CoChat

MoleculeNet: a benchmark for molecular machine learning

Open access

The DOI matches its Crossref record; 82 of the 82 references it cites were found in a registry or an index.

Abstract

A large scale benchmark for molecular machine learning consisting of multiple public datasets, metrics, featurizations and learning algorithms.

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.

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

Integrity signals

Registry

Crossref · 10.1039/c7sc02664a

Cites

82 of 82 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Example code for benchmark evaluation with DeepChem, multiple methods are provided for data splitting, featurization and learning.
Fig. 1 Example code for benchmark evaluation with DeepChem, multiple methods are provided for data splitting, featurization and learning.CC BY-NC · PubMed Central
Fig. 2 Tasks in different datasets focus on different levels of properties of molecules.
Fig. 2 Tasks in different datasets focus on different levels of properties of molecules.CC BY-NC · PubMed Central
Fig. 3 Representation of data splits in MoleculeNet.
Fig. 3 Representation of data splits in MoleculeNet.CC BY-NC · PubMed Central
Fig. 4 Receiver operating characteristic (ROC) curves and precision recall curves (PRC) for predictions of logistic regression and graph convolutional models under different class imbalance condition (details listed in Table 2). (A, B) task “FDA_APPROVED” from ClinTox, test subset; (C, D) task “Hepatobiliary disorders” from SIDER, test subset; (E, F) task “NR-ER” from Tox21, validation subset; (G, H): task “HIV_active” from HIV, test subset. Black dashed lines are performances of random classifiers.
Fig. 4 Receiver operating characteristic (ROC) curves and precision recall curves (PRC) for predictions of logistic regression and graph convolutional models under different class imbalance condition (details listed in Table 2). (A, B) task “FDA_APPROVED” from ClinTox, test subset; (C, D) task “Hepatobiliary disorders” from SIDER, test subset; (E, F) task “NR-ER” from Tox21, validation subset; (G, H): task “HIV_active” from HIV, test subset. Black dashed lines are performances of random classifiers.CC BY-NC · PubMed Central
Fig. 5 Diagrams of featurizations in MoleculeNet.
Fig. 5 Diagrams of featurizations in MoleculeNet.CC BY-NC · PubMed Central
Fig. 6 Core structures of graph-based models implemented in MoleculeNet. To build features for the central dark green atom: (A) graph convolutional model: features are updated by combination with neighbour atoms; (B) directed acyclic graph model: all bonds are directed towards the central atom, features are propagated from the farthest atom to the central atom through directed bonds; (C) Weave model: pairs are formed between each pair of atoms (including not directly bonded pairs), features for the central atom are updated using all other atoms and their corresponding pairs, pair features are also updated by combination of the two pairing atoms; (D) message passing neural network: neighbour atoms' features are input into bondtype-dependent neural networks, forming outputs (messages). Features of the central atom are then updated using the outputs; (E) deep tensor neural network: no explicit bonding information is included, features are updated using all other atoms based on their corresponding physical distances; (F) ANI-1: features are built on distance information between pairs of atoms (radial symmetry functions) and angular information between triplets of atoms (angular symmetry functions).
Fig. 6 Core structures of graph-based models implemented in MoleculeNet. To build features for the central dark green atom: (A) graph convolutional model: features are updated by combination with neighbour atoms; (B) directed acyclic graph model: all bonds are directed towards the central atom, features are propagated from the farthest atom to the central atom through directed bonds; (C) Weave model: pairs are formed between each pair of atoms (including not directly bonded pairs), features for the central atom are updated using all other atoms and their corresponding pairs, pair features are also updated by combination of the two pairing atoms; (D) message passing neural network: neighbour atoms' features are input into bondtype-dependent neural networks, forming outputs (messages). Features of the central atom are then updated using the outputs; (E) deep tensor neural network: no explicit bonding information is included, features are updated using all other atoms based on their corresponding physical distances; (F) ANI-1: features are built on distance information between pairs of atoms (radial symmetry functions) and angular information between triplets of atoms (angular symmetry functions).CC BY-NC · PubMed Central
