Environmental applications of molecular graph learning: Graph neural network based prediction of partition coefficients

Abstract

In cheminformatics, predicting molecular properties is crucial for enhancing material research, toxicity assessment, and drug discovery. This research investigates the use of graph neural networks (GNNs) for predicting molecular properties by examining three different architectures: graph convolutional networks (GCNs), graph isomorphism networks (GINs), and graph attention networks (GATs). Employing molecular graph information, these models are evaluated on the MUTEG dataset and measured against key metrics such as accuracy and area under the receiver operating characteristic curve (AUC). Our experimental findings show that GIN has the highest accuracy at 89.2%, exceeding GCN (87.5%) and GAT (88.3%). GIN also achieves the highest AUC of 0.89, whereas the AUCs of GCN and GAT are 0.84 and 0.86, respectively, indicating GIN’s enhanced ability to effectively model graph isomorphisms. We selected GIN for this study because of its proven theoretical and empirical strength in capturing graph-level representations, particularly in domains such as cheminformatics, where molecular structures are naturally modeled as graphs. These results highlight the efficacy of GNNs in predicting molecular properties and position GIN as a favored framework for tasks that demand accurate graph feature extraction. This study further plays a pivotal role in understanding the environmental fate and transport of chemical compounds. We used GIN to identify partition coefficients such as the octanol-water partition coefficient, air-water partition coefficient, and soil-water partition coefficient from the MoleculeNet dataset.

Keywords

Graph neural networks / Molecular property prediction / Cheminformatics / Drug discovery / Structure-property relationships

Funding

None.

Conflict of interest

The authors declare no competing interests.

References

[1] Mayr A, Klambauer G, Unterthiner T, Hochreiter S. DeepTox: Toxicity prediction using deep learning. Fron Environ Sci. 2016;3:80. doi: 10.3389/fenvs.2015.00080

[2] Wu Z, Ramsundar B, Feinberg EN, et al. MoleculeNet: A benchmark for molecular machine learning. Chem Sci. 2018; 9(2):513-530. doi: 10.1039/c7sc02664a

[3] Debnath AK, Lopez de Compadre RL, Debnath G, Shusterman AJ, Hansch C. Structure-activity relationship of mutagenic aromatic and heteroaromatic nitro compounds. Correlation with molecular orbital energies and hydrophobicity. J Med Chem. 1991; 34(2):786-797. doi: 10.1021/jm00106a046

[4] Kipf TN, Welling M. Semi-Supervised Classification with Graph Convolutional Networks. In: 5th International Conference on Learning Representations (ICLR); 2017.

[5] Xu K, Hu W, Leskovec J, Jegelka S. How Powerful are Graph Neural Networks? In: International Conference on Learning Representations; 2018.

[6] Ramakrishnan R, Dral PO, Rupp M, Von Lilienfeld OA. Quantum chemistry structures and properties of 134 kilo molecules. Sci Data. 2014; 1(1):140022. doi: 10.1038/sdata.2014.22

[7] Irwin JJ, Shoichet BK. ZINC-A free database of commercially available compounds for virtual screening. J Chem Inf Model. 2005; 45:177-182. doi: 10.1021/ci049714+

[8] Zeng XL, Wang HJ, Wang Y. QSPR models of n-octanol/ water partition coefficients and aqueous solubility of halogenated methyl-phenyl ethers by DFT method. Chemosphere. 2012; 86(6):619-625. doi: 10.1016/j.chemosphere.2011.10.051

[9] Wu Z, Ramsundar B, Feinberg EN, et al. MoleculeNet: A benchmark for molecular machine learning. Chem Sci. 2018; 9(2):513-530. doi: 10.1039/C7SC02664A

[10] Gilmer J, Schoenholz SS, Riley PF, Vinyals O, Dahl GE. Neural Message Passing for Quantum Chemistry. In: Proceedings of the 34th International Conference on Machine Learning; 2017.

