Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Targeted Perturb-seq enables genome-scale genetic screens in single cells

Nature Methods volume 17, pages 629–635 (2020)Cite this article

Subjects

Abstract

The transcriptome contains rich information on molecular, cellular and organismal phenotypes. However, experimental and statistical limitations constrain sensitivity and throughput of genetic screening with single-cell transcriptomics readout. To overcome these limitations, we introduce targeted Perturb-seq (TAP-seq), a sensitive, inexpensive and platform-independent method focusing single-cell RNA-seq coverage on genes of interest, thereby increasing the sensitivity and scale of genetic screens by orders of magnitude. TAP-seq permits routine analysis of thousands of CRISPR-mediated perturbations within a single experiment, detects weak effects and lowly expressed genes, and decreases sequencing requirements by up to 50-fold. We apply TAP-seq to generate perturbation-based enhancer–target gene maps for 1,778 enhancers within 2.5% of the human genome. We thereby show that enhancer–target association is jointly determined by three-dimensional contact frequency and epigenetic states, allowing accurate prediction of enhancer targets throughout the genome. In addition, we demonstrate that TAP-seq can identify cell subtypes with only 100 sequencing reads per cell.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: TAP-seq permits efficient expression profiling of target genes in single cells.
Fig. 2: TAP-seq sensitively detects gene-expression changes.
Fig. 3: A perturbation-based screen of enhancer targets across 2.5% of the human genome.
Fig. 4: TAP-seq permits efficient identification of cell types and differentiation states at very low read depths.

Similar content being viewed by others

Data availability

All data are available from GEO GSE135497. A summary of all genomics data used for each figure is provided in Supplementary Table 5. ENCODE bulk RNA- and ChIP-seq data is available from encodeproject.org (experiment IDs: ENCSR545DKY, ENCSR000AKP, ENCSR000EWC, ENCSR000EWA, ENCSR000EWB, ENCSR388QZF, ENCSR921NMD). Hi-C data from ref. 39 are available from GEO GSE63525. Mouse BM single-cell RNA-seq data are available from GEO GSE122465.

Code availability

The TAP-seq R package for primer design is available through Bioconductor (http://bioconductor.org/packages/release/bioc/html/TAPseq.html). Code required to reproduce the analyses of this paper, as well as a pipeline for TAP-seq data processing, is available at https://github.com/argschwind/TAPseq_manuscript and https://github.com/argschwind/TAPseq_workflow.

References

  1. Steinmetz, L. M. et al. Systematic screen for human disease genes in yeast. Nat. Genet. 31, 400–404 (2002).

    Article  CAS  PubMed  Google Scholar 

  2. Winzeler, E. A. et al. Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science 285, 901–906 (1999).

    Article  CAS  PubMed  Google Scholar 

  3. Shifrut, E. et al. Genome-wide CRISPR screens in primary human T cells reveal key regulators of immune function. Cell 175, 1958–1971.e15 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Dixit, A. et al. Perturb-Seq: dissecting molecular circuits with scalable single-cell RNA profiling of pooled genetic screens. Cell 167, 1853–1866.e17 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Jaitin, D. A. et al. Dissecting immune circuits by linking CRISPR-pooled screens with single-cell RNA-seq. Cell 167, 1883–1896.e15 (2016).

    Article  CAS  PubMed  Google Scholar 

  6. Datlinger, P. et al. Pooled CRISPR screening with single-cell transcriptome readout. Nat. Methods 14, 297–301 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Xie, S., Duan, J., Li, B., Zhou, P. & Hon, G. C. Multiplexed engineering and analysis of combinatorial enhancer activity in single cells. Mol. Cell 66, 285–299.e5 (2017).

    Article  CAS  PubMed  Google Scholar 

  8. Adamson, B. et al. A multiplexed single-cell CRISPR screening platform enables systematic dissection of the unfolded protein response. Cell 167, 1867–1882.e21 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Gasperini, M. et al. A genome-wide framework for mapping gene regulation via cellular genetic screens. Cell 176, 377–390 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Van Der Wijst, M. G. P. et al. Single-cell RNA sequencing identifies celltype-specific cis-eQTLs and co-expression QTLs. Nat. Genet. 50, 493–497 (2018).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  11. van Galen, P. et al. Single-cell RNA-seq reveals AML hierarchies relevant to disease progression and immunity. Cell 176, 1265–1281.e24 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  12. Cuomo, A. S. et al. Single-cell RNA-sequencing of differentiating iPS cells reveals dynamic genetic effects on gene expression. Nat. Commun. 11, 810 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Sanyal, A., Lajoie, B. R., Jain, G. & Dekker, J. The long-range interaction landscape of gene promoters. Nature 489, 109–113 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Krivega, I. & Dean, A. Enhancer and promoter interactions—long distance calls. Curr. Opin. Genet. Dev. 22, 79–85 (2012).

