Graph Neural Network Prediction of Antioxidant Activity of Polyphenols from Grape Pomace

References

[1] Tournour HH, Segundo MA, Magalhães LM, Barreiros L, Queiroz J, Cunha LM. Valorization of grape pomace: extraction of bioactive phenolics with antioxidant properties. Ind Crops Prod 2015;74:397-406. ScienceDirect

[2] Moutinho J, Gouvinhas I, Domínguez-Perles R, Barros A. Optimization of the extraction methodology of grape pomace polyphenols for food applications. Molecules 2023;28:3885. MDPI

[3] Zhu MT, Huang YS, Wang YL, Shi T, Zhang LL, Chen Y, et al. Comparison of (poly)phenolic compounds and antioxidant properties of pomace extracts from kiwi and grape juice. Food Chem 2019;271:425-32. ScienceDirect

[4] González-Centeno MR, Jourdes M, Femenia A, Simal S, Rosselló C, Teissedre PL. Characterization of polyphenols and antioxidant potential of white grape pomace byproducts (Vitis vinifera L.). J Agric Food Chem 2013;61:11579-87. ACS

[5] Rockenbach II, Rodrigues E, Gonzaga LV, Caliari V, Genovese MI, Gonçalves AESS, et al. Phenolic compounds content and antioxidant activity in pomace from selected red grapes (Vitis vinifera L. and Vitis labrusca L.) widely produced in Brazil. Food Chem 2011;127:174-9. ScienceDirect

[6] Rockenbach II, Gonzaga LV, Rizelio VM, Gonçalves AESS, Genovese MI, Fett R. Phenolic compounds and antioxidant activity of seed and skin extracts of red grape (Vitis vinifera and Vitis labrusca) pomace from Brazilian winemaking. Food Res Int 2011;44:897-901. ScienceDirect

[7] Pinelo M, Manzocco L, Nuñez MJ, Nicoli MC. Interaction among phenols in food fortification: negative synergism on antioxidant capacity. J Agric Food Chem 2004;52:1177-80. ACS

[8] Pinelo M, Rubilar M, Jerez M, Sineiro J, Núñez MJ. Effect of solvent, temperature, and solvent-to-solid ratio on the total phenolic content and antiradical activity of extracts from different components of grape pomace. J Agric Food Chem 2005;53:2111-7. ACS

[9] Ruberto G, Renda A, Daquino C, Amico V, Spatafora C, Tringali C, et al. Polyphenol constituents and antioxidant activity of grape pomace extracts from five Sicilian red grape cultivars. Food Chem 2007;100:203-10. ScienceDirect

[10] Negro C, Aprile A, Luvisi A, De Bellis L, Miceli A. Antioxidant activity and polyphenols characterization of four monovarietal grape pomaces from Salento (Apulia, Italy). Antioxidants 2021;10:1406. MDPI

[11] Radulescu C, Olteanu RL, Buruleanu CL, Nechifor (Tudorache) M, Dulama ID, Stirbescu RM, et al. Polyphenolic screening and the antioxidant activity of grape pomace extracts of Romanian white and red grape varieties. Antioxidants 2024;13:1133. MDPI

[12] Di Lorenzo C, Colombo F, Biella S, Orgiu F, Frigerio G, Regazzoni L, et al. Phenolic profile and antioxidant activity of different grape (Vitis vinifera L.) varieties. BIO Web Conf 2019;12:04005. BIO Web of Conferences

[13] Bors W, Michel C. Chemistry of the antioxidant effect of polyphenols. Ann N Y Acad Sci 2002;957:57-69. Wiley

[14] Baliyan S, Mukherjee R, Priyadarshini A, Vibhuti A, Gupta A, Pandey RP, et al. Determination of antioxidants by DPPH radical scavenging activity and quantitative phytochemical analysis of Ficus religiosa. Molecules 2022;27:1326. MDPI

[15] Blois MS. Antioxidant determinations by the use of a stable free radical. Nature 1958;181:1199-200. Nature

[16] Brand-Williams W, Cuvelier ME, Berset C. Use of a free radical method to evaluate antioxidant activity. LWT Food Sci Technol 1995;28:25-30. ScienceDirect

[17] Munteanu IG, Apetrei C. Analytical methods used in determining antioxidant activity: a review. Int J Mol Sci 2021;22:3380. MDPI

[18] Pérez-Jiménez J, Saura-Calixto F. Effect of solvent and certain food constituents on different antioxidant capacity assays. Food Res Int 2006;39:791-800. ScienceDirect

[19] Xie J, Schaich KM. Re-evaluation of the 2,2-diphenyl-1-picrylhydrazyl free radical (DPPH) assay for antioxidant activity. J Agric Food Chem 2014;62:4251-60. ACS

[20] Zhao D, Zhang Y, Chen Y, Li B, Zhou W, Wang L. Highly accurate and explainable predictions of small-molecule antioxidants for eight in vitro assays simultaneously through an alternating multitask learning strategy. J Chem Inf Model 2024;64:9098-110. ACS

[21] Rice-Evans CA, Miller NJ, Paganga G. Structure-antioxidant activity relationships of flavonoids and phenolic acids. Free Radic Biol Med 1996;20:933-56. ScienceDirect

[22] Moriwaki H, Tian YS, Kawashita N, Takagi T. Mordred: a molecular descriptor calculator. J Cheminform 2018;10:4. Springer

[23] Ghironi S, Viganò EL, Selvestrel G, Benfenati E. QSAR models for predicting the antioxidant potential of chemical substances. J Xenobiot 2025;15:80. MDPI

[24] Kim HC, Ha SY, Yang JK. Modeling radical scavenging activity using molecular descriptors of unclassified compounds. J King Saud Univ Sci 2025;37:5552024. Journal of King Saud University - Science

