Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat
Sign Up
Cited by 292

The DOI matches its Crossref record; 8 of the 8 references it cites were found in a registry or an index.

De novo molecular design and generative models

Open access
Meyers, Joshua, Fabian, Benedek, Brown, Nathan·2021·Drug Discovery Today
DOI 10.1016/j.drudis.2021.05.019
Open this paper in CoChat — free account

Abstract

Molecular design strategies are integral to therapeutic progress in drug discovery. Computational approaches for de novo molecular design have been developed over the past three decades and, recently, thanks in part to advances in machine learning (ML) and artificial intelligence (AI), the drug discovery field has gained practical experience. Here, we review these learnings and present de novo approaches according to the coarseness of their molecular representation: that is, whether molecular design is modeled on an atom-based, fragment-based, or reaction-based paradigm. Furthermore, we emphasize the value of strong benchmarks, describe the main challenges to using these methods in practice, and provide a viewpoint on further opportunities for exploration and challenges to be tackled in the upcoming years.

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 · 10.1016/j.drudis.2021.05.019

Cites

8 of 8 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

Show all 8 references

Research works printed but not linked: 83

  1. Emmert-Streib F, Yang Z, Feng H, Tripathi S, Dehmer M. An introductory review of deep learning for prediction models with big data. Front Artif Intell Published online February 28, 2020. https://doi.org/10.3389/frai.2020.00004. · DOI · Paper page
  2. Bickerton, 2012, Quantifying the chemical beauty of drugs, Nat Chem, 4, 90, 10.1038/nchem.1243 · DOI · Paper page
  3. Kusner MJ, Paige B, Hernández-Lobato JM. Grammar variational autoencoder. arXiv [statML]. Published online March 6, 2017. http://arxiv.org/abs/1703.01925. · arXiv
  4. Firth, 2015, MOARF, an integrated workflow for multiobjective optimization: implementation, synthesis, and biological evaluation, J Chem Inf Model, 55, 1169, 10.1021/acs.jcim.5b00073 · DOI
  5. Polykovskiy D, Zhebrak A, Sanchez-Lengeling B, Golovanov S, Tatanov O, Belyaev S, et al. Molecular Sets (MOSES): a benchmarking platform for molecular generation models. arXiv [csLG]. Published online November 29, 2018. http://arxiv.org/abs/1811.12823. · arXiv
  6. Sanchez-Lengeling, 2018, Inverse molecular design using machine learning: Generative models for matter engineering, Science, 361, 360, 10.1126/science.aat2663 · DOI · Paper page
  7. Xu, 2019, Deep learning for molecular generation, Future Med Chem, 11, 567, 10.4155/fmc-2018-0358 · DOI
  8. Fabian B, Edlich T, Gaspar H, Segler M, Meyers J, Fiscato M, et al. Molecular representation learning with language models and domain-relevant auxiliary tasks. arXiv [csLG]. Published online November 26, 2020. http://arxiv.org/abs/2011.13230. · arXiv
  9. Schneider, 2015, Get your atoms in order–an open-source implementation of a novel and robust molecular canonicalization algorithm, J Chem Inf Model, 55, 2111, 10.1021/acs.jcim.5b00543 · DOI
  10. Bjerrum, 2018, Improving chemical autoencoder latent space and molecular de novo generation diversity with heteroencoders, Biomolecules, 8, 131, 10.3390/biom8040131 · DOI
  11. Prykhodko, 2019, A de novo molecular generation method using latent vector based generative adversarial network, J Cheminform, 11, 74, 10.1186/s13321-019-0397-9 · DOI · Paper page
  12. Arús-Pous, 2019, Randomized SMILES strings improve the quality of molecular generative models, J Cheminform, 11, 71, 10.1186/s13321-019-0393-0 · DOI · Paper page
  13. O’Boyle N, Dalke A. DeepSMILES: an adaptation of SMILES for use in machine-learning of chemical structures. ChemRxiv, Published online September 18, 2018. https://doi.org/10.26434/chemrxiv.7097960.v1. · DOI
  14. Krenn, 2020, SELFIES: a robust representation of semantically constrained graphs with an example application in chemistry, Mach Learn Sci Technol, 1
  15. Brown, 2004, A graph-based genetic algorithm and its application to the multiobjective evolution of median molecules, J Chem Inf Comput Sci, 44, 1079, 10.1021/ci034290p · DOI