    Article  CAS  PubMed  Google Scholar 

  15. Zheng, G. X. Y. et al. Massively parallel digital transcriptional profiling of single cells. Nat. Commun. 8, 14049 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Macosko, E. Z. et al. Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets. Cell 161, 1202–1214 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Horlbeck, M. A. et al. Compact and highly active next-generation libraries for CRISPR-mediated gene repression and activation. Elife 5, e19760 (2016).

  18. Hilton, I. B. et al. Epigenome editing by a CRISPR–Cas9-based acetyltransferase activates genes from promoters and enhancers. Nat. Biotechnol. 33, 510–517 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Fulco, C. P. et al. Systematic mapping of functional enhancer–promoter connections with CRISPR interference. Science 354, 769–773 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Hill, A. J. et al. On the design of CRISPR-based single-cell molecular screens. Nat. Methods 15, 271–274 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Zacher, B. et al. Accurate promoter and enhancer identification in 127 ENCODE and roadmap epigenomics cell types and tissues by GenoSTAN. PLoS One 12, e0169249 (2017).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  22. Klann, T. S. et al. CRISPR–Cas9 epigenome editing enables high-throughput screening for functional regulatory elements in the human genome. Nat. Biotechnol. 35, 561–568 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Tusi, B. K. et al. Population snapshots predict early haematopoietic and erythroid hierarchies. Nature 555, 54–60 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Paul, F. et al. Transcriptional heterogeneity and lineage commitment in myeloid progenitors. Cell 163, 1663–1677 (2015).

    Article  CAS  PubMed  Google Scholar 

  25. Baccin, C. et al. Combined single-cell and spatial transcriptomics reveal the molecular, cellular and spatial bone marrow niche organization. Nat. Cell Biol. 22, 38–48 (2020).

    Article  CAS  PubMed  Google Scholar 

  26. Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902 (2019).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Fan, H. C., Fu, G. K. & Fodor, S. P. A. Combinatorial labeling of single cells for gene expression cytometry. Science 347, 1258367 (2015).

    Article  PubMed  CAS  Google Scholar 

  28. Shum, E. Y., Walczak, E. M., Chang, C. & Fan, H. C. in Advances in Experimental Medicine and Biology 63–79 (Springer, 2019).

  29. Vallejo, A. F.et al. Resolving cellular systems by ultra-sensitive and economical single-cell transcriptome filtering. Preprint at bioRxiv https://doi.org/10.1101/800631 (2019).

  30. Uzbas, F. et al. BART-Seq: cost-effective massively parallelized targeted sequencing for genomics, transcriptomics, and single-cell analysis. Genome Biol. 20, 155 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  31. Dunn, S.-J., Martello, G., Yordanov, B., Emmott, S. & Smith, A. G. Defining an essential transcription factor program for naïve pluripotency. Science 344, 1156–1160 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Horn, T. et al. Mapping of signaling networks through synthetic genetic interaction analysis by RNAi. Nat. Methods 8, 341–346 (2011).

    Article  CAS  PubMed  Google Scholar 

  33. Zetsche, B. et al. Multiplex gene editing by CRISPR–Cpf1 using a single crRNA array. Nat. Biotechnol. 35, 31–34 (2016).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  34. Subramanian, A. et al. A next generation connectivity Map: L1000 platform and the first 1,000,000 profiles. Cell 171, 1437–1452.e17 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Kearns, N. A. et al. Functional annotation of native enhancers with a Cas9-histone demethylase fusion. Nat. Methods 12, 401–403 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Li, K. et al. Interrogation of enhancer function by enhancer-targeting CRISPR epigenetic editing. Nat. Commun. 11, 485 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  37. 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).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Datlinger, P. et al. Ultra-high throughput single-cell RNA sequencing by combinatorial fluidic indexing. Preprint at bioRxiv https://doi.org/10.1101/2019.12.17.879304 (2019).

  39. Rao, S. S. P. et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 159, 1665–1680 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Dunham, I. et al. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012).

