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:

Interpretable neural architecture search and transfer learning for understanding CRISPR–Cas9 off-target enzymatic reactions

Nature Computational Science volume 3, pages 1056–1066 (2023)Cite this article

Subjects

A preprint version of the article is available at arXiv.

Abstract

Finely tuned enzymatic pathways control cellular processes, and their dysregulation can lead to disease. Developing predictive and interpretable models for these pathways is challenging because of the complexity of the pathways and of the cellular and genomic contexts. Here we introduce Elektrum, a deep learning framework that addresses these challenges with data-driven and biophysically interpretable models for determining the kinetics of biochemical systems. First, it uses in vitro kinetic assays to rapidly hypothesize an ensemble of high-quality kinetically interpretable neural networks (KINNs) that predict reaction rates. It then employs a transfer learning step, where the KINNs are inserted as intermediary layers into deeper convolutional neural networks, fine-tuning the predictions for reaction-dependent in vivo outcomes. We apply Elektrum to predict CRISPR–Cas9 off-target editing probabilities and demonstrate that Elektrum achieves improved performance, regularizes neural network architectures and maintains physical interpretability.

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: Overview of Elektrum framework.
Fig. 2: Searching KINN architectures for in vitro Cas9 cleavage kinetics.
Fig. 3: Integration of transfer learning in NAS for in vivo cleavage prediction.
Fig. 4: Interpreting sequence context for in vivo Cas9 cleavage rate prediction.

Similar content being viewed by others

Data availability

All training and testing data presented in this paper are publicly available. The in vitro kinetic data were downloaded from ref. 20 (https://github.com/finkelsteinlab/nucleaseq/tree/master). The in vivo CIRSPR–Cas9 off-targeting dataset was downloaded from CRISPR-Net29 (https://codeocean.com/capsule/9553651/tree/v2). Source data is available with this paper.

Code availability

The Elektrum code is available on GitHub at https://github.com/zj-zhang/Elektrum and Zenodo at https://doi.org/10.5281/zenodo.8044859 ref. 51.

References

  1. Gebauer, F., Schwarzl, T., Valcárcel, J. & Hentze, M. W. RNA-binding proteins in human genetic disease. Nat. Rev. Genet. 22, 185–198 (2021).

    Article  Google Scholar 

  2. Masoud, G. N. & Li, W. Hif-1α pathway: role, regulation and intervention for cancer therapy. Acta Pharm. Sin. B 5, 378–389 (2015).

    Article  Google Scholar 

  3. Santamaria, S. & Groot, R. ADAMTS proteases in cardiovascular physiology and disease. Open Biol. 10, 200333 (2020).

    Article  Google Scholar 

  4. Flinn, A.M. & Gennery, A.R. Adenosine deaminase deficiency: a review. Orphanet J. Rare Dis. 13, 65 (2018).

    Article  Google Scholar 

  5. Kim, R. Q. et al. Kinetic analysis of multistep USP7 mechanism shows critical role for target protein in activity. Nat. Commun. 10, 231 (2019).

    Article  Google Scholar 

  6. Persikov, A. V. et al. A systematic survey of the Cys2His2 zinc finger DNA-binding landscape. Nucleic Acids Res. 43, 1965–1984 (2015).

    Article  Google Scholar 

  7. Liepelt, S. & Lipowsky, R. Kinesin’s network of chemomechanical motor cycles. Phys. Rev. Lett. 98, 258102 (2007).

    Article  Google Scholar 

  8. Schreiber, G. Kinetic studies of protein–protein interactions. Curr. Opin. Struct. Biol. 12, 41–47 (2002).

    Article  Google Scholar 

  9. Zhou, J. & Troyanskaya, O. G. Predicting effects of noncoding variants with deep learning-based sequence model. Nat. Methods 12, 931–934 (2015).

    Article  Google Scholar 

  10. Fudenberg, G., Kelley, D. R. & Pollard, K. S. Predicting 3D genome folding from DNA sequence with Akita. Nat. Methods 17, 1111–1117 (2020).

