CoChat

Multiplexed single-cell transcriptional response profiling to define cancer vulnerabilities and therapeutic mechanism of action

Open access

Published version of a preprint

The DOI matches its Crossref record; 77 of the 77 references it cites were found in a registry or an index.

Abstract

Assays to study cancer cell responses to pharmacologic or genetic perturbations are typically restricted to using simple phenotypic readouts such as proliferation rate. Information-rich assays, such as gene-expression profiling, have generally not permitted efficient profiling of a given perturbation across multiple cellular contexts. Here, we develop MIX-Seq, a method for multiplexed transcriptional profiling of post-perturbation responses across a mixture of samples with single-cell resolution, using SNP-based computational demultiplexing of single-cell RNA-sequencing data. We show that MIX-Seq can be used to profile responses to chemical or genetic perturbations across pools of 100 or more cancer cell lines. We combine it with Cell Hashing to further multiplex additional experimental conditions, such as post-treatment time points or drug doses. Analyzing the high-content readout of scRNA-seq reveals both shared and context-specific transcriptional response components that can identify drug mechanism of action and enable prediction of long-term cell viability from short-term transcriptional responses to treatment.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-020-17440-w

Cites

77 of 77 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Multiplexed transcriptional profiling across pools of cell lines. a Schematic diagram illustrating the MIX-Seq platform. b Heatmap showing likelihoods assigned by the SNP classification model for each cell coming from each parental cell line. The model consistently picks out a single cell line, from among the 24 “in-pool” cell lines, with high confidence. c UMAP representation of cells treated with DMSO control (gray) or nutlin (red) across a pool of 24 cell lines. Arrows indicate the shift in the population median coordinates for each cell line. d Heatmap showing average log fold-change estimates for each cell line for top differentially expressed genes. Nutlin sensitivity is given by 1 − area under dose response curve (AUC, see “Methods”). e Volcano plot showing strong gene expression changes in response to nutlin treatment across TP53 WT cell lines (n = 7). Effect size estimates and p values (not corrected for multiple comparisons) for this and subsequent differential expression analyses are estimated using the limma-trend pipeline49,50 (“Methods”). Vertical lines indicate a logFC threshold of 1. f Same as e for TP53 mutant cell lines (n = 17), showing little gene expression change in response to nutlin treatment. g Gene set analysis identifies gene sets that are upregulated (right) and downregulated (left) by nutlin treatment in the TP53 WT cell lines.
Fig. 1 Multiplexed transcriptional profiling across pools of cell lines. a Schematic diagram illustrating the MIX-Seq platform. b Heatmap showing likelihoods assigned by the SNP classification model for each cell coming from each parental cell line. The model consistently picks out a single cell line, from among the 24 “in-pool” cell lines, with high confidence. c UMAP representation of cells treated with DMSO control (gray) or nutlin (red) across a pool of 24 cell lines. Arrows indicate the shift in the population median coordinates for each cell line. d Heatmap showing average log fold-change estimates for each cell line for top differentially expressed genes. Nutlin sensitivity is given by 1 − area under dose response curve (AUC, see “Methods”). e Volcano plot showing strong gene expression changes in response to nutlin treatment across TP53 WT cell lines (n = 7). Effect size estimates and p values (not corrected for multiple comparisons) for this and subsequent differential expression analyses are estimated using the limma-trend pipeline49,50 (“Methods”). Vertical lines indicate a logFC threshold of 1. f Same as e for TP53 mutant cell lines (n = 17), showing little gene expression change in response to nutlin treatment. g Gene set analysis identifies gene sets that are upregulated (right) and downregulated (left) by nutlin treatment in the TP53 WT cell lines.CC BY · PubMed Central
Fig. 2 Viability-related and -independent transcriptional response components. a UMAP representation of single-cell expression profiles in a 99 cell line pool treated with vehicle control (gray) or trametinib (red). Arrows indicate trametinib-induced shift of population median coordinates for each cell line. b Histogram showing the number of cells recovered in each cell line and condition. c Volcano plot showing the viability-independent response for each gene, representing the “y-intercept” of a linear fit of expression change to drug sensitivity (1-AUC; “Methods”). Inset at left shows this relationship for an example gene: EGR1. The blue line shows the linear regression trend line (with the 95% CI interval shown in shaded gray). Inset below shows top upregulated (right, red) and downregulated (left, blue) gene sets. d Same as c for the viability-related response.