    Article  CAS  Google Scholar 

  41. Chen, S. et al. Genome-wide CRISPR screen in a mouse model of tumor growth and metastasis. Cell 160, 1246–1260 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Mendenhall, E. M. et al. Locus-specific editing of histone modifications at endogenous enhancers. Nat. Biotechnol. 31, 1133–1136 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Deliu, E. et al. Haploinsufficiency of the intellectual disability gene SETD5 disturbs developmental gene expression and cognition. Nat. Neurosci. 21, 1717–1727 (2018).

    Article  CAS  PubMed  Google Scholar 

  44. Velten, L. et al. Single-cell polyadenylation site mapping reveals 3′ isoform choice variability. Mol. Syst. Biol. 11, 812 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  45. Xie, T. et al. Single-cell deconvolution of fibroblast heterogeneity in mouse pulmonary fibrosis article Single-Cell deconvolution of fibroblast heterogeneity in mouse pulmonary fibrosis. Cell Rep. 22, 3625–3640 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Untergasser, A. et al. Primer3—new capabilities and interfaces. Nucleic Acids Res. 40, e115 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Koressaar, T. & Remm, M. Enhancements and modifications of primer design program Primer3. Bioinformatics 23, 1289–1291 (2007).

    Article  CAS  PubMed  Google Scholar 

  48. Schraivogel, D., Velten, L., Gschwind, A. R. & Steinmetz, L. M. A protocol for Targeted Perturb (TAP)-seq and targeted single-cell RNA-seq. Protoc. Exch. https://doi.org/10.21203/rs.3.pex-864/v1 (2020).

  49. Tilgner, H. et al. Microfluidic isoform sequencing shows widespread splicing coordination in the human transcriptome. Genome Res. 28, 231–242 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. Wickham, H. ggplot2: Elegant Graphics for Data Analysis. (Springer, 2009).

  51. Santos, A., Wernersson, R. & Jensen, L. J. Cyclebase 3.0: a multi-organism database on cell-cycle regulation and phenotypes. Nucleic Acids Res. 43, D1140–D1144 (2015).

    Article  CAS  PubMed  Google Scholar 

  52. Finak, G. et al. MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome Biol. 16, 278 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  53. Miao, Z., Deng, K., Wang, X. & Zhang, X. DEsingle for detecting three types of differential expression in single-cell RNA-seq data. Bioinformatics 34, 3223–3224 (2018).

    Article  CAS  PubMed  Google Scholar 

  54. Korthauer, K. D. et al. A statistical approach for identifying differential distributions in single-cell RNA-seq experiments. Genome Biol. 17, 222 (2016).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  55. Li, P. et al. Coordinated regulation of IFITM1, 2 and 3 genes by an IFN-responsive enhancer through long-range chromatin interactions. Biochim. Biophys. Acta Gene Regul. Mech. 1860, 885–893 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

This work was supported by grants from the European Research Council Advanced Investigator Grant (AdG-294542 and AdG-742804 to L.M.S.) and the Emerson Collective (award 643577 to L.V. and L.M.S.). D.S. was supported by a fellowship from the EMBL Interdisciplinary Postdoc (EIPOD) program under Marie Sklodowska-Curie Actions COFUND (grant agreement number 664726). A.R.G. was supported by an Early Postdoc.Mobility fellowship (project number P2LAP3_171806) from the Swiss National Science Foundation (SNSF). We thank P. Collier, I. Gupta, H. Tilgner and D. Pavlinic for advice on 10X Chromium; S. Vonesch, K. Roy and J. Smith for advice on gRNA library cloning; V. Benes and D. Pavlinic for Illumina sequencing; M. Paulsen and team for flow cytometry service; S. Haas, J. Al-Sabah, C. Baccin, L. Ballenberger, L. Martins, C. Scholl and K.-M. Noh for providing cell samples; and A. Rabinowitz for advice on the analysis of Hi-C data.