[11] Amjath M, Henna S, Rathnayake U. Graph representation federated learning for malware detection in Internet of health things. Results Eng. 2025;25:103651. doi: 10.1016/j.rineng.2024.103651

[12] Ellis LM, Mobley PE. Machine learning approaches for predicting partition coefficients: A comparison of deep learning and classical methods. J Chem Inf Model. 2021; 61(9):4451-4464.

[13] Nevolianis T, Rittig JG, Mitsos A, Leonhard K. Multi-fidelity Graph Neural Networks for Predicting Toluene/ Water Partition Coefficients [Chemrxiv Preprint]; 2024. doi: 10.26434/chemrxiv-2024-3t818

[14] Sudhakar M, Jai A. Explainable AI for molecular property prediction: Enhancing interpretability of deep learning models. ACS Omega. 2022; 7(4):2931-2942.

[15] Duvenaud D, Maclaurin D, Aguilera-Iparraguirre J, et al. Convolutional Networks on Graphs for Learning Molecular Fingerprints. In: NIPS’15: Proceedings of the 29th International Conference on Neural Information Processing Systems; 2015.

[16] Scarselli F, Gori M, Tsoi AC, Hagenbuchner M, Monfardini G. The graph neural network model. IEEE Trans Neural Netw. 2009; 20(1):61-80.

[17] Shervashidze N, Schweitzer P, Van Leeuwen EJ, Mehlhorn K, Borgwardt KM. Weisfeiler-lehman graph kernels. J Mach Learn Res. 2011; 12:2539-2561.

[18] Veličković P, Casanova A, Liò P, Cucurull G, Romero A, Bengio Y. Graph Attention Networks. In: 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings; 2018. p. 1-12. doi: 10.1007/978-3-031-01587-8_7

[19] Schütt KT, Arbabzadah F, Chmiela S, Müller KR, Tkatchenko A. Quantum-chemical insights from deep tensor neural networks. Nat Commun. 2017; 8(1):13890. doi: 10.1038/ncomms13890

[20] Kang X, Hu B, Perdana MC, Zhao Y, Chen Z. Extreme learning machine models for predicting the n-octanol/ water partition coefficient (Kow) data of organic compounds. J Environ Chem Eng. 2022; 10(6):108552. doi: 10.1016/j.jece.2022.108552

[21] Kenney DH, Paffenroth RC, Timko MT, Teixeira AR. Dimensionally reduced machine learning model for predicting single component octanol-water partition coefficients. J Cheminform. 2023; 15(1):9. doi: 10.1186/s13321-022-00660-1

[22] Yokogawa D, Suda K. Interpretable Attribution Assignment for Octanol-Water Partition Coefficient. [Chemrxiv Preprint]; 2023. doi: 10.26434/chemrxiv-2023-2lwsh-v2

[23] Zamora WJ, Viayna A, Pinheiro S, et al. Prediction of toluene/water partition coefficients in the SAMPL9 blind challenge: Assessment of machine learning and IEF-PCM/MST continuum solvation models. Phys Chem Chem Physics. 2023; 25(27):17952-17965. doi: 10.1039/D3CP01428B

[24] Ma W, Wang M, Jiang R, Chen W. A machine learning based approach for estimating site-specific partition coefficient Kd of organic compounds: Application to nonionic pesticides. Environ Pollut. 2023;323:121297. doi: 10.1016/j.envpol.2023.121297

[25] Zhu Q, Jia Q, Liu Z, et al. Molecular partition coefficient from machine learning with polarization and entropy embedded atom-centered symmetry functions. Phys Chem Chem Phys. 2022; 24(38):23082-23088. doi: 10.1039/D2CP02648A

[26] Ebert RU, Kühne R, Schüürmann G. Octanol/air partition coefficient-a general-purpose fragment model to predict log Koa from molecular structure. Environ Sci Technol. 2023; 57(2):976-984. doi: 10.1021/acs.est.2c06170

[27] Stojanova M, Barré P, Clivot H, et al. A New Machine-Learning Model to Partition Soil Organic Carbon into its Centennially Stable and Active Fractions Based on Rock-Eval(r) Thermal Analysis. Germany: European Geosciences Union; 2025.doi: 10.5194/egusphere-egu24-11107

