CoChat
Sign Up
PreprintCited by 15Protein Structure and DynamicsCRISPR and Genetic EngineeringMachine Learning in Bioinformatics

The DOI matches its Crossref record; 1 of the 1 references it cites were found in a registry or an index.

Computational design of novel Cas9 PAM-interacting domains using evolution-based modelling and structural quality assessment

Open access
Malbranke, Cyril, Rostain, William, Depardieu, Florence, Cocco, Simona, Monasson, Rémi, Bikard, David·2023·PLoS Computational Biology
DOI 10.1101/2023.03.20.533501PMC PMC10729993

Preprint — published as 10.1371/journal.pcbi.1011621 · Paper page

Open this paper in CoChat — free accountOpen PDF

Abstract

We present here an approach to protein design that enables to leverage (i) scarce functional information such as experimental data (ii) evolutionary information learned from a natural sequence variants and (iii) physics-grounded modeling. Using a Restricted Boltzmann Machine (RBM), we learned a sequence model of a protein family. We use semi-supervision to leverage available functional information during the RBM training. We then propose a strategy to explore the protein representation space that can be informed by external models such as an empirical force field method (FoldX). This method was applied to a domain of the Cas9 protein responsible for recognition of a short DNA motif. We experimentally assessed the functionality of 71 variants that were generated to explore a range of RBM and FoldX energies. We show how a combination of functional, structural and evolutionary information can identify functional variants with high accuracy. Sequences with as many as 50 differences (20% of the protein domain) to the wild-type retained functionality. Overall, 21/71 sequences designed with our method were functional. Interestingly, 6/71 sequences showed an improved activity in comparison with the original wild-type protein sequence. These results demonstrate the interest in further exploring the synergies between machine-learning of protein sequence representations and physics grounded modeling strategies informed by structural information.

Overview

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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

Integrity signals

Registry

Crossref · 10.1101/2023.03.20.533501

Cites

1 of 1 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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