Author information

Author notes

  1. Lars Velten

    Present address: Center for Genomic Regulation, The Barcelona Institute of Science and Technology, Barcelona, Spain

  2. These authors contributed equally: Daniel Schraivogel, Andreas R. Gschwind.

Authors and Affiliations

  1. European Molecular Biology Laboratory (EMBL), Genome Biology Unit, Heidelberg, Germany

    Daniel Schraivogel, Jennifer H. Milbank, Daniel R. Leonce, Petra Jakob, Lukas Mathur, Jan O. Korbel, Christoph A. Merten, Lars Velten & Lars M. Steinmetz

  2. Department of Genetics, Stanford University School of Medicine, Stanford, CA, USA

    Andreas R. Gschwind & Lars M. Steinmetz

  3. Stanford Genome Technology Center, Palo Alto, CA, USA

    Lars M. Steinmetz

Authors
  1. Daniel Schraivogel
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Andreas R. Gschwind
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jennifer H. Milbank
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Daniel R. Leonce
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Petra Jakob
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Lukas Mathur
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Jan O. Korbel
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Christoph A. Merten
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Lars Velten
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Lars M. Steinmetz
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

D.S., L.V., A.R.G. and L.M.S. conceptualized the project, with contributions by J.O.K. D.S., D.R.L., P.J. and J.H.M. performed experiments. A.R.G. developed primer design pipeline. A.R.G. and L.V. performed data analysis. L.M. and C.M. implemented Drop-seq. D.S., A.R.G., L.V. and L.M.S. wrote the manuscript. All authors commented on the manuscript.

Corresponding authors

Correspondence to Lars Velten or Lars M. Steinmetz.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Lei Tang was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Sample bioanalyzer traces for libraries from TAP-seq and 10X Genomics.

a, Standard 10X Genomics protocol. b, TAP-seq library using panel 1 and cDNA from 10X Genomics K562 cells as input material. Strong peaks in TAP-seq profile correspond to highly expressed genes in the primer panel (HBG1, HBG2, HBE1), as validated by sub-cloning of bands and Sanger sequencing (not shown). c-f, Remaining target gene panels applied to different cell types and cell lines.

Extended Data Fig. 2 Choice of target genes and single cell capture platform.

a, Left panel: Expression ranks of genes used for the three test panels. An x axis value of 1 refers to highest expression. See y axis labels for number of target genes, and refer to Methods section on ‘data visualization’ for a description of violin plot elements. Right panel: Fraction of the transcriptome covered by these panels, computed from a whole transcriptome reference data set from the same cell type (y axis labels). For panels 1+2, K562 cells were used, panel 3 was applied to mouse embryonic stem cells (ESCs), mouse 3T3 cells, mouse neutrophils (Neutr.), and mouse lung mesenchymal cells (Lung). b, Mean gene expression levels for 10X Genomics and Drop-seq based TAP-seq. Panel 1 was used in both cases. n = 684 genes are shown. c, Number of UMIs observed per cell for both TAP-seq and whole transcriptome readout (Whole Tx), using 10X Genomics or Drop-seq for RNA capture and reverse transcription. Experiments were downsampled to an average of 1,000 reads per cell.

Extended Data Fig. 3 Analysis of the L1000 panel.

a, Spearman correlation in mean gene expression levels between TAP-seq and whole transcriptome readout (Whole Tx) for a panel targeting the L1000 gene set34. n = 6,963 genes were covered in TAP-seq and are included in the plot. b, Principal component analysis of the TAP-seq dataset and the whole transcriptome dataset. Principal component loadings of all genes annotated in cyclebase v351 are shown, with the peak-time of expression color-coded. In the whole transcriptome dataset, PC1-3 were not significantly associated with GO-terms (not shown). n = 68,734 cells (TAP-seq) and 8,282 cells (Whole Tx).

Extended Data Fig. 4 Analysis of library complexity in TAP-seq and whole transcriptome 10x Genomics.

a, Deeply sequenced TAP-seq and whole transcriptome (Whole Tx) libraries were downsampled to a given average number of reads per cell (x axis). The average number of UMIs observed on the target panel (solid lines, shown for both methods) or across the entire genome (dashed line, only shown for whole transcriptome readout) is shown. See also Fig. 1e. b, Deeply sequenced TAP-seq and whole transcriptome libraries were down-sampled to a given average number of reads per cell (x axis). The ratio in UMIs observed on the target gene panel between TAP-seq and whole transcriptome sequencing is plotted as a measure of enrichment efficiency c, For K562 cells and panel 1, gene detection levels were compared between genes of different expression levels. See also Fig. 1f. d, Number of molecules observed per cell in different cell types at 160,000 reads per cell. n = 6,109 3T3 cells, 160 ESCs, 130 Lung cells and 55 Neutrophils. See Methods section on ‘data visualization’ for a description of box plot elements.