    Article  Google Scholar 

  11. Li, V. R., Zhang, Z. & Troyanskaya, O. G. CROTON: an automated and variant-aware deep learning framework for predicting CRISPR/Cas9 editing outcomes. Bioinformatics 37, 342 (2021).

    Article  Google Scholar 

  12. Avsec, Ž. et al. Effective gene expression prediction from sequence by integrating long-range interactions. Nat. Methods 18, 1196–1203 (2021).

    Article  Google Scholar 

  13. Wong, A. K., Sealfon, R. S., Theesfeld, C. L. & Troyanskaya, O. G. Decoding disease: from genomes to networks to phenotypes. Nat. Rev. Genet. 22, 774–790 (2021).

    Article  Google Scholar 

  14. Ching, T. et al. Opportunities and obstacles for deep learning in biology and medicine. J. R. Soc. Interface 15, 20170387 (2018).

    Article  Google Scholar 

  15. Tareen, A. & Kinney, J. B. Biophysical models of cis-regulation as interpretable neural networks. Preprint at bioRxiv https://doi.org/10.1101/835942 (2019).

  16. Tareen, A. et al. MAVE-NN: learning genotype–phenotype maps from multiplex assays of variant effect. Genome Biol. 23, 1–27 (2022).

    Article  Google Scholar 

  17. Fowler, D. M. & Fields, S. Deep mutational scanning: a new style of protein science. Nat. Methods 11, 801–807 (2014).

    Article  Google Scholar 

  18. Faure, A. J. et al. Mapping the energetic and allosteric landscapes of protein binding domains. Nature 604, 175–183 (2022).

    Article  Google Scholar 

  19. Kretz, C. A. et al. Massively parallel enzyme kinetics reveals the substrate recognition landscape of the metalloprotease ADAMTS13. Proc. Natl Acad. Sci. USA 112, 9328–33 (2015).

    Article  Google Scholar 

  20. Jones, S. K. et al. Massively parallel kinetic profiling of natural and engineered CRISPR nucleases. Nat. Biotechnol. 39, 84–93 (2021).

    Article  Google Scholar 

  21. Zhang, Z., Park, C. Y., Theesfeld, C. L. & Troyanskaya, O. G. An automated framework for efficiently designing deep convolutional neural networks in genomics. Nat. Mach. Intell. 3, 392–400 (2021).

    Article  Google Scholar 

  22. Tsai, S. Q. et al. Circle-seq: a highly sensitive in vitro screen for genome-wide CRISPR–Cas9 nuclease off-targets. Nat. Methods 14, 607–614 (2017).

    Article  Google Scholar 

  23. Haapaniemi, E., Botla, S., Persson, J., Schmierer, B. & Taipale, J. CRISPR–Cas9 genome editing induces a p53-mediated DNA damage response. Nat. Med. 24, 927–930 (2018).

    Article  Google Scholar 

  24. Cancellieri, S. et al. Human genetic diversity alters off-target outcomes of therapeutic gene editing. Nat. Genet. 55, 34–43 (2023).

    Article  Google Scholar 

  25. Eslami-Mossallam, B. et al. A kinetic model predicts SpCas9 activity, improves off-target classification, and reveals the physical basis of targeting fidelity. Nat. Commun. 13, 1–10 (2022).

    Article  Google Scholar 

  26. Klein, M., Eslami-Mossallam, B., Arroyo, D. G. & Depken, M. Hybridization kinetics explains CRISPR-Cas off-targeting rules. Cell Rep. 22, 1413–1423 (2018).

    Article  Google Scholar 

  27. Fu, R. et al. Systematic decomposition of sequence determinants governing CRISPR/Cas9 specificity. Nat. Commun. https://doi.org/10.1038/s41467-022-28028-x (2022).

  28. Concordet, J.-P. & Haeussler, M. CRISPOR: intuitive guide selection for CRISPR/Cas9 genome editing experiments and screens. Nucleic Acids Res. 46, 242–245 (2018).

