A Graph Neural Network Approach for Molecule Carcinogenicity Prediction

Overview

Journal Bioinformatics

Publisher Oxford University Press

Specialty Biology

Date 2022 Jun 27

PMID 35758812

Authors

Philip Fradkin

Adamo Young

Lazar Atanackovic

Brendan Frey

Leo J Lee

Bo Wang

Affiliations

Soon will be listed here.

Abstract

Motivation: Molecular carcinogenicity is a preventable cause of cancer, but systematically identifying carcinogenic compounds, which involves performing experiments on animal models, is expensive, time consuming and low throughput. As a result, carcinogenicity information is limited and building data-driven models with good prediction accuracy remains a major challenge.

Results: In this work, we propose CONCERTO, a deep learning model that uses a graph transformer in conjunction with a molecular fingerprint representation for carcinogenicity prediction from molecular structure. Special efforts have been made to overcome the data size constraint, such as multi-round pre-training on related but lower quality mutagenicity data, and transfer learning from a large self-supervised model. Extensive experiments demonstrate that our model performs well and can generalize to external validation sets. CONCERTO could be useful for guiding future carcinogenicity experiments and provide insight into the molecular basis of carcinogenicity.

Availability And Implementation: The code and data underlying this article are available on github at https://github.com/bowang-lab/CONCERTO.

Citing Articles

Deep active learning with high structural discriminability for molecular mutagenicity prediction.

Xu H, Zhao Y, Zhang Y, Han J, Zan P, He S Commun Biol. 2024; 7(1):1071.

PMID: 39217273 PMC: 11366013. DOI: 10.1038/s42003-024-06758-6.

Transformer models in biomedicine.

Madan S, Lentzen M, Brandt J, Rueckert D, Hofmann-Apitius M, Frohlich H BMC Med Inform Decis Mak. 2024; 24(1):214.

PMID: 39075407 PMC: 11287876. DOI: 10.1186/s12911-024-02600-5.

Advancing Drug Safety in Drug Development: Bridging Computational Predictions for Enhanced Toxicity Prediction.

Amorim A, Piochi L, Gaspar A, Preto A, Rosario-Ferreira N, Moreira I Chem Res Toxicol. 2024; 37(6):827-849.

PMID: 38758610 PMC: 11187637. DOI: 10.1021/acs.chemrestox.3c00352.

Using test-time augmentation to investigate explainable AI: inconsistencies between method, model and human intuition.

Hartog P, Kruger F, Genheden S, Tetko I J Cheminform. 2024; 16(1):39.

PMID: 38576047 PMC: 10993590. DOI: 10.1186/s13321-024-00824-1.

Graph Neural Networks in Cancer and Oncology Research: Emerging and Future Trends.

Gogoshin G, Rodin A Cancers (Basel). 2023; 15(24).

PMID: 38136405 PMC: 10742144. DOI: 10.3390/cancers15245858.

References

Wolf D, Cohen S, Boobis A, Dellarco V, Fenner-Crisp P, Moretto A . Chemical carcinogenicity revisited 1: A unified theory of carcinogenicity based on contemporary knowledge. Regul Toxicol Pharmacol. 2019; 103:86-92. DOI: 10.1016/j.yrtph.2019.01.021. View

Singh K, Gupta S, Rai P . Predicting carcinogenicity of diverse chemicals using probabilistic neural network modeling approaches. Toxicol Appl Pharmacol. 2013; 272(2):465-75. DOI: 10.1016/j.taap.2013.06.029. View

Onakpoya I, Heneghan C, Aronson J . Post-marketing withdrawal of 462 medicinal products because of adverse drug reactions: a systematic review of the world literature. BMC Med. 2016; 14:10. PMC: 4740994. DOI: 10.1186/s12916-016-0553-2. View

Gold L, Manley N, Slone T, Rohrbach L, Garfinkel G . Supplement to the Carcinogenic Potency Database (CPDB): results of animal bioassays published in the general literature through 1997 and by the National Toxicology Program in 1997-1998. Toxicol Sci. 2005; 85(2):747-808. DOI: 10.1093/toxsci/kfi161. View

