On failure modes in molecule generation and optimization
Open accessThe DOI matches its Crossref record; 18 of the 18 references it cites were found in a registry or an index.
Abstract
There has been a wave of generative models for molecules triggered by advances in the field of Deep Learning. These generative models are often used to optimize chemical compounds towards particular properties or a desired biological activity. The evaluation of generative models remains challenging and suggested performance metrics or scoring functions often do not cover all relevant aspects of drug design projects. In this work, we highlight some unintended failure modes in molecular generation and optimization and how these evade detection by current performance metrics.
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.
Integrity signals
Registry
Cites
Statistics
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.
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
Breiman, L. (2001) · Machine Learning · doi:10.1023/a:1010933404324 · Cited by 133,538
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318
Self-citationLeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
The graph neural network model
Scarselli, F., Gori, M., Tsoi, A. C., Hagenbuchner, M., & Monfardini, G. (2009) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2008.2005605 · Cited by 9,582
Intriguing properties of neural networks
Szegedy, C., Zaremba, W., Sutskever, I., Bruna, J., Erhan, D., Goodfellow, I., & Fergus, R. (2013) · arXiv · doi:10.48550/arxiv.1312.6199 · Cited by 8,143
Weininger, D. (1988) · Journal of Chemical Information and Computer Sciences · doi:10.1021/ci00057a005 · Cited by 7,436 · cited 1×
“The first wave of Deep Learning methods for molecule generation [1,2] used the SMILES [13] representation of compounds in combination with recurrent neural networks (RNNs) [14,15], which are well suited to process sequences.”
MethodDeep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images
Nguyen, A., Yosinski, J., & Clune, J. (2014) · arXiv · doi:10.48550/arxiv.1412.1897 · Cited by 3,258
Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules
Gómez-Bombarelli, R., Wei, J. N., Duvenaud, D., Hernández-Lobato, J. M., Sánchez-Lengeling, B., Sheberla, D., Aguilera-Iparraguirre, J., Hirzel, T. D., Adams, R. P., & Aspuru-Guzik, A. (2018) · ACS Central Science · doi:10.1021/acscentsci.7b00572 · Cited by 3,106
Deep learning enables rapid identification of potent DDR1 kinase inhibitors
Zhavoronkov, A., Ivanenkov, Y. A., Aliper, A., Veselov, M. S., Aladinskiy, V. A., Aladinskaya, A. V., Terentiev, V. A., Polykovskiy, D. A., Kuznetsov, M. D., Asadulaev, A., et al. (2019) · Nature Biotechnology · doi:10.1038/s41587-019-0224-x · Cited by 1,573 · cited 1×
“Zhavoronkov et al. [31] identified a DDR1 kinase inhibitor using generative Deep Learning models.”
MethodGenerating Focused Molecule Libraries for Drug Discovery with Recurrent Neural Networks
Segler, M. H. S., Kogej, T., Tyrchan, C., & Waller, M. P. (2018) · ACS Central Science · doi:10.1021/acscentsci.7b00512 · Cited by 1,537
GuacaMol: Benchmarking Models for de Novo Molecular Design
Brown, N., Fiscato, M., Segler, M. H. S., & Vaucher, A. C. (2019) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.8b00839 · Cited by 895
Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models
Polykovskiy, D., Zhebrak, A., Sanchez-Lengeling, B., Golovanov, S., Tatanov, O., Belyaev, S., Kurbanov, R., Artamonov, A., Aladinskiy, V., Veselov, M., et al. (2020) · Frontiers in Pharmacology · doi:10.3389/fphar.2020.565644 · Cited by 849 · cited 3×
“The FCD is also included in more comprehensive benchmarking suites [20,22].”
BackgroundMethodDeep learning for molecular design—a review of the state of the art
Elton, D. C., Boukouvalas, Z., Fuge, M. D., & Chung, P. W. (2019) · Molecular Systems Design & Engineering · doi:10.1039/c9me00039a · Cited by 683 · cited 1×
“For a more detailed overview on current approaches, see [3,19].”
BackgroundKadurin, A., Nikolenko, S., Khrabrov, K., Aliper, A., & Zhavoronkov, A. (2017) · Molecular Pharmaceutics · doi:10.1021/acs.molpharmaceut.7b00346 · Cited by 587 · cited 1×
“Examples of distribution-learning models for molecules include auto-regressive approaches [1,32], variational auto-encoders [2,23,33], generative adversarial networks [34,35], and adversarial auto-encoders [36,37].”
BackgroundLarge-scale comparison of machine learning methods for drug target prediction on ChEMBL
Mayr, A., Klambauer, G., Unterthiner, T., Steijaert, M., Wegner, J. K., Ceulemans, H., Clevert, D.-A., & Hochreiter, S. (2018) · Chemical Science · doi:10.1039/c8sc00148k · Cited by 584
Self-citationDe Novo Design of Bioactive Small Molecules by Artificial Intelligence
Merk, D., Friedrich, L., Grisoni, F., & Schneider, G. (2018) · Molecular Informatics · doi:10.1002/minf.201700153 · Cited by 373
Lehman, J., Clune, J., Misevic, D., Adami, C., Altenberg, L., Beaulieu, J., Bentley, P. J., Bernard, S., Beslon, G., Bryson, D. M., et al. (2020) · Artificial Life · doi:10.1162/artl_a_00319 · Cited by 181 · cited 1×
“This phenomenon of optimization of a score in unexpected ways, has been observed in other applications [39].”
ResultMachine Learning in Drug Discovery
Hochreiter, S., Klambauer, G., & Rarey, M. (2018) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.8b00478 · Cited by 36
Self-citation
Cited by
Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction
Passaro, S., Corso, G., Wohlwend, J., Reveiz, M., Thaler, S., Somnath, V. R., Getz, N., Portnoi, T., Roy, J., Stark, H., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.06.14.659707 · Cited by 665
Artificial intelligence in drug development
Zhang, K., Yang, X., Wang, Y., Yu, Y., Huang, N., Li, G., Li, X., Wu, J. C., & Yang, S. (2025) · Nature Medicine · doi:10.1038/s41591-024-03434-4 · Cited by 498
REINVENT 2.0: An AI Tool for De Novo Drug Design
Blaschke, T., Arús-Pous, J., Chen, H., Margreitter, C., Tyrchan, C., Engkvist, O., Papadopoulos, K., & Patronov, A. (2020) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.0c00915 · Cited by 496
Generative models for molecular discovery: Recent advances and challenges
Bilodeau, C., Jin, W., Jaakkola, T., Barzilay, R., & Jensen, K. F. (2022) · WIREs Computational Molecular Science · doi:10.1002/wcms.1608 · Cited by 318
Generative Models as an Emerging Paradigm in the Chemical Sciences
Anstine, D. M., & Isayev, O. (2023) · Journal of the American Chemical Society · doi:10.1021/jacs.2c13467 · Cited by 313
Metadata
- Authors
- Renz, P., Van Rompaey, D., Wegner, J. K., Hochreiter, S., & Klambauer, G.
- Venue
- Drug Discovery Today: Technologies
- Publisher
- Elsevier BV
- Published
- 2019-12
- Volume
- 32-33
- Pages
- 55-63
- ISSN
- 1740-6749 (ISSN-L)
- OpenAlex
- W3094553402
- Cited by
- 154
- References
- 18
- FWCI
- 9.99
- Type
- journal article
- Language
- en
- Open access
- hybrid
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.