    Article  Google Scholar 

  29. Lin, J., Zhang, Z., Zhang, S., Chen, J. & Wong, K.-C. CRISPR-Net: a recurrent convolutional network quantifies CRISPR off-target activities with mismatches and indels. Adv. Sci. 7, 1903562 (2020).

    Article  Google Scholar 

  30. Listgarten, J. et al. Prediction of off-target activities for the end-to-end design of CRISPR guide RNAs. Nat. Biomed. Eng. 2, 38–47 (2018).

    Article  Google Scholar 

  31. Doench, J. G. et al. Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR–cas9. Nat. Biotechnol. 34, 184–191 (2016).

    Article  Google Scholar 

  32. Haeussler, M. et al. Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA selection tool CRISPOR. Genome Biol. 17, 1–12 (2016).

    Article  Google Scholar 

  33. Cameron, P. et al. Mapping the genomic landscape of CRISPR-Cas9 cleavage. Nat. Methods 14, 600–606 (2017).

    Article  Google Scholar 

  34. Kleinstiver, B. P. et al. Engineered CRISPR–Cas9 nucleases with altered PAM specificities. Nature 523, 481–485 (2015).

    Article  Google Scholar 

  35. Zhuo, C. et al. Spatiotemporal control of CRISPR/Cas9 gene editing. Signal Transduct. Target. Ther. 6, 1–18 (2021).

    Google Scholar 

  36. Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J. & Wojna, Z. Rethinking the inception architecture for computer vision. In Proc. IEEE Conference on Computer Vision and Pattern Recognition 2818–2826 (IEEE, 2016).

  37. Kelley, D. R. et al. Sequential regulatory activity prediction across chromosomes with convolutional neural networks. Genome Res. 28, 739–750 (2018).

    Article  Google Scholar 

  38. Lundberg, S. M. & Lee, S.-I. A unified approach to interpreting model predictions. Adv. Neural Inf. Process. Syst. 30, 4765–4774 (2017).

  39. Moreno-Mateos, M. A. et al. CRISPRscan: designing highly efficient sgRNAs for CRISPR–Cas9 targeting in vivo. Nat. Methods 12, 982–988 (2015).

    Article  Google Scholar 

  40. Moreb, E. & Lynch, M. Genome dependent Cas9/gRNA search time underlies sequence dependent gRNA activity. Nat. Commun. 12, 5034 (2021).

    Article  Google Scholar 

  41. Moreb, E. A., Hutmacher, M. & Lynch, M. D. CRISPR–Cas “non-target" sites inhibit on-target cutting rates. CRISPR J. 3, 550–561 (2020).

    Article  Google Scholar 

  42. Shen, Y., Pressman, A., Janzen, E. & Chen, I. A. Kinetic sequencing (k-seq) as a massively parallel assay for ribozyme kinetics: utility and critical parameters. Nucleic Acids Res. 49, 67 (2021).

    Article  Google Scholar 

  43. King, E. L. & Altman, C. A schematic method of deriving the rate laws for enzyme-catalyzed reactions. J. Phys. Chem. 60, 1375–1378 (1956).

    Article  Google Scholar 

  44. Cornish Bowden, A. An automatic method for deriving steady-state rate equations. Biochem. J. 165, 55–59 (1977).

    Article  Google Scholar 

  45. Lam, C. F. & Priest, D. G. Enzyme kinetics: systematic generation of valid King–Altman patterns. Biophys. J. 12, 248–256 (1972).

    Article  Google Scholar 

  46. Pelikan, M. Probabilistic model-building genetic algorithms. In Proc. 13th Annual Conference Companion on Genetic and Evolutionary Computation 913–940 (2011).

  47. Wang, Wei, et al. "Backpropagation-friendly eigendecomposition." Advances in Neural Information Processing Systems 32, (2019).

  48. Bae, S., Park, J. & Kim, J.-S. Cas-OFFinder: a fast and versatile algorithm that searches for potential off-target sites of Cas9 RNA-guided endonucleases. Bioinformatics 30, 1473–1475 (2014).

