Synergistic drug combinations for cancer identified in a CRISPR screen for pairwise genetic interactions


Identification of effective combination therapies is critical to address the emergence of drug-resistant cancers, but direct screening of all possible drug combinations is infeasible. Here we introduce a CRISPR-based double knockout (CDKO) system that improves the efficiency of combinatorial genetic screening using an effective strategy for cloning and sequencing paired single guide RNA (sgRNA) libraries and a robust statistical scoring method for calculating genetic interactions (GIs) from CRISPR-deleted gene pairs. We applied CDKO to generate a large-scale human GI map, comprising 490,000 double-sgRNAs directed against 21,321 pairs of drug targets in K562 leukemia cells and identified synthetic lethal drug target pairs for which corresponding drugs exhibit synergistic killing. These included the BCL2L1 and MCL1 combination, which was also effective in imatinib-resistant cells. We further validated this system by identifying known and previously unidentified GIs between modifiers of ricin toxicity. This work provides an effective strategy to screen synergistic drug combinations in high-throughput and a CRISPR-based tool to dissect functional GI networks.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Figure 1: Development of a CDKO platform to identify novel cancer drug combinations in high throughput.
Figure 2: 490,000-element DrugTarget-CDKO library can be efficiently assembled and delivered into cells and shows no positional bias.
Figure 3: Strategy to calculate quantitative GI scores.
Figure 4: A CRISPR-based GI map of ricin pathway regulators validates the CDKO platform.
Figure 5: Validation of predicted GIs using individual sgRNAs.
Figure 6: Predicted synergistic gene pairs translate to synergistic drug combinations.

Accession codes

Primary accessions


Sequence Read Archive


  1. 1

    Jia, J. et al. Mechanisms of drug combinations: interaction and network perspectives. Nat. Rev. Drug Discov. 8, 111–128 (2009).

    Article  CAS  Google Scholar 

  2. 2

    Ashburn, T.T. & Thor, K.B. Drug repositioning: identifying and developing new uses for existing drugs. Nat. Rev. Drug Discov. 3, 673–683 (2004).

    Article  CAS  Google Scholar 

  3. 3

    Al-Lazikani, B., Banerji, U. & Workman, P. Combinatorial drug therapy for cancer in the post-genomic era. Nat. Biotechnol. 30, 679–692 (2012).

    Article  CAS  Google Scholar 

  4. 4

    Sun, X., Vilar, S. & Tatonetti, N.P. High-throughput methods for combinatorial drug discovery. Sci. Transl. Med. 5, 205rv1 (2013).

    Article  CAS  Google Scholar 

  5. 5

    Collins, S.R., Weissman, J.S. & Krogan, N.J. From information to knowledge: new technologies for defining gene function. Nat. Methods 6, 721–723 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. 6

    Costanzo, M. et al. A global genetic interaction network maps a wiring diagram of cellular function. Science (2016).

  7. 7

    Schuldiner, M. et al. Exploration of the function and organization of the yeast early secretory pathway through an epistatic miniarray profile. Cell 123, 507–519 (2005).

    Article  CAS  Google Scholar 

  8. 8

    Jonikas, M.C. et al. Comprehensive characterization of genes required for protein folding in the endoplasmic reticulum. Science 323, 1693–1697 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. 9

    Bandyopadhyay, S. et al. Rewiring of genetic networks in response to DNA damage. Science 330, 1385–1389 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. 10

    Collins, S.R. et al. Functional dissection of protein complexes involved in yeast chromosome biology using a genetic interaction map. Nature 446, 806–810 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. 11

    Dixon, S.J., Costanzo, M., Baryshnikova, A., Andrews, B. & Boone, C. Systematic mapping of genetic interaction networks. Annu. Rev. Genet. 43, 601–625 (2009).