  16. Reutlinger, 2014, Multi-objective molecular de novo design by adaptive fragment prioritization, Angew Chem Int Ed Engl, 53, 4244, 10.1002/anie.201310864 · DOI
  17. Jensen, 2019, A graph-based genetic algorithm and generative model/Monte Carlo tree search for the exploration of chemical space, Chem Sci, 10, 3567, 10.1039/C8SC05372C · DOI · Paper page
  18. Yoshikawa, 2018, Population-based de novo molecule generation, using grammatical evolution, Chem Lett, 47, 1431, 10.1246/cl.180665 · DOI
  19. Winter, 2019, Efficient multi-objective molecular optimization in a continuous latent space, Chem Sci, 10, 8016, 10.1039/C9SC01928F · DOI · Paper page
  20. Kwon, 2021, MolFinder: an evolutionary algorithm for the global optimization of molecular properties and the extensive exploration of chemical space using SMILES, J Cheminform, 13, 24, 10.1186/s13321-021-00501-7 · DOI
  21. Verhellen, 2020, Illuminating elite patches of chemical space, Chem Sci, 11, 11485, 10.1039/D0SC03544K · DOI
  22. Lewell, 1998, RECAP–retrosynthetic combinatorial analysis procedure: a powerful new technique for identifying privileged molecular fragments with useful applications in combinatorial chemistry, J Chem Inf Comput Sci, 38, 511, 10.1021/ci970429i · DOI · Paper page
  23. Polishchuk, 2020, CReM: chemically reasonable mutations framework for structure generation, J Cheminform, 12, 28, 10.1186/s13321-020-00431-w · DOI
  24. Vinkers, 2003, SYNOPSIS: SYNthesize and OPtimize system in silico, J Med Chem, 46, 2765, 10.1021/jm030809x · DOI
  25. Spiegel, 2020, AutoGrow4: an open-source genetic algorithm for de novo drug design and lead optimization, J Cheminform, 12, 25, 10.1186/s13321-020-00429-4 · DOI
  26. Hartenfeller, 2011, A collection of robust organic synthesis reactions for in silico molecule design, J Chem Inf Model, 51, 3093, 10.1021/ci200379p · DOI
  27. Ghiandoni, 2020, Enhancing reaction-based de novo design using a multi-label reaction class recommender, J Comput Aided Mol Des, 34, 783, 10.1007/s10822-020-00300-6 · DOI
  28. Segler, 2018, Generating focused molecule libraries for drug discovery with recurrent neural networks, ACS Cent Sci, 4, 120, 10.1021/acscentsci.7b00512 · DOI · Paper page
  29. Simonovsky M, Komodakis N. GraphVAE: towards generation of small graphs using variational autoencoders. In: Artificial Neural Networks and Machine Learning – ICANN 2018. London: Springer International Publishing; 2018: 412–422. · DOI · Paper page
  30. De Cao N, Kipf T. MolGAN: An implicit generative model for small molecular graphs. arXiv [statML]. Published online May 30, 2018. http://arxiv.org/abs/1805.11973. · arXiv
  31. Liu, 2018, Constrained graph variational autoencoders for molecule design, Adv Neural Inf Process Syst, 31, 7795
  32. Li Y, Vinyals O, Dyer C, Pascanu R, Battaglia P. Learning deep generative models of graphs. arXiv [csLG]. Published online March 8, 2018. http://arxiv.org/abs/1803.03324. · arXiv
  33. Zhou, 2019, Optimization of molecules via deep reinforcement learning, Sci Rep, 9, 10752, 10.1038/s41598-019-47148-x · DOI · Paper page
  34. Mercado, 2020, Graph networks for molecular design, Mach Learn Sci Technol, 2
  35. Lynch, 1996, The Sheffield generic structures project: a retrospective review, J Chem Inf Comput Sci, 36, 930, 10.1021/ci950173l · doi:10.1021/ci950173l
  36. Pogány, 2019, De novo molecule design by translating from reduced graphs to SMILES, J Chem Inf Model, 59, 1136, 10.1021/acs.jcim.8b00626 · DOI
  37. Jin W, Barzilay R, Jaakkola T. Junction tree variational autoencoder for molecular graph generation. arXiv [csLG]. Published online February 12, 2018. http://arxiv.org/abs/1802.04364. · arXiv
  38. Ståhl, 2019, Deep reinforcement learning for multiparameter optimization in de novo drug design, J Chem Inf Model, 59, 3166, 10.1021/acs.jcim.9b00325 · DOI
  39. Button, 2019, Automated de novo molecular design by hybrid machine intelligence and rule-driven chemical synthesis, Nat Mach Intell, 1, 307, 10.1038/s42256-019-0067-7 · DOI
  40. Bradshaw, 2019, A model to search for synthesizable molecules, Adv Neural Inf Process Syst, 32, 7937