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

Flagged first, then by citations

Research works printed but not linked: 63

  1. 10.1093/nar/gky1004 · DOI
  2. 10.1093/nar/gkx272 · DOI
  3. 10.1073/pnas.1111471108 · DOI · Paper page
  4. 10.1038/nbt.3769 · DOI · Paper page
  5. 10.1126/science.aba3304 · DOI · Paper page
  6. 10.1371/journal.pcbi.1008736 · DOI
  7. 10.1016/j.jcp.2014.07.024 · DOI
  8. 10.1073/pnas.0805923106 · DOI · Paper page
  9. 10.1038/s41467-021-25756-4 · DOI
  10. 2019, Learning protein constitutive motifs from sequence data, Elife, 8, e39397., 10.7554/eLife.39397 · DOI
  11. 10.1038/s41467-021-26529-9 · DOI
  12. Rives A , Meier J , Sercu T , Goyal S , Lin Z , Liu J , et al. Biological Structure and Function Emerge from Scaling Unsupervised Learning to 250 Million Protein Sequences. Proceedings of the National Academy of Sciences. 2021;118(15). (Back to p.2, 14) · DOI · Paper page
  13. 10.1109/TPAMI.2021.3095381 · DOI · Paper page
  14. 10.1038/s41586-021-04184-w · DOI · Paper page
  15. Baldassarre F , Hurtado DM , Elofsson A , Azizpour H. GraphQA: Protein Model Quality Assessment Using Graph Convolutional Network. Submitted to Bioinformatics. 2020;. (Back to p.2) · DOI
  16. 10.1093/bioinformatics/btab442 · DOI
  17. 10.1038/s41592-020-0848-2 · DOI · Paper page
  18. 10.1093/bioinformatics/btz184 · DOI
  19. 10.1126/science.abj8754 · DOI · Paper page
  20. Moffat L , Greener JG , Jones DT . Using AlphaFold for Rapid and Accurate Fixed Backbone Protein Design; 2021. (Back to p.2) · DOI · Paper page
  21. Jendrusch M , Korbel JO , Sadiq SK . AlphaDesign: A de Novo Protein Design Framework Based on AlphaFold; 2021. (Back to p.2) · DOI · Paper page
  22. 2012, Learning Algorithms for the Classification Restricted Boltzmann Machine, The Journal of Machine Learning Research, 13, 643
  23. 2021, RBM-MHC: a semi-supervised machine-learning method for sample-specific prediction of antigen presentation by HLA-I alleles, Cell systems, 12, 195, 10.1016/j.cels.2020.11.005 · DOI
  24. 2021, Computational Saturation Screen Reveals the Landscape of Mutations in Human Fumarate Hydratase, Journal of Chemical Information and Modeling, 61, 1970, 10.1021/acs.jcim.1c00063 · DOI
  25. 10.1002/jmv.26626 · DOI
  26. 10.1016/j.bbrc.2020.05.028 · DOI
  27. 2020, Enhancing the Thermostability of Rhizopus Chinensis Lipase by Rational Design and MD Simulations, International Journal of Biological Macromolecules, 160, 1189, 10.1016/j.ijbiomac.2020.05.243 · DOI
  28. 10.1126/science.1258096 · DOI · Paper page
  29. 10.1038/s41467-020-20633-y · DOI · Paper page
  30. 10.1038/s41467-020-16117-8 · DOI
  31. 10.1126/science.aba8853 · DOI · Paper page
  32. 10.1126/science.aas9129 · DOI · Paper page
  33. 10.1016/j.cell.2014.02.001 · DOI · Paper page
  34. 10.1038/s41467-019-08395-8 · DOI
  35. 10.1016/j.molcel.2018.12.003 · DOI · Paper page
  36. 10.7554/eLife.77825 · DOI
  37. 2008, Learning and Evaluating Boltzmann Machines. Utml Tr, 2, 21
  38. Tieleman T. Training Restricted Boltzmann Machines Using Approximations to the Likelihood Gradient. In: Proceedings of the 25th International Conference on Machine Learning. ICML ‘08. Helsinki, Finland: Association for Computing Machinery; 2008. p. 1064–1071. (Back to p.4, 17) · DOI
  39. 2021, Low-N protein engineering with data-efficient deep learning, Nature methods, 18, 389, 10.1038/s41592-021-01100-y · DOI · Paper page
  40. 10.1186/s13059-021-02495-9 · DOI
  41. 2021, Barriers and Dynamical Paths in Alternating Gibbs Sampling of Restricted Boltzmann Machines, Physical Review E, 104, 034109, 10.1103/PhysRevE.104.034109 · DOI
  42. Kumar S , Paria B , Tsvetkov Y. Constrained Sampling from Language Models via Langevin Dynamics in Embedding Spaces; 2022. (Back to p.7) · DOI
  43. Mirdita M , Schütze K , Moriwaki Y , Heo L , Ovchinnikov S , Steinegger M. ColabFold: Making Protein Folding Accessible to All. Nature Methods. 2022; p. 1–4. (Back to p.12) · DOI · Paper page
  44. Klausen MS , Jespersen MC , Nielsen H , Kjærgaard Jensen K, Jurtz VI , Sønderby CK , et al. NetSurfP-2.0: Improved Prediction of Protein Structural Features by Integrated Deep Learning. Bioinformatics; 2018. (Back to p.14) · DOI
  45. Asgari E , Poerner N , McHardy AC , Mofrad MRK . DeepPrime2Sec: Deep Learning for Protein Secondary Structure Prediction from the Primary Sequences; 2019. (Back to p.14) · DOI
  46. 10.1093/nar/gky439 · DOI
  47. Hekkelman ML , de Vries I , Joosten RP , Perrakis A. AlphaFill: Enriching the AlphaFold Models with Ligands and Co-Factors; 2021. (Back to p.14) · DOI · Paper page
  48. Lin Z , Akin H , Rao R , Hie B , Zhu Z , Lu W , et al. Language models of protein sequences at the scale of evolution enable accurate structure prediction. BioRxiv. 2022;. (Back to p.14)
  49. 10.1002/wcms.1121 · DOI
  50. Nijkamp E , Ruffolo J , Weinstein EN , Naik N , Madani A. Progen2: exploring the boundaries of protein language models. arXiv preprint arXiv:220613517. 2022;. (Back to p.14) · DOI · Paper page
  51. Rao RM , Liu J , Verkuil R , Meier J , Canny J , Abbeel P , et al. MSA transformer. In: International Conference on Machine Learning. PMLR; 2021. p. 8844–8856. (Back to p.14) · DOI · Paper page
  52. Elnaggar A , Essam H , Salah-Eldin W , Moustafa W , Elkerdawy M , Rochereau C , et al. Ankh: Optimized Protein Language Model Unlocks General-Purpose Modelling. bioRxiv. 2023; p. 2023–01. (Back to p.14) · DOI
  53. Gligorijević V , Berenberg D , Ra S , Watkins A , Kelow S , Cho K , et al. Function-guided protein design by deep manifold sampling. bioRxiv. 2021; p. 2021–12. (Back to p.14) · DOI
  54. 2021, Deep extrapolation for attribute-enhanced generation, Advances in Neural Information Processing Systems, 34, 14084
  55. Madani A , Krause B , Greene ER , Subramanian S , Mohr BP , Holton JM , et al. Large language models generate functional protein sequences across diverse families. Nature Biotechnology. 2023; p. 1–8. (Back to p.14) · DOI · Paper page
  56. 2023, Machine learning for evolutionary-based and physics-inspired protein design: Current and future synergies, Current Opinion in Structural Biology, 80, 102571, 10.1016/j.sbi.2023.102571 · DOI
  57. 10.1126/science.add2187 · DOI · Paper page
  58. 2023, PeSTo: parameter-free geometric deep learning for accurate prediction of protein binding interfaces, Nature Communications, 14, 2175, 10.1038/s41467-023-37701-8 · DOI
  59. 10.1093/bioinformatics/btw006 · DOI
  60. Glorot X , Bengio Y. Understanding the Difficulty of Training Deep Feedforward Neural Networks.In: Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics. JMLR Workshop and Conference Proceedings; 2010. p. 249–256. (Back to p.16)
  61. Loshchilov I , Hutter F. Fixing Weight Decay Regularization in Adam. open review. 2018;. (Back to p.17)
  62. 2021, The impact of genetic diversity on gene essentiality within the Escherichia coli species, Nature microbiology, 6, 301, 10.1038/s41564-020-00839-y · DOI
  63. 10.1093/MOLBEV/MSAB120 · DOI