    Article  CAS  Google Scholar 

  12. 12

    Frost, A. et al. Functional repurposing revealed by comparing S. pombe and S. cerevisiae genetic interactions. Cell 149, 1339–1352 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. 13

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

    CAS  Article  Google Scholar 

  14. 14

    Roguev, A. et al. Quantitative genetic-interaction mapping in mammalian cells. Nat. Methods 10, 432–437 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. 15

    Bassik, M.C. et al. A systematic mammalian genetic interaction map reveals pathways underlying ricin susceptibility. Cell 152, 909–922 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. 16

    Srivas, R. et al. A network of conserved synthetic lethal interactions for exploration of precision cancer therapy. Mol. Cell 63, 514–525 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. 17

    Wong, A.S.L., Choi, G.C.G., Cheng, A.A., Purcell, O. & Lu, T.K. Massively parallel high-order combinatorial genetics in human cells. Nat. Biotechnol. 33, 952–961 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. 18

    Jinek, M. et al. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337, 816–821 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

  19. 19

    Mali, P. et al. RNA-guided human genome engineering via Cas9. Science 339, 823–826 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. 20

    Cong, L. et al. Multiplex genome engineering using CRISPR/Cas systems. Science 339, 819–823 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. 21

    Tsai, S.Q. et al. Dimeric CRISPR RNA-guided FokI nucleases for highly specific genome editing. Nat. Biotechnol. 32, 569–576 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. 22

    Xie, K., Minkenberg, B. & Yang, Y. Boosting CRISPR/Cas9 multiplex editing capability with the endogenous tRNA-processing system. Proc. Natl. Acad. Sci. USA 112, 3570–3575 (2015).

    Article  CAS  Google Scholar 

  23. 23

    Wong, A.S.L. et al. Multiplexed barcoded CRISPR-Cas9 screening enabled by CombiGEM. Proc. Natl. Acad. Sci. USA 113, 2544–2549 (2016).

    Article  CAS  Google Scholar 

  24. 24

    Vidigal, J.A. & Ventura, A. Rapid and efficient one-step generation of paired gRNA CRISPR-Cas9 libraries. Nat. Commun. 6, 8083 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. 25

    Sack, L.M., Davoli, T., Xu, Q., Li, M.Z. & Elledge, S.J. Sources of error in mammalian genetic screens. G3 (Bethesda) 6, 2781–2790 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. 26

    Zhu, F. et al. Therapeutic target database update 2012: a resource for facilitating target-oriented drug discovery. Nucleic Acids Res. 40, D1128–D1136 (2012).

    Article  CAS  Google Scholar 

  27. 27

    Wishart, D.S. et al. DrugBank: a comprehensive resource for in silico drug discovery and exploration. Nucleic Acids Res. 34, D668–D672 (2006).

    Article  CAS  Google Scholar 

  28. 28

    Pawson, A.J. et al. The IUPHAR/BPS Guide to PHARMACOLOGY: an expert-driven knowledgebase of drug targets and their ligands. Nucleic Acids Res. 42, D1098–D1106 (2014).

    Article  CAS  Google Scholar 

  29. 29

    Morgens, D.W. et al. Genome-scale measurement of off-target activity using Cas9 toxicity in high-throughput screens. Nat. Commun (in press). (2017).

  30. 30

    Deans, R.M. et al. Parallel shRNA and CRISPR-Cas9 screens enable antiviral drug target identification. Nat. Chem. Biol. 12, 361–366 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. 31

    Kampmann, M., Bassik, M.C. & Weissman, J.S. Integrated platform for genome-wide screening and construction of high-density genetic interaction maps in mammalian cells. Proc. Natl. Acad. Sci. USA 110, E2317–E2326 (2013).

    Article  PubMed  PubMed Central  Google Scholar 

  32. 32

    Gilbert, L.A. et al. Genome-scale CRISPR-mediated control of gene repression and activation. Cell 159, 647–661 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. 33

    Collins, S.R., Schuldiner, M., Krogan, N.J. & Weissman, J.S. A strategy for extracting and analyzing large-scale quantitative epistatic interaction data. Genome Biol. 7, R63 (2006).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. 34

    Fischer, B. et al. A map of directional genetic interactions in a metazoan cell. eLife 4, 1–21 (2015).

    Google Scholar 

  35. 35

    Blomen, V.A. et al. Gene essentiality and synthetic lethality in haploid human cells. Science 350, 1092–1096 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. 36

    Brazil, D.P., Yang, Z.Z. & Hemmings, B.A. Advances in protein kinase B signalling: AKTion on multiple fronts. Trends Biochem. Sci. 29, 233–242 (2004).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. 37

    Saurabh, K. et al. The PIM family of oncoproteins: small kinases with huge implications in myeloid leukemogenesis and as therapeutic targets. Oncotarget 5, 8503–8514 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  38. 38

    Chen, J.L., Limnander, A. & Rothman, P.B. Pim-1 and Pim-2 kinases are required for efficient pre-B-cell transformation by v-Abl oncogene. Blood 111, 1677–1685 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. 39

    Johannes, L. & Popoff, V. Tracing the retrograde route in protein trafficking. Cell 135, 1175–1187 (2008).

