs (a) and genes (b) per cell in each of six treatment conditions. Cell number n = 1,054 (0h), 1,049 (2h), 949 (4h), 1,262 (6h), 1,041 (8h), and 1,325 (10h). For all violin plots: thick middle lines, medians; upper and lower box edges, first and third quartiles, respectively; whiskers, 1.5 times the interquartile range; circles, outliers. (c) Barplot showing the distribution of T > C mutation counts across labeled reads, estimated using reads from 100 randomly sampled cells and normalized by the total read number. (d) Violin plot showing the fraction of 4sU unlabeled reads per cell, split out by the subsets that map to exons vs. introns (Cell number n = 6,680). (e) Violin plot showing the fraction of exonic reads per cell, split out by the subsets that are labeled or unlabeled by 4sU (Cell number n = 6,680). (f) The x and y coordinates correspond to the joint information UMAP space (Cell number n = 6,680) shown in the rightmost panel of Fig. 1e, colored by DEX treatment time. Grey lines represent inferred cell state transition directions by RNA velocity with intronic reads (left) or newly synthesised reads (right). (g) Correlation plot showing the Pearson’s correlation between different treatment conditions, using either whole transcriptomes (upper right; circles, gene number n = 43,167) or newly synthesized transcriptomes (bottom left; squares, gene number n = 43,167). For each condition, we generated pseudobulk transcriptomes for either the newly synthesized or whole transcriptomes (i.e. aggregating across cells), and compared these in a pairwise fashion between conditions (e.g. whole transcriptome at 0 hrs vs. 4 hrs; newly synthesized transcriptome at 2 hrs vs. 6 hrs, etc.). The lowest correlations corresponded to the newly synthesized transcriptome with no DEX treatment (0 hrs) vs. the newly synthesized transcriptomes of any DEX treated condition. (h) Bar plot showing the expression fold-change before and after 2 hour DEX treatment for FGD4 and FKBP5, calculated using either whole RNA and newly synthesised RNA. Error bars represent standard deviation estimated by sampling (without replacement) 200 cells five times from each time point (n = 5, centre: mean). (i) Scatter plot showing the expression fold-change before and after 2 hour DEX treatment for differentially expressed (DE) genes identified using either whole transcriptomes or newly synthesized transcriptomes. The blue line represents y = x. (j-k) UMAP visualization of A549 cells (n = 6,680) via joint information of the whole and newly synthesized transcriptomes, colored by normalized expression of S phase marker genes in the whole (j) and newly synthesized (k) transcriptomes. UMI counts for these genes are scaled for library size, log-transformed, aggregated and then mapped to Z-scores.
(a) Scatter plot showing the Spearman’s correlations between either the overall expression (x-axis) or newly synthesized RNA expression (y-axis) for each TF-gene gene pair identified by correlation analysis using newly synthesized mRNA (blue, n = 448), full mRNA expression (red, n = 1,606) or found in both approaches (green, n = 538) (Methods). (b) Line plot showing the number of identified TF-gene links with varied numbers of sampled UMIs per cell, by the same TF-gene linkage analysis using newly synthesized mRNA or full mRNA expression. (c) Identified gene targets (grey, n = 5) of CEBPB (orange). Only links with regularized regression coefficients from LASSO > 0.06 are shown. (d) UMAP visualization of A549 cells (n = 6,680) via joint information of the whole and newly synthesized transcriptomes, colored by CEBPB expression (left) and activity (right). (e) similar to panel d, colored by the YOD1 expression (left) or activity (right). (f) similar to panel d, colored by the GTF2IRD1 expression (left) or activity (right). (g) similar to panel d, colored by the E2F1 expression (left), activity (middle) or aggregated whole transcriptome expression of E2F1 linked genes (right). (h) similar to panel d, colored by the expression and activity of KLF6 (left two panels), GATA3 (middle two panels) or NRF1 (right two panels).
Supplementary Figure 4 Twenty-seven cell states defined by combinations of TF module-defined states.
(a) Schematic of strategy for characterizing cell states as the combination of TF modules. (b) Supplementary Table showing the observed proportion of cells (black numbers) falling into each of 27 cell states that each correspond to one of 3 states defined by GR response (rows) and one of 9 states defined by the cell cycle module (columns). The red numbers in parentheses correspond to the expected proportions, assuming that the distributions are independent.