Metadata

Authors
Malbranke, C., Rostain, W., Depardieu, F., Cocco, S., Monasson, R., & Bikard, D.
Affiliations
Centre National de la Recherche Scientifique (FR)
Corresponding author
Malbranke, Cyril
ORCID
Venue
PLoS Computational Biology
Publisher
openRxiv
Published
2023-03-22
Volume
19
Issue
11
Pages
e1011621-e1011621
ISSN
1553-734X (ISSN-L)
DOI
10.1101/2023.03.20.533501
PMC
PMC10729993
OpenAlex
W4388775348
Cited by
15
References
1
Citations by year
2026: 3 · 2025: 9 · 2024: 3
FWCI
1.34
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • PLoS Computational Biology (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • HAL (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Greater South Information System
Funders
  • Agence Nationale de la Recherche (ANR-17-CE30-0021, 10-LABX-62-IBEID, ANR‐10‐LABX‐62‐IBEID, ANR‐19, LABX-62-IBEID, 101044479, ANR-19-CE30-0021, ANR-17, ANR-10, ANR-10-LABX, ANR-10-LABX-62)
  • Fondation Bettencourt Schueller
  • European Research Council (677823, 101044479)
  • École Polytechnique, Université Paris-Saclay
  • H2020 European Research Council (677823)
  • HORIZON EUROPE European Research Council (101044479)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › CRISPR and Genetic Engineering
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
Keywords
Cas9, protein design, restricted Boltzmann machine, semi-supervised learning, FoldX, evolutionary information, protein structure prediction, CRISPR/Cas9
MeSH terms
Machine Learning, Amino Acid Sequence, Learning, Proteins, CRISPR-Cas Systems

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