    Article  CAS  Google Scholar 

  40. 40

    Moreau, D. et al. Genome-wide RNAi screens identify genes required for Ricin and PE intoxications. Dev. Cell 21, 231–244 (2011).

    Article  CAS  Google Scholar 

  41. 41

    Morgens, D.W., Deans, R.M., Li, A. & Bassik, M.C. Systematic comparison of CRISPR/Cas9 and RNAi screens for essential genes. Nat. Biotechnol. 34, 634–636 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. 42

    Wang, W. et al. Mannosidase 2, alpha 1 deficiency is associated with ricin resistance in embryonic stem (ES) cells. PLoS One 6, e22993 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. 43

    Bar-Peled, L. et al. A Tumor suppressor complex with GAP activity for the Rag GTPases that signal amino acid sufficiency to mTORC1. Science 340, 1100–1106 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. 44

    Margueron, R. & Reinberg, D. The Polycomb complex PRC2 and its mark in life. Nature 469, 343–349 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. 45

    Szklarczyk, D. et al. STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 43, D447–D452 (2015).

    Article  CAS  Google Scholar 

  46. 46

    Parnas, O. et al. A genome-wide CRISPR screen in primary immune cells to dissect regulatory networks. Cell 162, 675–686 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. 47

    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 

  48. 48

    Brinkman, E.K., Chen, T., Amendola, M. & van Steensel, B. Easy quantitative assessment of genome editing by sequence trace decomposition. Nucleic Acids Res. 42, e168 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. 49

    Lord, C.J., Tutt, A.N.J. & Ashworth, A. Synthetic lethality and cancer therapy: lessons learned from the development of PARP inhibitors. Annu. Rev. Med. 66, 455–470 (2015).

    Article  CAS  Google Scholar 

  50. 50

    Sultana, R. et al. Synthetic lethal targeting of DNA double-strand break repair deficient cells by human apurinic/apyrimidinic endonuclease inhibitors. Int. J. Cancer 131, 2433–2444 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. 51

    Riabinska, A. et al. Therapeutic targeting of a robust non-oncogene addiction to PRKDC in ATM-defective tumors. Sci. Transl. Med. 5, 189ra78 (2013).

    Article  CAS  Google Scholar 

  52. 52

    Austin, C.J.D., Kahlert, J., Kassiou, M. & Rendina, L.M. The translocator protein (TSPO): a novel target for cancer chemotherapy. Int. J. Biochem. Cell Biol. 45, 1212–1216 (2013).

    Article  CAS  Google Scholar 

  53. 53

    Placzek, W.J. et al. A survey of the anti-apoptotic Bcl-2 subfamily expression in cancer types provides a platform to predict the efficacy of Bcl-2 antagonists in cancer therapy. Cell Death Dis. 1, e40 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. 54

    Bixby, D. & Talpaz, M. Seeking the causes and solutions to imatinib-resistance in chronic myeloid leukemia. Leukemia 25, 7–22 (2011).

    Article  CAS  Google Scholar 

  55. 55

    Delbridge, A.R.D., Grabow, S., Strasser, A. & Vaux, D.L. Thirty years of BCL-2: translating cell death discoveries into novel cancer therapies. Nat. Rev. Cancer 16, 99–109 (2016).

    Article  CAS  Google Scholar 

  56. 56

    Aichberger, K.J. et al. Identification of mcl-1 as a BCR/ABL-dependent target in chronic myeloid leukemia (CML): evidence for cooperative antileukemic effects of imatinib and mcl-1 antisense oligonucleotides. Blood 105, 3303–3311 (2005).

    Article  CAS  Google Scholar 

  57. 57

    Dai, Y., Rahmani, M., Corey, S.J., Dent, P. & Grant, S. A Bcr/Abl-independent, Lyn-dependent form of imatinib mesylate (STI-571) resistance is associated with altered expression of Bcl-2. J. Biol. Chem. 279, 34227–34239 (2004).

    Article  CAS  Google Scholar 

  58. 58

    Berns, K. et al. A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 428, 431–437 (2004).