[28] Torralba-Sanchez TL, Di Toro DM, Dmitrenko O, Murillo-Gelvez J, Tratnyek PG. Modeling the partitioning of anionic carboxylic and perfluoroalkyl carboxylic and sulfonic acids to octanol and membrane lipid. Environ Toxicol Chem. 2023; 42(11):2317-2328. doi: 10.1002/etc.5716

[29] Khawar MI, Mahmood A, Nabi D. Exploring the role of octanol-water partition coefficient and Henry’s law constant in predicting the lipid-water partition coefficients of organic chemicals. Sci Rep. 2022; 12(1):14936. doi: 10.1038/s41598-022-19452-6

[30] Patel C, Roy D. Octanol-water partition coefficients of fluorinated drug molecules with continuum solvation models. J Phys Chem A. 2022; 126(26):4185-4190. doi: 10.1021/acs.jpca.2c02172

[31] Baskaran S, Podagatlapalli A, Sangion A, Wania F. Predicting the temperature dependence of the octanol-air partition ratio: A new model for estimating δU circOAδUOA circ. J Solut Chem. 2023; 52(1):51-69. doi: 10.1007/s10953-022-01214-7

[32] Singh B, Crasto M, Ravi K, Singh S. Pharmaceutical advances: Integrating artificial intelligence in QSAR, combinatorial and green chemistry practices. Intell Pharm. 2024; 2:598-608. doi: 10.1016/j.ipha.2024.05.005

[33] Arab M, Faramarz MG, Hashim K. Applications of computational and statistical models for optimizing the electrochemical removal of cephalexin antibiotic from water. Water (Switzerland). 2022; 14(3):344. doi: 10.3390/w14030344

[34] Ågerstrand M, Berg C, Björlenius B, et al. Improving environmental risk assessment of human pharmaceuticals. Environ Sci Technol. 2015; 49(9):5336-5345. doi: 10.1021/acs.est.5b00302

[35] Soares TA, Nunes-Alves A, Mazzolari A, Ruggiu F, Wei GW, Merz K. The (Re)-evolution of quantitative structure-activity relationship (QSAR) studies propelled by the surge of machine learning methods. J Chem Inf Model. 2022; 62(22):5317-5320. doi: 10.1021/acs.jcim.2c01422

[36] Shen X, Wang R, Xiong X, et al. Metabolic reaction network-based recursive metabolite annotation for untargeted metabolomics. Nat Commun. 2019; 10(1):1516. doi: 10.1038/s41467-019-09550-x

[37] Luo Y, Zhao X, Zhou J, et al. A network integration approach for drug-target interaction prediction and computational drug repositioning from heterogeneous information. Nat Commun. 2017; 8(1):573. doi: 10.1038/s41467-017-00680-8

[38] Vora LK, Gholap AD, Jetha K, Thakur RRS, Solanki HK, Chavda VP. Artificial intelligence in pharmaceutical technology and drug delivery design. Pharmaceutics. 2023; 15(7): 1916. doi: 10.3390/pharmaceutics15071916

[39] Han R, Yoon H, Kim G, Lee H, Lee Y. Revolutionizing medicinal chemistry: The application of artificial intelligence (AI) in early drug discovery. Pharmaceuticals (Basel). 2023; 16(9):1259. doi: 10.3390/ph16091259

[40] Environment news futures. Asian J Water Environ Pollut. 2024; 21(5):95-98. doi: 10.3233/AJW240064

[41] Lallawmzuali G, Devi AS, Liana T, Hriatsaka V, Singh AP, Lalhriatpuia C. Assessment of the heavy metal contaminations of roadside soil in Aizawl, Mizoram (India): An in-depth analysis utilising advanced scientific methodologies. Asian J Water Environ Pollut. 2024; 21(5):37-47. doi: 10.3233/AJW240058

[42] Kaur I, Gulati A, Lamba PS, Jain A, Taneja H, Syal JS. Water quality assessment using machine learning: A focus on coliform prediction in water. Asian J Water Environ Pollut. 2024; 21(5):19-26. doi: 10.3233/AJW240056