Fig. 2 Viability-related and -independent transcriptional response components. a UMAP representation of single-cell expression profiles in a 99 cell line pool treated with vehicle control (gray) or trametinib (red). Arrows indicate trametinib-induced shift of population median coordinates for each cell line. b Histogram showing the number of cells recovered in each cell line and condition. c Volcano plot showing the viability-independent response for each gene, representing the “y-intercept” of a linear fit of expression change to drug sensitivity (1-AUC; “Methods”). Inset at left shows this relationship for an example gene: EGR1. The blue line shows the linear regression trend line (with the 95% CI interval shown in shaded gray). Inset below shows top upregulated (right, red) and downregulated (left, blue) gene sets. d Same as c for the viability-related response.CC BY · PubMed Central
Fig. 3 Machine learning analysis powered by large-scale transcriptional profiling. a Accuracy of models trained to predict sensitivity for each individual drug, using either measured transcriptional responses or baseline “omics” features of the cell lines (using the same set of cell lines). Predictions based on transcriptional profiling at 6 and 24 h posttreatment are indicated by the gold and blue dots, respectively. b Plot of the Pearson correlation between each gene’s transcriptional response (24 h posttreatment) and drug sensitivity (1 − AUC) across cell lines and drugs, compared with its feature importance for the random forest predictive models. Gene set enrichment analysis of highlighted genes shown above. c Matrix of measured transcriptional responses across cell lines 24 h after trametinib treatment. d Eigenvalue spectrum of PCA applied to the matrix from c. e The projection of each cell line’s response onto PC1 is plotted against its measured trametinib sensitivity (red and green points indicate cell lines with activating BRAF or KRAS mutations respectively). Linear regression trend line (with the shaded gray region showing the 95% CI interval) is shown in blue. f Comparison of PC2 scores with expression of the melanoma-specific transcription factor SOX10 across cell lines. g UMAP embedding of transcriptional response profiles across drugs, cell lines, and posttreatment time points. Points are colored by treatment condition (drug and time point). Inset below shows zoomed view of the region indicated by the rectangle, with the larger green dots representing responses of BRAF mutant melanoma lines to the BRAFi dabrafenib.
Fig. 3 Machine learning analysis powered by large-scale transcriptional profiling. a Accuracy of models trained to predict sensitivity for each individual drug, using either measured transcriptional responses or baseline “omics” features of the cell lines (using the same set of cell lines). Predictions based on transcriptional profiling at 6 and 24 h posttreatment are indicated by the gold and blue dots, respectively. b Plot of the Pearson correlation between each gene’s transcriptional response (24 h posttreatment) and drug sensitivity (1 − AUC) across cell lines and drugs, compared with its feature importance for the random forest predictive models. Gene set enrichment analysis of highlighted genes shown above. c Matrix of measured transcriptional responses across cell lines 24 h after trametinib treatment. d Eigenvalue spectrum of PCA applied to the matrix from c. e The projection of each cell line’s response onto PC1 is plotted against its measured trametinib sensitivity (red and green points indicate cell lines with activating BRAF or KRAS mutations respectively). Linear regression trend line (with the shaded gray region showing the 95% CI interval) is shown in blue. f Comparison of PC2 scores with expression of the melanoma-specific transcription factor SOX10 across cell lines. g UMAP embedding of transcriptional response profiles across drugs, cell lines, and posttreatment time points. Points are colored by treatment condition (drug and time point). Inset below shows zoomed view of the region indicated by the rectangle, with the larger green dots representing responses of BRAF mutant melanoma lines to the BRAFi dabrafenib.CC BY · PubMed Central
Fig. 4 Population heterogeneity in pre- and post-perturbation transcriptional programs. a UMAP representation of gene expression profiles of DMSO (smaller points) and nutlin-treated (larger points) cell populations for a pool of 24 cell lines (as in Fig. 1). Cells are colored by their inferred cell cycle phase. TP53 WT cell lines (red arrows) show predominance of G0/G1-phase cells after nutlin treatment not observed in TP53 mutant cell lines (black arrows). b Quantification of the change in proportion of cells in G0/G1 in each cell line (y-axis) shows that nutlin treatment elicits G1 arrest selectively among the TP53 WT cell lines. Error bars show the 95% CI for estimates of the change in G0/G1 cell proportions. c Average change in the fraction of cells in each cell cycle phase for