Fig. 7 Benchmark performances for biophysics tasks: PCBA, 4 models are evaluated by AUC-PRC on random split; MUV, 8 models are evaluated by AUC-PRC on random split; HIV, 8 models are evaluated by AUC-ROC on scaffold split; BACE, 9 models are evaluated by AUC-ROC on scaffold split. For AUC-ROC and AUC-PRC, higher value indicates better performance (to the right).
Fig. 7 Benchmark performances for biophysics tasks: PCBA, 4 models are evaluated by AUC-PRC on random split; MUV, 8 models are evaluated by AUC-PRC on random split; HIV, 8 models are evaluated by AUC-ROC on scaffold split; BACE, 9 models are evaluated by AUC-ROC on scaffold split. For AUC-ROC and AUC-PRC, higher value indicates better performance (to the right).CC BY-NC · PubMed Central
Fig. 8 Benchmark performances for physiology tasks: ToxCast, 8 models are evaluated by AUC-ROC on random split; Tox21, 9 models are evaluated by AUC-ROC on random split; BBBP, 9 models are evaluated by AUC-ROC on scaffold split; SIDER, 9 models are evaluated by AUC-ROC on random split. For AUC-ROC, higher value indicates better performance (to the right).
Fig. 8 Benchmark performances for physiology tasks: ToxCast, 8 models are evaluated by AUC-ROC on random split; Tox21, 9 models are evaluated by AUC-ROC on random split; BBBP, 9 models are evaluated by AUC-ROC on scaffold split; SIDER, 9 models are evaluated by AUC-ROC on random split. For AUC-ROC, higher value indicates better performance (to the right).CC BY-NC · PubMed Central
Fig. 9 Benchmark performances for physiology tasks: ClinTox, 9 models are evaluated by AUC-ROC on random split.
Fig. 9 Benchmark performances for physiology tasks: ClinTox, 9 models are evaluated by AUC-ROC on random split.CC BY-NC · PubMed Central
Fig. 10 Out-of-sample performances with different training set sizes on Tox21. Each datapoint is the average of 5 independent runs, with standard deviations shown as error bars.
Fig. 10 Out-of-sample performances with different training set sizes on Tox21. Each datapoint is the average of 5 independent runs, with standard deviations shown as error bars.CC BY-NC · PubMed Central
Fig. 11 Benchmark performances of PDBbind: 5 models are evaluated by RMSE on the three subsets: core, refined and full. Time split is applied to all three subsets. Noe that for RMSE, lower value indicates better performance (to the right).
Fig. 11 Benchmark performances of PDBbind: 5 models are evaluated by RMSE on the three subsets: core, refined and full. Time split is applied to all three subsets. Noe that for RMSE, lower value indicates better performance (to the right).CC BY-NC · PubMed Central
Fig. 12 Benchmark performances for physical chemistry tasks: ESOL, 8 models are evaluated by RMSE on random split; FreeSolv, 8 models are evaluated by RMSE on random split; lipophilicity, 8 models are evaluated by RMSE on random split. Note that for RMSE, lower value indicates better performance (to the right).
Fig. 12 Benchmark performances for physical chemistry tasks: ESOL, 8 models are evaluated by RMSE on random split; FreeSolv, 8 models are evaluated by RMSE on random split; lipophilicity, 8 models are evaluated by RMSE on random split. Note that for RMSE, lower value indicates better performance (to the right).CC BY-NC · PubMed Central
Fig. 13 Out-of-sample performances with different training set sizes on FreeSolv. Each datapoint is the average of 5 independent runs, with standard deviations shown as error bars.
Fig. 13 Out-of-sample performances with different training set sizes on FreeSolv. Each datapoint is the average of 5 independent runs, with standard deviations shown as error bars.CC BY-NC · PubMed Central
Fig. 14 Out-of-sample performances with different training set sizes on QM7. Each datapoint is the average of 5 independent runs, with standard deviations shown as error bars.
Fig. 14 Out-of-sample performances with different training set sizes on QM7. Each datapoint is the average of 5 independent runs, with standard deviations shown as error bars.CC BY-NC · PubMed Central
Fig. 15 Benchmark performances for quantum mechanics tasks: QM7, 8 models are evaluated by MAE on stratified split; QM7b, 3 models (QM7b only provides 3D coordinates) are evaluated by MAE on random split; QM8, 7 models are evaluated by MAE on random split; QM9, 5 models are evaluated by MAE on random split. Note that for MAE, lower value indicates better performance (to the right).
Fig. 15 Benchmark performances for quantum mechanics tasks: QM7, 8 models are evaluated by MAE on stratified split; QM7b, 3 models (QM7b only provides 3D coordinates) are evaluated by MAE on random split; QM8, 7 models are evaluated by MAE on random split; QM9, 5 models are evaluated by MAE on random split. Note that for MAE, lower value indicates better performance (to the right).CC BY-NC · PubMed Central