[43] Xu L, Dong Z, Fang L, et al. OrthoVenn2: A web server for whole-genome comparison and annotation of orthologous clusters across multiple species. Nucleic Acids Res. 2019; 47(W1):W52-W58. doi: 10.1093/nar/gkz333

[44] Huerta-Cepas J, Szklarczyk D, Heller D, et al. EggNOG 5.0: A hierarchical, functionally and phylogenetically annotated orthology resource based on 5090 organisms and 2502 viruses. Nucleic Acids Res. 2019; 47(D1):D309-D314. doi: 10.1093/nar/gky1085

[45] Altenhoff AM, Vesztrocy AW, Bernard C, et al. OMA orthology in 2024: Improved prokaryote coverage, ancestral and extant GO enrichment, a revamped synteny viewer and more in the OMA ecosystem. Nucleic Acids Res. 2024; 52(D1):D513-D521. doi: 10.1093/nar/gkad1020

[46] Zdobnov EM, Kuznetsov D, Tegenfeldt F, Manni M, Berkeley M, Kriventseva EV. OrthoDB in 2020: Evolutionary and functional annotations of orthologs. Nucleic Acids Res. 2021; 49(D1):D389-D393. doi: 10.1093/nar/gkaa1009

[47] Neves BJ, Braga RC, Melo-Filho CC, Moreira-Filho JT, Muratov EN, Andrade CH. QSAR-based virtual screening: Advances and applications in drug discovery. Front Pharmacol. 2018;9:1275. doi: 10.3389/fphar.2018.01275

[48] Dintakurthy Y, Krishna Innmuri R, Vanteru A, ThotakurixA. Emerging Applications of Artificial Intelligence in Edge Computing: A Comprehensive Review. Vol. 1. Germany: Springer; 2024.

[49] Zielezinski A, Dziubek M, Sliski J, Karlowski WM. ORCAN - a web-based meta-server for real-time detection and functional annotation of orthologs. Bioinformatics. 2017; 33(8):1224-1226. doi: 10.1093/bioinformatics/btw825

[50] Mohamed SK, Nounu A, Nováček V. Biological applications of knowledge graph embedding models. Brief Bioinform. 2021; 22(2):1679-1693. doi: 10.1093/bib/bbaa012

[51] Price OR, Hughes GO, Roche NL, Mason PJ. Improving emissions estimates of home and personal care products ingredients for use in EU risk assessments. Integr Environ Assess Manag. 2010; 6(4):677-684. doi: 10.1002/ieam.88

[52] Martinelli DD. Machine learning for metabolomics research in drug discovery. Intell Based Med. 2023;8:100101. doi: 10.1016/j.ibmed.2023.100101

[53] Solanke AV, Kumar Verma S, Kumar S, Oyinna B, Okedu KE. MPPT for hybrid energy system using machine learning techniques. J Mod Technol. 2024; 1:19-37. doi: 10.71426/jmt.v1.i1.pp19-37

[54] Nguyen G, Šipková V, Dlugolinsky S, Nguyen BM, Tran V, Hluchý L. A comparative study of operational engineering for environmental and compute-intensive applications. Array. 2021;12:100096. doi: 10.1016/j.array.2021.100096

[55] Yin H, Duo H, Li S, et al. Unlocking biological insights from differentially expressed genes: Concepts, methods, and future perspectives. J Adv Res. 2024. doi: 10.1016/j.jare.2024.12.004

[56] Badawi AK, Hassan R. Optimizing sludge extract reuse from physico-chemical processes for zero-waste discharge: A critical review. Desalination Water Treat. 2024;319:100527. doi: 10.1016/j.dwt.2024.100527

[57] Badawi AK, Hassan R, Alghamdi AM, Ismail B, Osman RM, Salama RS. Advancing cobalt ferrite-supported activated carbon from orange peels for real pulp and paper mill wastewater treatment. Desalination Water Treat. 2024;318:100331. doi: 10.1016/j.dwt.2024.100331

[58] Kingma DP, Ba J. Adam: A Method for Stochastic Optimization; 2014.