    Article  Google Scholar 

  49. Lewandowski, D., Kurowicka, D. & Joe, H. Generating random correlation matrices based on vines and extended onion method. J. Multivar. Anal. 100, 1989–2001 (2009).

    Article  MathSciNet  Google Scholar 

  50. Salvatier, J., Wiecki, T. V. & Fonnesbeck, C. Probabilistic programming in Python using PyMC3. PeerJ Comput. Sci. 2, 55 (2016).

    Article  Google Scholar 

  51. Zhang, F. Z. & Lamson, A. R. zj-zhang/Elektrum: frozen publication version. Zenodo https://doi.org/10.5281/zenodo.8044859 (2023).

  52. Liu, Q., He, D. & Xie, L. Prediction of off-target specificity and cell-specific fitness of CRISPR–Cas system using attention boosted deep learning and network-based gene feature. PLoS Comput. Biol. 15, 1007480 (2019).

    Article  Google Scholar 

  53. Peng, Hui., Zheng, Yi., Zhao, Zhixun., Liu, Tao. & Li, Jinyan. Recognition of CRISPR/Cas9 off-target sites through ensemble learning of uneven mismatch distributions. Bioinformatics 34, 757–765 (2018).

    Article  Google Scholar 

  54. Lin, J. & Wong, K.-C. Off-target predictions in CRISPR–Cas9 gene editing using deep learning. Bioinformatics 34, 656–663 (2018).

    Article  Google Scholar 

  55. Alkan, F., Wenzel, A., Anthon, C., Havgaard, J. H. & Gorodkin, J. CRISPR–Cas9 off-targeting assessment with nucleic acid duplex energy parameters. Genome Biol. 19, 1–13 (2018).

    Article  Google Scholar 

Download references

Acknowledgements

We thank C. Theesfeld and E. Thiede for useful conversations. O.T. is supported by National Institutes of Health (NIH) grant R01GM071966, and Simons Foundation grant no. 395506.

Author information

Author notes

  1. These authors contributed equally: Zijun Zhang, Adam R. Lamson.

Authors and Affiliations

  1. Division of Artificial Intelligence in Medicine, Cedars-Sinai Medical Center, Los Angeles, CA, USA

    Zijun Zhang

  2. Center for Computational Biology, Flatiron Institute, New York City, NY, USA

    Adam R. Lamson, Michael Shelley & Olga Troyanskaya

  3. Courant Institute of Mathematical Sciences, New York University, New York City, NY, USA

    Michael Shelley

  4. Lewis Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ, USA

    Olga Troyanskaya

Authors
  1. Zijun Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Adam R. Lamson
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Michael Shelley
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Olga Troyanskaya
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Z.Z. and A.R.L. conceived the study, wrote the code and performed the analysis. M.S. and O.T. supervised the study. Z.Z., A.R.L., M.S. and O.T. wrote the paper.

Corresponding authors

Correspondence to Michael Shelley or Olga Troyanskaya.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Computational Science thanks Xin Gao, Christina Theodoris and Ka-Chun Wong for their contribution to the peer review of this work. Primary Handling Editor: Kaitlin McCardle, in collaboration with the Nature Computational Science team.

Additional information

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

Supplementary information

Supplementary Information

Supplementary Figs. 1–10 and Tables 1-3.

Reporting Summary

Source data

Source Data Fig. 2

Statistical source data for Fig. 2.

Source Data Fig. 3

Statistical source data for Fig. 3.

Source Data Fig. 4

Statistical source data for Fig. 4.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhang, Z., Lamson, A.R., Shelley, M. et al. Interpretable neural architecture search and transfer learning for understanding CRISPR–Cas9 off-target enzymatic reactions. Nat Comput Sci 3, 1056–1066 (2023). https://doi.org/10.1038/s43588-023-00569-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s43588-023-00569-1

This article is cited by

Search

Advanced search

Quick links

Nature Briefing AI and Robotics

Sign up for the Nature Briefing: AI and Robotics newsletter — what matters in AI and robotics research, free to your inbox weekly.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing: AI and Robotics