    Article  CAS  Google Scholar 

  59. 59

    Friedman, A. & Perrimon, N. A functional RNAi screen for regulators of receptor tyrosine kinase and ERK signalling. Nature 444, 230–234 (2006).

    Article  CAS  Google Scholar 

  60. 60

    Mendes-Pereira, A.M. et al. Genome-wide functional screen identifies a compendium of genes affecting sensitivity to tamoxifen. Proc. Natl. Acad. Sci. USA 109, 2730–2735 (2012).

    Article  Google Scholar 

  61. 61

    Eichhorn, P.J.A. et al. Phosphatidylinositol 3-kinase hyperactivation results in lapatinib resistance that is reversed by the mTOR/phosphatidylinositol 3-kinase inhibitor NVP-BEZ235. Cancer Res. 68, 9221–9230 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. 62

    Corcoran, R.B. et al. Synthetic lethal interaction of combined BCL-XL and MEK inhibition promotes tumor regressions in KRAS mutant cancer models. Cancer Cell 23, 121–128 (2013).

    Article  CAS  Google Scholar 

  63. 63

    Prahallad, A. et al. Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature 483, 100–103 (2012).

    Article  CAS  Google Scholar 

  64. 64

    Opferman, J.T. Attacking cancer's Achilles heel: antagonism of anti-apoptotic BCL-2 family members. FEBS J. 283, 2661–2675 (2016).

    Article  CAS  Google Scholar 

  65. 65

    Leverson, J.D. et al. Potent and selective small-molecule MCL-1 inhibitors demonstrate on-target cancer cell killing activity as single agents and in combination with ABT-263 (navitoclax). Cell Death Dis. 6, e1590 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  66. 66

    Chonghaile, T.N. et al. Maturation stage of T-cell acute lymphoblastic leukemia determines BCL-2 versus BCL-XL dependence and sensitivity to ABT-199. Cancer Discov. 4, 1074–1087 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  67. 67

    Ko, T.K., Chuah, C.T.H., Huang, J.W.J., Ng, K.-P. & Ong, S.T. The BCL2 inhibitor ABT-199 significantly enhances imatinib-induced cell death in chronic myeloid leukemia progenitors. Oncotarget 5, 9033–9038 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  68. 68

    Hong, H.-Y., Choi, J., Cho, Y.-W. & Kim, B.-C. Cdc25A promotes cell survival by stimulating NF-κB activity through IκB-α phosphorylation and destabilization. Biochem. Biophys. Res. Commun. 420, 293–296 (2012).

    Article  CAS  Google Scholar 

  69. 69

    Bassi, Z.I., Audusseau, M., Riparbelli, M.G., Callaini, G. & D'Avino, P.P. Citron kinase controls a molecular network required for midbody formation in cytokinesis. Proc. Natl. Acad. Sci. USA 110, 9782–9787 (2013).

    Article  Google Scholar 

  70. 70

    Chen, B. et al. Dynamic imaging of genomic loci in living human cells by an optimized CRISPR/Cas system. Cell 155, 1479–1491 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  71. 71

    Shalem, O., Sanjana, E.N., Hartenian, E. & Zhang, F. Genome-Scale CRISPR-Cas9 Knockout. Science 343, 84–88 (2014).

    Article  CAS  Google Scholar 

  72. 72

    Langmead, B., Trapnell, C., Pop, M. & Salzberg, S.L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  73. 73

    Kampmann, M., Bassik, M.C. & Weissman, J.S. Functional genomics platform for pooled screening and generation of mammalian genetic interaction maps. Nat. Protoc. 9, 1825–1847 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  74. 74

    Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  75. 75

    Chatr-aryamontri, A. et al. The BioGRID interaction database: 2015 update. Nucleic Acids Res. 43, D470–D478 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  76. 76

    Eisen, M.B., Spellman, P.T., Brown, P.O. & Botstein, D. Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA 95, 14863–14868 (1998).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  77. 77

    Saldanha, A.J. Java Treeview--extensible visualization of microarray data. Bioinformatics 20, 3246–3248 (2004).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references