  41. Korovina, 2020, ChemBO: Bayesian optimization of small organic molecules with synthesizable recommendations, Proc Machine Learn Res, 108, 3393
  42. Horwood J, Noutahi E. Molecular design in synthetically accessible chemical space via deep reinforcement learning. arXiv [physics.chem–ph]. Published online April 29, 2020. http://arxiv.org/abs/2004.14308. · arXiv
  43. Gottipati SK, Sattarov B, Niu S, Pathak Y, Wei H, Liu S, et al. Learning to navigate the synthetically accessible chemical space using reinforcement learning. arXiv [csLG]. Published online April 26, 2020. http://arxiv.org/abs/2004.12485. · arXiv
  44. Bradshaw J, Paige B, Kusner MJ, Segler MHS, Hernández-Lobato JM. Barking up the right tree: an approach to search over molecule synthesis DAGs. arXiv [csLG]. Published online December 21, 2020. http://arxiv.org/abs/2012.11522. · arXiv
  45. Boitreaud, 2020, OptiMol: Optimization of binding affinities in chemical space for drug discovery, J Chem Inf Model, 60, 5658, 10.1021/acs.jcim.0c00833 · DOI
  46. von Lilienfeld, 2020, Exploring chemical compound space with quantum-based machine learning, Nat Rev Chem, 4, 347, 10.1038/s41570-020-0189-9 · DOI · Paper page
  47. Ghanakota, 2020, Combining cloud-based free-energy calculations, synthetically aware enumerations, and goal-directed generative machine learning for rapid large-scale chemical exploration and optimization, J Chem Inf Model, 60, 4311, 10.1021/acs.jcim.0c00120 · DOI
  48. Gillet, 1994, SPROUT: recent developments in the de novo design of molecules, J Chem Inf Comput Sci, 34, 207, 10.1021/ci00017a027 · DOI
  49. Imrie, 2020, Deep generative models for 3D linker design, J Chem Inf Model, 60, 1983, 10.1021/acs.jcim.9b01120 · DOI · Paper page
  50. Skalic, 2019, Shape-based generative modeling for de novo drug design, J Chem Inf Model, 59, 1205, 10.1021/acs.jcim.8b00706 · DOI · Paper page
  51. Simm GNC, Pinsler R, Hernández-Lobato JM. Reinforcement learning for molecular design guided by quantum mechanics. arXiv [statML]. Published online February 18, 2020. http://arxiv.org/abs/2002.07717. · arXiv
  52. Schneider, 2019, Automated de novo drug design: are we nearly there yet?, Angew Chem Int Ed Engl, 58, 10792, 10.1002/anie.201814681 · DOI · Paper page
  53. Assmann M, Bal M, Craig M, D’Oyley J, Phillips L, Triendl H, et al. A novel machine learning approach uncovers new and distinctive inhibitors for cyclin-dependent kinase 9. bioRxiv. Published online March 19, 2020. http://dx.doi/org/10.1101/2020.03.18.996538. · DOI
  54. Perron Q, Mirguet O, Tajmouati H, Skiredj A, Rojas A, Gohier A, et al. Deep generative models for ligand-based de Novo design applied to multi–parametric optimization. ChemRxiv. Published online January 25, 2021. http://dx.doi/org/10.26434/chemrxiv.13622417.v2. · doi:10.26434/chemrxiv.13622417.v1
  55. Li, 2020, Chemical space exploration based on recurrent neural networks: applications in discovering kinase inhibitors, J Cheminform, 12, 42, 10.1186/s13321-020-00446-3 · DOI
  56. Grisoni F, Huisman B, Button A, Moret M, Atz K, Merk D, et al. Combining generative artificial intelligence and on-chip synthesis for de novo drug design. ChemRxiv. Published online December 30, 2020. http://dx.doi/org/10.26434/chemrxiv.13498587.v1. · doi:10.26434/chemrxiv.13498587.v1
  57. Kaushik, 2020, AI-driven drug discovery: a boon against COVID-19?, AI Open, 1, 1, 10.1016/j.aiopen.2020.07.001 · DOI
  58. Arús-Pous, 2020, SMILES-based deep generative scaffold decorator for de-novo drug design, J Cheminform, 12, 38, 10.1186/s13321-020-00441-8 · DOI
  59. Jin, 2020, Multi-objective molecule generation using interpretable substructures, Proc Mach Learn Res, 119, 4849
  60. Lim, 2020, Scaffold-based molecular design with a graph generative model, Chem Sci, 11, 1153, 10.1039/C9SC04503A · DOI
  61. Gao, 2020, The synthesizability of molecules proposed by generative models, J Chem Inf Model, 60, 5714, 10.1021/acs.jcim.0c00174 · DOI · Paper page
  62. Makara, 2021, Derivatization design of synthetically accessible space for optimization: in silico synthesis vs deep generative design, ACS Med Chem Lett, 12, 185, 10.1021/acsmedchemlett.0c00540 · DOI
  63. Vaucher, 2020, Automated extraction of chemical synthesis actions from experimental procedures, Nat Commun, 11, 3601, 10.1038/s41467-020-17266-6 · DOI