Extended Data Fig. 5 Analysis of reproducibility in TAP-seq and whole transcriptome 10X Genomics.

a, Pearson correlation in mean gene expression levels across all genes of panel 2 (n = 674 genes) between three biological replicates. b, Pearson correlation between whole transcriptome 10X Genomics and TAP-seq for various panels and cell lines/cell types (see Extended Data Fig. 2a for number of genes per panel). c, Pearson correlation between whole transcriptome 10X Genomics and bulk RNA-seq (GEO: GSM2343836), across the n = 684 genes of panel 1 and the n = 674 genes of panel 2.

Extended Data Fig. 6 Technical properties of the ground truth perturbation experiment.

a, Gene expression level in K562 cells of the various gRNA target genes used. b, Enhancer gRNAs were validated by pooled transduction of K562 dCas9-KRAB cells with all four enhancer-targeting guides, and the effect on target gene expression was quantified by qPCR. HBE1 was analyzed as target gene for the HS2 enhancer. n = 63 replicates. c, Histogram of the number of gRNAs identified per cell in the TAP-seq experiment of Fig. 2. d, The number of gRNAs observed per cell (see also in c) was fitted with a generative model of gRNA capture efficiency and multiplicity of infection4,20. Log-likelihood is plotted as a function of the parameters; the maximum likelihood estimate is marked by a cross. Data from n = 621,977 (TAP-seq), n = 67,994 (Perturb-Seq) or n = 637,971 cells (Perturb-seq + gRNA amp.) was used. e, Mean expression per gene for whole transcriptome 10X Genomics compared to TAP-seq, with perturbation target genes highlighted. n = 674 genes from panel 2 are shown. Two genes for which perturbation effects were detected with a lower efficiency in TAP-seq are highlighted in red.

Extended Data Fig. 7 Comparison of differential expression testing methods.

a, Comparison using Precision-Recall curves, as in Fig. 2f. TAP-seq data were downsampled to 10, 25, 50 or 100 cells per gRNA. For each sampling run, differential expression testing was performed using a simple (two-sided) Wilcoxon test, MAST52, DEsingle53 and scDD54, as well as MAST with the number of genes observed as an additional covariate. Precision-Recall curves were computed assuming that the intended gRNA targets constitute the full set of true positives. Data were normalized across cells using the censored mean, that is division with the mean expression of all genes not part of the highest decile. b, Performance comparison in terms of area under the Precision-Recall curve for different data normalization strategies and tests. c, Performance comparison in terms of area under the ROC curve.

Extended Data Fig. 8 Additional analyses of the ground truth perturbation dataset.

a, Precision-Recall curves, as in Fig. 2f. Potentially true gRNA off-target or downstream effects were identified by differential expression testing across all cells, and then excluded from the analysis. Points indicate performance at a nominal FDR of 0.05. See Note S3 section ‘Sensitivity analysis (differential expression)’ for detail on the statistical test used. b, Comparison of Area under the precision-recall curves (AUPRC) for n = 6,100 cells per perturbation, sampled to various read depths. Potential gRNA off-target and downstream effects were treated as false positives (solid lines, same as in Fig. 2g) or excluded (dashed lines). c, The absolute effect of a gRNA-mediated perturbation in UMIs/cell was quantified from non-downsampled whole transcriptome data (x Axis). The probability of observing these effects as significant was the quantified by drawing 100 samples using 150 cells per sample and 1,000 average reads per cell (y Axis). Lines derive from a logistic regression. The UMI difference required for achieving a 50% detection probability was used as a measure of molecular sensitivity (dotted line). Data from n = 660,106 cells and 9,750 sampling runs. d, Like Fig. 2g, but using molecular sensitivity as defined in panel c as the measure of sensitivity. Down-sampling was restricted to 50–150 cells per perturbation, since estimates of molecular sensitivity were otherwise driven by excessive sampling noise. Data from n = 660,106 cells and 7,150 sampling runs. e, AUPRC plotted in relationship to number of cells per perturbation and total number of reads (data from Fig. 2g). f, For of n = 656 each gRNA targets, the absolute and relative expression change elicited by the perturbation, as well as the expression baseline, were computed from whole transcriptome data without subsampling (x axis). Data from both methods were then downsampled repeatedly to 150 cells per perturbation and 10,000 (Perturb-seq) or 1,000 (TAP-seq) reads per cell to determine the probability of detecting a change (y axis). Refer to methods section on ‘data visualization’ for a definition of box plot elements.