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.

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

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

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 82 references

Research works printed but not linked: 11

  1. Mitchell, 2014, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 4, 468
  2. T. Unterthiner , A.Mayr , G.Klambauer , M.Steijaert , J.Wenger , H.Ceulemans and S.Hochreiter , Deep Learning and Representation Learning Workshop (NIPS 2014) , 2014
  3. Delaney, 2004, J. Chem. Inf. Model., 44, 1000
  4. J. Deng , W.Dong , R.Socher , L.-J.Li , K.Li and L.Fei-Fei , CVPR09 , 2009
  5. A. Krizhevsky , I.Sutskever and G. E.Hinton , NIPS Proceedings , 2012
  6. Pedregosa, 2011, J. Mach. Learn. Res., 12, 2825
  7. Quantum Machine , http://quantum-machine.org/datasets/ , accessed 2017-09-27
  8. Medical Dictionary for Regulatory Activities , http://www.meddra.org/ , accessed 2017-09-27
  9. Artemov, 2016, bioRxiv, Biochem., 095653
  10. Aggregate Analysis of ClincalTrials.gov (AACT) Database , http://www.ctti-clinicaltrials.org/aact-database , accessed 2017-09-27
  11. Cortes, 1995, Mach. Learn., 20, 273

Cited by

Metadata

Authors
Wu, Z., Ramsundar, B., Feinberg, E. N., Gomes, J., Geniesse, C., Pappu, A. S., Leswing, K., & Pande, V.
Affiliations
  • Stanford University (US)
  • Stanford Medicine (US)
  • Schrodinger (United States) (US)
Corresponding author
Pande, Vijay
Venue
Chemical Science
Published
2018
Volume
9
Issue
2
Pages
513-530
ISSN
  • 2041-6520 (ISSN-L)
  • 2041-6539
OpenAlex
W2594183968
Cited by
3,107
References
62
Citations by year
2026: 553 · 2025: 561 · 2024: 485 · 2023: 387 · 2022: 328 · 2021: 309 · 2020: 251 · 2019: 144 · 2018: 73
FWCI
98.85
Type
journal article
Language
en
Open access
Locations
  • Chemical Science (publishedVersion, cc-by-nc)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • The HKU Scholars Hub (University of Hong Kong) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • Hertz Foundation
  • NIH (R01 GM062868, U19 AI109662)
Topics
  • Life Sciences › Pharmacology, Toxicology and Pharmaceutics › Drug Discovery › Computational Drug Discovery Methods
  • Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
  • Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Advanced Chemical Physics Studies
Keywords
molecular machine learning, benchmark, MoleculeNet, molecular property prediction, DeepChem, open source software, data scarcity, imbalanced classification, quantum mechanical properties, biophysical properties

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.