each drug treatment (averages are weighted by measured drug sensitivity, all for 24 h time points). d UMAP plot showing two subpopulations of RCC10RGB cells emerging 24 h after treatment with bortezomib. Dot color depicts inferred cell cycle phase and dot size represents treatment condition. e Heatmap showing LFC relative to control cells of top differentially expressed genes between the two clusters of bortezomib-treated cells for each of the 10/24 cell lines that showed a bimodal response pattern. In nearly all cases, one of the clusters (labeled cluster 2 by convention) was characterized by a predominance of S-phase cells, compared with mostly G0/G1 cells in cluster 1. f Volcano plot showing differentially expressed genes in cluster 2 vs cluster 1 averaged across the nine cell lines depicted in e. Genes that are part of the S-phase signature are highlighted in red.
Fig. 4 Population heterogeneity in pre- and post-perturbation transcriptional programs. a UMAP representation of gene expression profiles of DMSO (smaller points) and nutlin-treated (larger points) cell populations for a pool of 24 cell lines (as in Fig. 1). Cells are colored by their inferred cell cycle phase. TP53 WT cell lines (red arrows) show predominance of G0/G1-phase cells after nutlin treatment not observed in TP53 mutant cell lines (black arrows). b Quantification of the change in proportion of cells in G0/G1 in each cell line (y-axis) shows that nutlin treatment elicits G1 arrest selectively among the TP53 WT cell lines. Error bars show the 95% CI for estimates of the change in G0/G1 cell proportions. c Average change in the fraction of cells in each cell cycle phase for each drug treatment (averages are weighted by measured drug sensitivity, all for 24 h time points). d UMAP plot showing two subpopulations of RCC10RGB cells emerging 24 h after treatment with bortezomib. Dot color depicts inferred cell cycle phase and dot size represents treatment condition. e Heatmap showing LFC relative to control cells of top differentially expressed genes between the two clusters of bortezomib-treated cells for each of the 10/24 cell lines that showed a bimodal response pattern. In nearly all cases, one of the clusters (labeled cluster 2 by convention) was characterized by a predominance of S-phase cells, compared with mostly G0/G1 cells in cluster 1. f Volcano plot showing differentially expressed genes in cluster 2 vs cluster 1 averaged across the nine cell lines depicted in e. Genes that are part of the S-phase signature are highlighted in red.CC BY · PubMed Central
Fig. 5 Dual-multiplexed transcriptional profiling across cell lines and time points. a Schematic diagram illustrating experiment using Cell Hashing to multiplex scRNA-seq of cell line pools sampled at different time points following drug treatment. b UMAP plot showing 13,713 cells across a pool of 24 cell lines at different times following treatment with trametinib (shades of blue) or DMSO control (pink). c Single-cell expression levels of EGR1 at different time points following trametinib treatment for an example insensitive/sensitive cell line (left/right). Red dots depict the mean expression levels at each time point, and error bars show the interval +/− s.e.m. d Same as c, for MCM7. e (Top) Time course of the viability-independent response for top downregulated genes. (Bottom) Enrichment of HALLMARK_KRAS_SIGNALING_UP genes in the downregulated viability-independent response at each time point. f (Top) Same as e, showing time course of the viability-related response for top downregulated genes. (Bottom) Enrichment of HALLMARK_G2M_CHECKPOINT genes in the viability-related response at each time posttreatment. g Average time course of G0/G1 arrest across cell lines (n = 24 cell lines). Error bars indicate interval +/− s.e.m.
Fig. 5 Dual-multiplexed transcriptional profiling across cell lines and time points. a Schematic diagram illustrating experiment using Cell Hashing to multiplex scRNA-seq of cell line pools sampled at different time points following drug treatment. b UMAP plot showing 13,713 cells across a pool of 24 cell lines at different times following treatment with trametinib (shades of blue) or DMSO control (pink). c Single-cell expression levels of EGR1 at different time points following trametinib treatment for an example insensitive/sensitive cell line (left/right). Red dots depict the mean expression levels at each time point, and error bars show the interval +/− s.e.m. d Same as c, for MCM7. e (Top) Time course of the viability-independent response for top downregulated genes. (Bottom) Enrichment of HALLMARK_KRAS_SIGNALING_UP genes in the downregulated viability-independent response at each time point. f (Top) Same as e, showing time course of the viability-related response for top downregulated genes. (Bottom) Enrichment of HALLMARK_G2M_CHECKPOINT genes in the viability-related response at each time posttreatment. g Average time course of G0/G1 arrest across cell lines (n = 24 cell lines). Error bars indicate interval +/− s.e.m.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 77 references