We thank S. Collins and members of the Bassik laboratory for helpful discussions and critical reading of the manuscript, A. Sockell and L. Tonkin for technical assistance in deep sequencing, and M. Porteus and S. Mantri for CD34+ HSPCs. We thank M. Cleary, J. Duque-Afonso, and P. Jackson for helpful discussions regarding drug combination assays. We thank M. Kampmann, M. Horlbeck, L. Gilbert, and J. Weissman for helpful discussions, and L. Bruhn, C. Carstens, P. Sheffield, and B. Borgo of Agilent technologies for oligonucleotide synthesis and helpful discussions. GM12892 cells were a gift from S.B. Montgomery. The work was funded by the NIH Director's New Innovator Award Program (project no. 1DP2HD084069-01) NIH/NHGRI (training grant T32 HG000044 to D.W.M.), NIH/NCI 1U01CA199261-02, and a seed grant from Stanford ChEM-H. K.H. is supported by the Walter V. and Idun Berry award. E.E.J. was supported by the NIH (2T32CA009302) and a Hubert Shaw and Sandra Lui Stanford Graduate Fellowship.

Author information




K.H. and M.C.B. conceived and designed the study. K.H. designed the CDKO system and the scoring systems for GI map. K.H. analyzed the screen data and performed the GI and PPI analyses. K.H., A.L., and E.E.J. performed the CDKO screens. G.T.H., A.L., and D.W.M. performed the genome-wide screens for ricin modulators. D.W.M. selected the best-working sgRNAs to design the CDKO libraries. K.H. and E.E.J. validated the hits from the CDKO screens. E.E.J. performed the drug validations and related experiments. K.H., E.E.J., and M.C.B. wrote the manuscript. All authors discussed the results and the manuscript. M.C.B. supervised the study.

Corresponding author

Correspondence to Michael C Bassik.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 A scalable CDKO system with minimized recombination compatible with deep sequencing.

(a) Generation of PCR amplicons for deep sequencing (see methods). Double-sgRNA cassettes were directly amplified from genomic DNA and adapters were added during two rounds of PCRs. (b) Paired-end sequencing to directly read double-sgRNAs. PCR amplicons (around 640 bp) were clustered efficiently in flow cells and three custom sequencing primers used to read a front sgRNA, a rear sgRNA, and an index in order. (c) Double-sgRNA vectors with either two identical mU6 promoters, or mU6 and hU6 promoters driving expression of mCherry and GFP-targeting sgRNAs, were PCR-amplified using the primer pairs indicated. (d) The double-sgRNA cassettes in panel c were PCR-amplified from either purified plasmids or genomic DNA isolated from K562 cells infected with the corresponding double-sgRNA vectors. Expected size of the PCR amplicons are marked with red arrows and PCR amplicons from recombination-affected vectors are marked with blue arrows. (e) Flow cytometry analysis of GFP and mCherry knockout efficiency in cells infected with the vectors in panel. (f) Sequencing of the recombination-affected PCR amplicon from the double-sgRNA vector with two identical mU6 promoters shows that the recombination happens between two mU6 promoters.

Supplementary Figure 2 Assessment of the DrugTarget-CDKO library quality.

(a) Estimating the minimally required read count for a double-sgRNA. From the representation of single-sgRNAs in the hU6 and mU6 single-sgRNA library, the expected frequency of double-sgRNAs were calculated and compared to the observed frequency of double-sgRNAs in the DrugTarget-CDKO library. Ratios of the two frequencies showed that under 50 read counts, the observed frequencies markedly fell below the expected. Based on this data, double-sgRNAs with less than 50 read counts were removed from further analyses. (b) Frequencies of double-sgRNAs were compared between the Plasmid library and the T0 sample. Frequencies of Safe_Safe sgRNAs were slightly enriched in the T0 sample since most double-sgRNAs have negative γ-phenotypes. (c) Minimal positional bias in DrugTarget-CDKO library. γ-phenotypes of gene pairs were compared between both orientations. PCC, Pearson correlation coefficient. (d) High reproducibility of measured γ-phenotypes of gene pairs between two experimental replicates. PCC, Pearson correlation coefficient. (e) Single knockout phenotypes of genes are highly correlated between the DrugTarget-CDKO screen and a previous genome-wide single-sgRNA screen10. PCC, Pearson correlation coefficient. (f) Sequencing depth affects the phenotypic plateau observed in Figure 3a.

Supplementary Figure 3 Comparison of GI scores.