Doe J, Boobis A, Dellarco V, Fenner-Crisp P, Moretto A, Pastoor T . Chemical carcinogenicity revisited 2: Current knowledge of carcinogenesis shows that categorization as a carcinogen or non-carcinogen is not scientifically credible. Regul Toxicol Pharmacol. 2019; 103:124-129. DOI: 10.1016/j.yrtph.2019.01.024. View

Maggiora G, Vogt M, Stumpfe D, Bajorath J . Molecular similarity in medicinal chemistry. J Med Chem. 2013; 57(8):3186-204. DOI: 10.1021/jm401411z. View

Fjodorova N, Vracko M, Tusar M, Jezierska A, Novic M, Kuhne R . Quantitative and qualitative models for carcinogenicity prediction for non-congeneric chemicals using CP ANN method for regulatory uses. Mol Divers. 2009; 14(3):581-94. DOI: 10.1007/s11030-009-9190-4. View

Hansen K, Mika S, Schroeter T, Sutter A, Ter Laak A, Steger-Hartmann T . Benchmark data set for in silico prediction of Ames mutagenicity. J Chem Inf Model. 2009; 49(9):2077-81. DOI: 10.1021/ci900161g. View

Li S, Zhang L, Feng H, Meng J, Xie D, Yi L . MutagenPred-GCNNs: A Graph Convolutional Neural Network-Based Classification Model for Mutagenicity Prediction with Data-Driven Molecular Fingerprints. Interdiscip Sci. 2021; 13(1):25-33. DOI: 10.1007/s12539-020-00407-2. View

10.

Balmain A . The critical roles of somatic mutations and environmental tumor-promoting agents in cancer risk. Nat Genet. 2020; 52(11):1139-1143. PMC: 8360498. DOI: 10.1038/s41588-020-00727-5. View

11.

Wu X, Zhang Q, Wang H, Hu J . Predicting carcinogenicity of organic compounds based on CPDB. Chemosphere. 2015; 139:81-90. DOI: 10.1016/j.chemosphere.2015.05.056. View

12.

Smietana K, Siatkowski M, Moller M . Trends in clinical success rates. Nat Rev Drug Discov. 2016; 15(6):379-80. DOI: 10.1038/nrd.2016.85. View

13.

Durant J, Leland B, Henry D, Nourse J . Reoptimization of MDL keys for use in drug discovery. J Chem Inf Comput Sci. 2002; 42(6):1273-80. DOI: 10.1021/ci010132r. View

14.

Benigni R, Bossa C, Richard A, Yang C . A novel approach: chemical relational databases, and the role of the ISSCAN database on assessing chemical carcinogenicity. Ann Ist Super Sanita. 2008; 44(1):48-56. View

15.

Kazius J, McGuire R, Bursi R . Derivation and validation of toxicophores for mutagenicity prediction. J Med Chem. 2005; 48(1):312-20. DOI: 10.1021/jm040835a. View

16.

Bailer A, Portier C . An index of tumorigenic potency. Biometrics. 1993; 49(2):357-65. View

17.

Wang Y, Huang L, Jiang S, Li K, Zou J, Yang S . CapsCarcino: A novel sparse data deep learning tool for predicting carcinogens. Food Chem Toxicol. 2019; 135:110921. DOI: 10.1016/j.fct.2019.110921. View

18.

Scarselli F, Gori M, Tsoi A, Hagenbuchner M, Monfardini G . The graph neural network model. IEEE Trans Neural Netw. 2008; 20(1):61-80. DOI: 10.1109/TNN.2008.2005605. View

19.

Walmsley R, Billinton N . How accurate is in vitro prediction of carcinogenicity?. Br J Pharmacol. 2010; 162(6):1250-8. PMC: 3058158. DOI: 10.1111/j.1476-5381.2010.01131.x. View

20.

Wellawatte G, Seshadri A, White A . Model agnostic generation of counterfactual explanations for molecules. Chem Sci. 2022; 13(13):3697-3705. PMC: 8966631. DOI: 10.1039/d1sc05259d. View