Cited by

See all 204

Metadata

Authors
McFarland, J. M., Paolella, B. R., Warren, A., Geiger-Schuller, K., Shibue, T., Rothberg, M., Kuksenko, O., Colgan, W. N., Jones, A., Chambers, E., Dionne, D., Bender, S., Wolpin, B. M., Ghandi, M., Tirosh, I., Rozenblatt-Rosen, O., Roth, J. A., Golub, T. R., Regev, A., Aguirre, A. J., Vazquez, F., & Tsherniak, A.
Affiliations
  • Broad Institute (US)
  • Brigham and Women's Hospital (US)
Corresponding author
  • Aguirre, Andrew J.
  • Tsherniak, Aviad
Venue
Nature Communications
Published
2020-08-27
Volume
11
Issue
1
Pages
4296-4296
ISSN
2041-1723 (ISSN-L)
OpenAlex
W3080295962
Cited by
173
References
73
Citations by year
2026: 23 · 2025: 29 · 2024: 26 · 2023: 24 · 2022: 28 · 2021: 35 · 2020: 8
FWCI
7.70
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-nc)
Funders
  • Howard Hughes Medical Institute
  • Doris Duke Charitable Foundation
  • Pancreatic Cancer Action Network
  • Scheme for Promotion of Academic and Research Collaboration
  • National Institutes of Health (K08 CA21842002)
  • National Cancer Institute (P50CA127003, U01CA224146, CA224146, CA218420-02, U01CA250549, K08CA218420)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
Keywords
single-cell RNA sequencing, cancer cell lines, drug response, gene perturbation, cell hashing, mechanism of action, cancer vulnerability, drug dose-response
MeSH terms
Antineoplastic Agents, Base Sequence, Cell Survival, Humans, Neoplasms, Pyridones, Pyrimidinones, Models, Statistical, Gene Expression Regulation, Neoplastic, Polymorphism, Single Nucleotide, Gene Expression Profiling, Cell Line, Tumor, Single-Cell Analysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Multiplexed single-cell transcriptional response profiling to define cancer vulnerabilities and therapeutic mechanism of action. https://app.cochat.ai/papers/10.1038/s41467-020-17440-w
The paper
McFarland, J. M., Paolella, B. R., Warren, A., Geiger-Schuller, K., Shibue, T., Rothberg, M., Kuksenko, O., Colgan, W. N., Jones, A., Chambers, E., Dionne, D., Bender, S., Wolpin, B. M., Ghandi, M., Tirosh, I., Rozenblatt-Rosen, O., Roth, J. A., Golub, T. R., Regev, A., . . . Tsherniak, A. (2020). Multiplexed single-cell transcriptional response profiling to define cancer vulnerabilities and therapeutic mechanism of action. Nature Communications. https://doi.org/10.1038/s41467-020-17440-w

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.