(a) GIT scores were calculated based on Raw-GIs and compared between two experimental replicates. PCC, Pearson correlation coefficient. (b) GIM scores were calculated and compared between two experimental replicates (see methods). GIM scores of gene pairs comprised of two same genes are marked in pink. The 5 most synergistic gene pairs are marked in orange. PCC, Pearson correlation coefficient. (c) Histograms of GIT scores showed that most gene pairs comprised of two same genes are buffering (red solid line). GIT score of Safe_Safe pair (purple dotted line). (d) Norm GIs of double-sgRNA combinations targeting the most synergistic (AKT1_AKT2) and most buffering (SKP2_SKP2) gene pairs by GIT score were plotted against the expected γ-phenotypes. These two pairs showed distinct GI distributions that were well-separated in opposite directions from 0. Safe_Safe double-sgRNAs are marked as yellow dots and all other double-sgRNAs are marked as grey dots. (e) Expected and observed γ-phenotypes of gene pairs were plotted and color-coded by their GIT score. The most synergistic (AKT1_AKT2) and buffering (SKP2_SKP2) pairs are highlighted.

Supplementary Figure 4 Two different sgRNAs targeting the same gene are more synergistic than identical sgRNAs.

(a-c) Norm-GIs were measured for three groups - all double-sgRNAs, pairs of same guides, and pairs of two different sgRNAs targeting the same genes. Two sgRNAs targeting the same gene tend to be buffering. However, two different sgRNAs targeting the same gene tend to be less buffering than two identical sgRNAs. (a) Norm-GIs were plotted against the expected γ-phenotypes for all three groups (black: all double-sgRNAs, pink: double-sgRNAs comprised of two same guides, blue: double-sgRNAs comprised of two different guides targeting the same gene) (b) Plots in a were binned against the expected phenotype of double-sgRNAs. Data represent mean ± s.e.m. Data were binned in a way that each range includes at least 30 data points for any given group among the three. (c) Distribution of Norm-GIs for the three groups. Only double-sgRNAs with absolute γ-phenotypes greater than 3.7 were analyzed. The pink line indicates the mean, the blue box represents the 25th-75th percentile, and the dotted bar indicates the 10th-90th percentile. (d) The Norm-GI patterns in 3 x 3 double-sgRNA combinations of pairs targeting the same gene. Top panel shows the distribution of GI scores for pairs targeting the same gene. The blue dotted lines mark the 5 most synergistic pairs in the distribution. The brown dotted lines represent the 35th-39th buffering pairs, and the yellow dotted lines mark the 5 most buffering pairs in the distribution. All double-sgRNA combinations for each group were averaged over the 3 x 3 grid and color-coded by their Norm-GIs. (e) In the dual sgRNA validation assay, two different sgRNAs targeting TK1 showed synergistic GI whereas two identical sgRNAs did not show significant GI. Data represent mean ± s.d. (n=3) from replicate cultures.

Supplementary Figure 5 γ-phenotype-based DrugTarget-CDKO GI map shows sparse genetic interactions.

(a) Correlations of GI profiles between two sgRNAs were compared in two experimental replicates: sgRNAs targeting the same gene are marked in pink. Due to the low GI frequencies, correlations of GI profiles between sgRNA pairs were very low. PCC, Pearson correlation coefficient. (b) The distributions of correlations of GI profiles for all sgRNA pairs (blue) and for sgRNAs pairs targeting the same gene (orange). Medians of the distributions are marked by dotted lines. (c) Sparse genetic interactions in the DrugTarget-CDKO GI map. GIM scores of all gene pairs were calculated and color-coded by a yellow-cyan heatmap. Genes were hierarchically clustered by their correlation of GI profiles. γ-phenotypes of individual genes are marked in sidebars with a red-blue heatmap.

Supplementary Figure 6 Ricin-CDKO screen.

(a) Selection of genes modulating ricin-sensitivity and resulting Ricin-CDKO library. (b) Schematic of the Ricin-CDKO screen. Infected cells were split into two and one group was treated with 4 pulses of ricin for 14 days, while the other was cultured for 14 days without ricin treatment. (c) Cumulative distribution of sequencing reads for double-sgRNAs. (d) Histogram plotting the number of double-sgRNAs per gene pair. 99.6% of the 6,063 detected gene pairs have more than 6 double-sgRNA combinations. (e) Minimal positional bias in Ricin-CDKO library for ρ-phenotype. PCC, Pearson correlation coefficient. (f) Measuring Norm-GIs of double-sgRNAs. Expected and observed ρ-phenotypes of double-sgRNAs were plotted and deviations from the median line (blue) were processed from Raw-GIs into Norm-GIs as previously done for DrugTarget-CDKO screen. (g) t-value-based GIT scores calculated for ρ-phenotypes were plotted between two experimental replicates. The 5 most synergistic and buffering pairs by rank-sum of GIT scores of two replicates are marked in orange dots. The Pearson correlation after same-gene targeting pairs were removed is reported in parentheses. PCC, Pearson correlation coefficient.