(a) We selected genes showing higher differences in normalized newly synthesis rate between 0 hrs and 2 hrs time points. For this, we first tested a series of thresholds for gene filtering and calculated the detection rate (α) for all genes. We then plotted the relationship between threshold and the ratio of genes with out-range α values (< 0 or > 1). Blue line represents loess smooth line for the data. We selected the threshold that was at the knee point of the plot, resulting in 186 genes selected. (b) Scatter plot (Gene number n = 186) showing differences between the normalized transcriptomes with no DEX treatment vs. 2 hours DEX treatment for each of 186 genes exhibiting the largest differences in new transcription between the two conditions. X-axis shows absolute differences between the conditions in the whole transcriptome. Y-axis shows absolute differences between the conditions in the newly synthesized transcriptome. Blue line is the linear regression line. Both whole transcriptome and newly synthesized transcriptome of each time point are normalized by the library size of whole transcriptome. (c) Histogram showing the distribution of the estimated detection rate of newly synthesized mRNA for each of 186 genes. (d) Scatter plot showing the median detection rate computed with different number of genes ordered by absolute differences in normalized newly synthesis rate between 0 hrs and 2 hrs time points. (e-f) Line plot showing the estimated new RNA detection rate (e) or Pearson’s correlation of estimated gene degradation rate with the full data (f, gene number n = 13,343) using varied numbers of sampled UMIs per cell. (g) Below-left of diagonal: correlation plots showing the Pearson’s correlations (r) of gene degradation rates between treatment conditions (Gene number n = 14,587). Diagonal: plots showing the distribution of gene degradation rates at each time point. Top-right of diagonal: correlation plots showing the Pearson’s correlations (r) values and mean squared errors (MSE) of gene degradation rates between treatment conditions. (h) Scatter plot between published per-gene mRNA half-lives (log transformed) in K562 cells (Nat. Methods, 221–225, 2018) (x-axis) vs. as estimated by sci-fate in A549 cells (y-axis) (Shared gene number n = 4,963).
Supplementary Figure 6 Reconstructed cellular past states overlapped with real cell states in the corresponding past time point even without Seurat integration.
(a) UMAP visualization of cells of 0 hrs treatment (Cell number n = 1,054) and original states (left) or reconstructed past states (right) from 2 hrs treatment (Cell number n = 1,049). (b) UMAP visualization of cells for reconstructed cellular past states of 4 hrs cells (Cell number n = 949) and profiled cells from 0 hrs (left, cell number n = 1,054) or 2 hrs (right, cell number n = 1,049). (c) Similar to Fig. 3f, barplots showing the contributions of the 9 different cell cycle states to each of three cell trajectory clusters by simply aligning neighboring time points, but without knowledge of newly synthesized mRNA.
(a-b) Scatter plots of cell state transition probabilities (a) or cell state proportions (Cell state number n = 27) (b), comparing inferences made at different timepoints against one another. More specifically, we computed the cell state transition probability from each detected cell state (with at least 50 cells profiled) to all 27 cell states (shown in Fig. 4a) for the last three time intervals (4-6h, 6-8h, 8-10h), and then compared these against one another (Point number n = 91, 77 and 90 for plots on the left, middle and right in a). The first two time points (0h and 2h) were excluded as they had few cell states (with at least 50 cells profiled) shared with other time points. (c) Cell state transition network (Cell state number = 27) similar to Fig. 4a (left), but after permuting the parent and child cell links (middle), or by similar approaches but based on mature mRNA only to link cells from different time points (right). For the permutation-based control (middle), each child cell of each time point was linked to a randomly selected parent cell from the immediately preceding time point. (d) The x and y coordinates correspond to UMAP visualization of cells from 0h and 2h treatment groups (Cell number n = 1,054 for 0h and 1,049 for 2h) colored by DEX treatment time (top) or inferred cell cycle stage (bottom). For both panels, black lines and arrows represent inferred cells state transitions by either RNA velocity (left), treatment time-informed RNA velocity (middle) or sci-fate analysis (right). Reference: Schofield, J. A., Duffy, E. E., Kiefer, L., Sullivan, M. C. & Simon, M. D. TimeLapse-seq: adding a temporal dimension to RNA sequencing through nucleoside recoding. Nat. Methods 15, 221–225 (2018).
About this article
Cite this article
Cao, J., Zhou, W., Steemers, F. et al. Sci-fate characterizes the dynamics of gene expression in single cells. Nat Biotechnol (2020). https://doi.org/10.1038/s41587-020-0480-9
Cell Stem Cell (2020)