Extended Data Fig. 9 Additional analyses of the enhancer screen.

a, Number of detected gRNAs/perturbations per cell were plotted. b, Levenshtein edit distance between the consensus sequence of a gRNA in a given cell, and the template sequence, showing that in 93 % (chr. 8) or 95 % (chr. 11) of cases, there were no mismatches between consensus and template. c, Fold change in gene expression of enhancer targets is plotted in relation to the number of gRNAs supporting an enhancer-target gene pair (ETP). Number of ETPs per confidence level: 0 = 61, 1 = 621, 2 = 612, 3 = 611, 4 = 611. d, Zoom-in on a region surrounding the IFITM locus shows identification previously known enhancers55. e, Distance to transcription start site (TSS) was plotted against confidence level, as calculated from the number of individual gRNAs with a detected effect on the target gene. Number of ETPs per confidence level: 0 = 61, 1 = 621, 2 = 612, 3 = 611, 4 = 611. See Methods section on ‘data visualization’ for a definition of boxplot elements. f, Number of genes jumped between an enhancer and the identified target gene was plotted (main panel). Inset shows association strength, calculated from the proportion of gRNAs that support the ETP, plotted against the number of jumped genes. g, Histogram of log-fold expression differences between jumped genes and the respective true enhancer target. h, Like Fig. 3h, but including all 34,493 potential ETPs across the whole dataset, instead of just gene-proximal ETPs. i, Precision-Recall curves for classifiers trained on the dataset generated in this study, and applied to the dataset from ref. 9 (orange line), or classifiers trained on the dataset from ref. 9 and applied to this dataset.

Extended Data Fig. 10 Additional analyses of the mouse bone marrow experiment.

a, Heatmap depicting the expression of all 182 target genes across 11,794 cells, as measured by TAP-seq. Top row (‘Cluster’) depicts the result of unsupervised clustering, second row (‘Projection’) depicts the result of transferring labels26 from the whole transcriptome reference data set (see Fig. 4a for color code). b, Gene expression correlations across populations. Mean gene expression for each gene in the mouse bone marrow panel was computed for each cell type, and the Pearson correlation between TAP-seq and whole transcriptome readout (Whole Tx) across n = 618 cell types was computed. Main panel shows Pearson correlation coefficients for all tested genes across cell types. Inset shows expression of IFITM2 as measured by TAP-seq and whole transcriptome readout for each cell type (color code as described in Fig. 4a). c, Data from both methods were downsampled to various average read depths and an identical number of cells, and labels were transferred26 from the non-downsampled reference. For each cell type plotted on the x axis, the fraction of cells projected to the cell types plotted on the y axis was quantified (color code as described in Fig. 4a). d, Average read depth per cell is plotted against the fraction of cells correctly classified. e, The fold difference in sequencing reads between TAP-seq and whole transcriptome is plotted as a function of the fraction of cells correctly classified.

Supplementary information

Supplementary Information

Supplementary Notes 1–3 and Supplementary Protocol.

Reporting Summary

Supplementary Table 1

List of primer panel sequences used for TAP-seq. For each target gene, the sequences as well as the genome-wide rank in gene expression (from whole-transcriptome 10X Genomics) are shown.

Supplementary Table 2

List of gRNA sequences used for enhancer screens.

Supplementary Table 3

Significant enhancer-target pairs identified in this study. See Methods section ‘Enhancer screen analysis’ for description of the statistical test used to compute P values. A total of n = 231,667 cells were included into the tests.

Supplementary Table 4

Pairwise comparisons of chromatin activity between enhancers from strong, weak and non-significant ETPs. P values were calculated using Wilcoxon signed-rank tests. See Supplementary Note 3 section ‘Hi-C and chromatin analyses’ for description of chromatin activity and selection of non-significant ETPs.

Supplementary Table 5

Overview of sequencing datasets generated.

Supplementary Table 6

Oligonucleotide sequences for TAP-seq.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Schraivogel, D., Gschwind, A.R., Milbank, J.H. et al. Targeted Perturb-seq enables genome-scale genetic screens in single cells. Nat Methods 17, 629–635 (2020). https://doi.org/10.1038/s41592-020-0837-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41592-020-0837-5

This article is cited by

Search

Advanced search

Quick links

Nature Briefing: Translational Research

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

Get what matters in translational research, free to your inbox weekly. Sign up for Nature Briefing: Translational Research