Supplementary Figure 7 Comparison of a previous shRNA-based ricin GI map with the Ricin-CDKO GI map.

(a) Measurement of correlation of genetic interactions between the previously published shRNA-based Ricin GI map11 and the Ricin-CDKO map for the 91 common interactions present in both maps. GIs are scaled according to the system described in the corresponding manuscript. Two cases of genetic interactions which have opposite signs in the two maps are labeled in the plot. PCC, Pearson correlation coefficient. (b) Genetic interactions of the common protein complexes and genes between the two maps are used to generate two compact GI maps for comparison. Essential genes are marked in red.

Supplementary Figure 8 Gene pairs with similar GI profiles are enriched for known protein interactions (Online Methods).

(a) Data in Figure 4d were plotted as AUC (Area Under the Curve). Gene pairs were sorted by the different features of GI map and cumulative sums of the number of STRING interactions identified in the sorted gene pairs were plotted. (b) A genetic interaction network of the 66 most correlated gene pairs (GI correlation > 0.5) in terms of GI patterns was generated. Genes are grouped and colored by their biological functions and their known PPIs. Each edge between two genes indicates that they have a correlation of GI patterns over 0.5. A red edge indicates that this gene pair also has known protein interactions reported in STRING. 40% of the 66 gene pairs have reported protein interactions whereas on average, only 5% of 66 randomly selected gene pairs from the Ricin-CDKO map have reported protein interactions in STRING.

Supplementary Figure 9 Validation of synergistic gene pairs with individual sgRNAs.

(a) Predicted synergistic gene pairs were validated using individual sgRNAs (using two separate vectors). Data represent mean ± s.d. from 3 replicate cultures. (b) 3 of the synergistic gene pairs were validated using a second pair of sgRNAs. Data represent mean ± s.d. from 3 replicate cultures. (c) Example of GI calculation for individual sgRNA validations using double-sgRNA vector. Double-sgRNA vectors were cloned containing two safe-sgRNAs, one safe-sgRNA and one gene-targeting sgRNA (PIM1_Safe and Safe_PIM2), or two gene-targeting sgRNAs (PIM1_PIM2) and infected into Cas9-expressing K562 cells. Growth phenotypes of single and double gene knockouts are calculated by measuring the depletion of GFP+ cells relative to uninfected cells (PIM1_Safe and Safe_PIM2 for single knockout phenotypes and PIM1_PIM2 for double knockout) from T0 to T7, normalized to Safe_Safe cells. GIs are determined by comparing the observed double knockout phenotype to the expected from the single knockout phenotypes. (d) Using the double-sgRNA vector system, synergy was validated for sgRNA pairs predicted to be synergistic (PIM1_PIM2, BCL2L1_MCL1, PRKDC_ATM, PRKDC_TSPO) while sgRNA pairs not predicted to be synergistic did not show synergy in dual-sgRNA retests (PRKDC_PIM2, PRKDC_MCL1, PIM1_TSPO, BCL2L1_TSPO, BCL2L1_ATM, PIM1_MCL1, BCL2L1_PIM2). Data represent mean ± s.d. from 3 replicate cultures. (e) TIDE indel analysis for sgRNAs against indicated genes.

Supplementary Figure 10 Combination APEX1_ATM drug treatment synergistically induces DSBs and apoptosis.

(a-c) K562 cells were treated with APEX1 (CRT0044876) and ATM (KU-60019) inhibitors for 48 h, fixed, and stained for γH2AX. Cells were analyzed by flow cytometry and representative histograms from each sample are plotted in a. Median FL-1 ± s.d. from 3 replicate cultures are plotted in b. Cells were additionally stained with Hoescht and representative images are shown in c. Scale bars,10 μm. (d,e) K562 cells were treated with indicated drugs for 48 h and assessed by flow cytometry for Annexin V-FITC and propidium iodide (PI) staining. Plots in d are representative of three independent experiments and the percentages of Annexin V-positive cells are quantified in e (mean ± s.d., n = 3).