  64. Nicolaou, 2013, Multi-objective optimization methods in drug design, Drug Discov Today Technol, 10, e427, 10.1016/j.ddtec.2013.02.001 · DOI
  65. Grebner, 2020, Automated de novo design in medicinal chemistry: which types of chemistry does a generative neural network learn?, J Med Chem, 63, 8809, 10.1021/acs.jmedchem.9b02044 · DOI
  66. Ecemis, 2008, A drug candidate design environment using evolutionary computation, IEEE Trans Evol Comput, 12, 591, 10.1109/TEVC.2007.913131 · doi:10.1109/tevc.2007.913131
  67. Lameijer, 2006, The molecule evoluator. an interactive evolutionary algorithm for the design of drug-like molecules, J Chem Inf Model, 46, 545, 10.1021/ci050369d · DOI
  68. Renz, 2019, On failure modes in molecule generation and optimization, Drug Discov Today Technol, 32–33, 55, 10.1016/j.ddtec.2020.09.003 · DOI
  69. Brown, 2006, A novel workflow for the inverse QSPR problem using multiobjective optimization, J Comput Aided Mol Des, 20, 333, 10.1007/s10822-006-9063-1 · DOI
  70. Bush, 2020, A Turing test for molecular generators, J Med Chem, 63, 11964, 10.1021/acs.jmedchem.0c01148 · DOI
  71. Zhang J, Mercado R, Engkvist O, Chen H. Comparative study of deep generative models on chemical space coverage. ChemRxiv. Published online November 16, 2020. https://doi.org/10.26434/chemrxiv.13234289.v3. · doi:10.26434/chemrxiv.13234289.v1
  72. Coley, 2020, Autonomous discovery in the chemical sciences part II: outlook, Angew Chem Int Ed Engl, 59, 23414, 10.1002/anie.201909989 · DOI
  73. Cieplinski T, Danel T, Podlewska S, Jastrzebski S. We should at least be able to design molecules that dock well. arXiv [q-bioBM]. Published online June 20, 2020. http://arxiv.org/abs/2006.16955. · arXiv
  74. Nigam A, Friederich P, Krenn M, Aspuru-Guzik A. Augmenting genetic algorithms with deep neural networks for exploring the chemical space. arXiv [csNE]. Published online September 25, 2019. http://arxiv.org/abs/1909.11655. · arXiv
  75. Shi C, Xu M, Zhu Z, Zhang W, Zhang M, Tang J. GraphAF: a flow-based autoregressive model for molecular graph generation. arXiv [csLG]. Published online January 26, 2020. http://arxiv.org/abs/2001.09382. · arXiv
  76. Madhawa K, Ishiguro K, Nakago K, Abe M. GraphNVP: an invertible flow model for generating molecular graphs. arXiv [statML]. Published online May 28, 2019. http://arxiv.org/abs/1905.11600. · arXiv
  77. Xu M, Ran T, Chen H. De Novo molecule design through molecular generative model conditioned by 3D information of protein binding sites. ChemRxiv. Published online December 30, 2020. https://doi.org/10.26434/chemrxiv.13498332.v1. · doi:10.26434/chemrxiv.13498332.v1
  78. Simonovsky, 2020, DeeplyTough: learning structural comparison of protein binding sites, J Chem Inf Model, 60, 2356, 10.1021/acs.jcim.9b00554 · DOI
  79. Méndez-Lucio, 2020, De novo generation of hit-like molecules from gene expression signatures using artificial intelligence, Nat Commun, 11, 10, 10.1038/s41467-019-13807-w · DOI · Paper page
  80. Born, 2021, PaccMannRL: de novo generation of hit-like anticancer molecules from transcriptomic data via reinforcement learning, iScience, 24, 10.1016/j.isci.2021.102269 · DOI
  81. Zhavoronkov, 2020, Reply to “Assessing the impact of generative AI on medicinal chemistry”, Nat Biotechnol, 38, 10.1038/s41587-020-0417-3 · DOI
  82. Bajorath J. Faculty Opinions Recommendation of “Assessing the impact of generative AI on medicinal chemistry”. Faculty Opinions. Published online February 11, 2020; https://doi.org/10.3410/f.737286584.793570630. · doi:10.3410/f.737286584.793570630
  83. Chen, 2019, Has drug design augmented by artificial intelligence become a reality?, Trends Pharmacol Sci, 40, 806, 10.1016/j.tips.2019.09.004 · DOI

Cited by

Metadata

Authors
Meyers, J., Fabian, B., & Brown, N.
Venue
Drug Discovery Today
Publisher
Elsevier BV
Published
2021-11
Volume
26
Issue
11
Pages
2707-2715
ISSN
1359-6446 (ISSN-L)
DOI
10.1016/j.drudis.2021.05.019
OpenAlex
W3165171933
Cited by
292
References
8
FWCI
23.12
Type
journal article
Language
en
Open access
hybrid
Licence

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

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.