Supplementary Figure 11 DrugTarget-CDKO genetic interactions predict drug synergy.

(a-k) Cell viability and Bliss drug synergy plots for drug pairs in K562 cells (a-j) and MV4;11 cells (k). Additional true positives are shown in a-c: (a) ATM (KU-60019) and PRKDC (NU7441), (b) APEX1 (CRT0044876) and PRKDC, (c) TSPO (PK-11195) and PRKDC. Examples of true negatives are shown in d-g: (d) TXN (PX-12) and XPO1 (KPT-330), (e) MCL1 (A-1210477) and PRKDC, (f) TSPO and XPO1, (g) CARM1 (1-benzyl-3,5-bis-(3-bromo-4-hydroxybenzylidene)piperidin-4-one) and XPO1. (h) NAMPT (FK866) and XPO1 (false positive). (i) BCL2L1 (A-1155463) and XPO1 (false negative). (j) The BCL2L1 and MCL1 pair was tested using a different MCL1 inhibitor (UMI-77) and (k) in the MV4;11 AML cell line (using A-1155463 and A-1210477).

Supplementary Figure 12 Normalized GI improves reproducibility of genetic interactions between replicates.

(a) GIT scores calculated from Raw-GI and Norm-GI are compared for the ρ-phenotype-based Ricin-CDKO map. GIT scores calculated from Norm-GI show higher correlation between replicates than those from Raw-GI. PCC, Pearson correlation coefficient. (b) GIT scores calculated from Raw-GI and Norm-GI are compared for the γ-phenotype-based DrugTarget-CDKO map. GIT scores calculated from Norm-GI show slightly higher correlation between replicates than those from RawGI. PCC, Pearson correlation coefficient. (c) Normalization of GIs improves the uniformity of variance across the range of expected γ-phenotypes. Variance of Raw-GIs and Norm-GIs for the same guide pairs across two experimental replicates in DrugTarget-CDKO map was measured with respect to the expected γ-phenotype. Data are binned across the expected γ-phenotypes (bin number = 22, bin size = 1 pZ) and average variance on each bin is calculated and plotted in the graphs.

Supplementary Figure 13 Complete blot images.

Complete blot images presented in Figure 6l. Protein bands shown in Figure 6l are indicated by arrows. (a) Blots for BCL2L1 (left panel) and tubulin loading control (right panel). (b) Blots for MCL1 (left panel) and tubulin loading control (right panel).

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–13 and Supplementary Text (PDF 6026 kb)

Supplementary Table 1

Selected 207 genes for DrugTarget-CDKO library (XLSX 29 kb)

Supplementary Table 2

700 sgRNAs for DrugTarget-CDKO library (XLSX 29 kb)

Supplementary Table 3

Distribution of double-sgRNAs per gene pair after filtering (XLSX 23 kb)

Supplementary Table 4

GI scores of DrugTarget-CDKO screen (XLSX 2925 kb)

Supplementary Table 5

Selected 79 genes for Ricin-CDKO library (XLSX 11 kb)

Supplementary Table 6

284 sgRNAs for Ricin-CDKO library (XLSX 17 kb)

Supplementary Table 7

GI scores of Ricin-CDKO screen (XLSX 482 kb)

Supplementary Table 8

246 STRING interactions between 79 Ricin hits (XLSX 30 kb)

Supplementary Table 9

STRING interactions in the 66 most correlated gene pairs (XLSX 13 kb)

Supplementary Table 10

Selected 79 genes for DrugTarget Batch retest (XLSX 9 kb)

Supplementary Table 11

287 sgRNAs for DrugTarget Batch retest (XLSX 17 kb)

Supplementary Table 12

sgRNAs used for the validation of individual sgRNA pairs (XLSX 12 kb)

Supplementary Table 13

Summary of sgRNA and drug validations (XLSX 11 kb)

Supplementary Table 14

30 most synergistic DrugTarget pairs (XLSX 11 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Han, K., Jeng, E., Hess, G. et al. Synergistic drug combinations for cancer identified in a CRISPR screen for pairwise genetic interactions. Nat Biotechnol 35, 463–474 (2017).

Download citation

Further reading