[25] Goya Jorge E, Rayar AM, Barigye SJ, Jorge Rodríguez ME, Sylla-Iyarreta Veitía M. Development of an in silico model of DPPH free radical scavenging capacity: prediction of antioxidant activity of coumarin type compounds. Int J Mol Sci 2016;17:881. MDPI

[26] Muraro C, Polato M, Bortoli M, Aiolli F, Orian L. Radical scavenging activity of natural antioxidants and drugs: development of a combined machine learning and quantum chemistry protocol. J Chem Phys 2020;153:114117. AIP Publishing

[27] Ayres L, Benavidez T, Varillas A, Linton J, Whitehead DC, Garcia CD. Predicting antioxidant synergism via artificial intelligence and benchtop data. J Agric Food Chem 2023;71:15644-55. ACS

[28] Gilmer J, Schoenholz SS, Riley PF, Vinyals O, Dahl GE. Neural message passing for quantum chemistry. Proc 34th Int Conf Mach Learn, PMLR 2017;70:1263-72. arXiv

[29] Hakkal S, Ait Lahcen A. Leveraging graph neural network for learner performance prediction. Expert Syst Appl 2025;293:128724. ScienceDirect

[30] Cai J, Lou X, Chong CF, Alex D, Arrais JP, Wang Y, et al. Multi-AOP: a lightweight multi-view deep learning framework for antioxidant peptide discovery. Bioresour Bioprocess 2026;13:21. Springer

[31] Deng W, Chen Y, Sun X, Wang L. AODB: a comprehensive database for antioxidants including small molecules, peptides and proteins. Food Chem 2023;418:135992. ScienceDirect

[32] Weininger D. SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules. J Chem Inf Comput Sci 1988;28:31-6. ACS

[33] Heller SR, McNaught A, Pletnev I, Stein S, Tchekhovskoi D. InChI, the IUPAC International Chemical Identifier. J Cheminform 2015;7:23. Springer

[34] Kipf TN, Welling M. Semi-supervised classification with graph convolutional networks. Int Conf Learn Represent (ICLR) 2017. arXiv

[35] Veličković P, Cucurull G, Casanova A, Romero A, Liò P, Bengio Y. Graph attention networks. Int Conf Learn Represent (ICLR) 2018. arXiv

[36] Xu K, Hu W, Leskovec J, Jegelka S. How powerful are graph neural networks? Int Conf Learn Represent (ICLR) 2019. arXiv

[37] Alexander DLJ, Tropsha A, Winkler DA. Beware of R-squared: simple, unambiguous assessment of the prediction accuracy of QSAR and QSPR models. J Chem Inf Model 2015;55:1316-22. ACS

[38] Bemis GW, Murcko MA. The properties of known drugs. 1. Molecular frameworks. J Med Chem 1996;39:2887-93. ACS

[39] OECD. Guidance document on the validation of (quantitative) structure-activity relationship [(Q)SAR] models. OECD Series on Testing and Assessment No. 69. Paris: OECD Publishing; 2007. OECD

[40] Landrum G. RDKit: open-source cheminformatics software; 2024. Available from: RDKit

[41] Iacopini P, Baldi M, Storchi P, Sebastiani L. Catechin, epicatechin, quercetin, rutin and resveratrol in red grape: content, in vitro antioxidant activity and interactions. J Food Compos Anal 2008;21:589-98. ScienceDirect

[42] Lv Q, Luo F, Zhao X, Liu Y, Hu G, Sun C, et al. Identification of proanthocyanidins from litchi (Litchi chinensis Sonn.) pulp by LC-ESI-Q-TOF-MS and their antioxidant activity. PLoS One 2015;10:e0120480. PLOS

[43] Luchian CE, Cotea VV, Vlase L, Toiu AM, Colibaba LC, Raschip IE, et al. Antioxidant and antimicrobial effects of grape pomace extracts. BIO Web Conf 2019;15:04006. BIO Web of Conferences

[44] Sy B, Krisa S, Richard T, Courtois A. Resveratrol, ε-viniferin, and vitisin B from vine: comparison of their in vitro antioxidant activities and study of their interactions. Molecules 2023;28:7521. MDPI

[45] Bioactive depside and anthocyanin compounds, compositions, and methods of use. US Patent 8,846,757 B2; 2014. Available from: Google Patents

[46] Moazzen A, Öztinen N, Ak-Sakalli E, Koşar M. Structure-antiradical activity relationships of 25 natural antioxidant phenolic compounds from different classes. Heliyon 2022;8:e10467. ScienceDirect

[47] Xiong Z, Wang D, Liu X, Zhong F, Wan X, Li X, et al. Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. J Med Chem 2020;63:8749-60. ACS

@article{1723177,
      author = {Aditi Ekhande},
      title = {Graph Neural Network Prediction of Antioxidant Activity of Polyphenols from Grape Pomace},
      journal = {Iconic Research And Engineering Journals},
      year = {2026},
      volume = {10},
      number = {3},
      pages = {2145-2168},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1723177.pdf},
      abstract = {Grape pomace is the main solid by-product of winemaking and a cheap source of polyphenols. However, the antioxidant strength of most compounds in grape pomace has never been measured on its own. Mixtures do not behave additively, so extract results cannot be separated into compound-level values. This paper reviews the chemistry and the modelling problem, sets out a full pipeline from data curation to molecular graphs to a trained network, and then runs it on measured data. Three results concern the data. Unit conversion matters, since epicatechin measured in two laboratories in two-unit systems agrees to 0.021 pIC50 log units once converted. Donor count does not order activity, since gallic acid and trans-resveratrol carry the same three phenolic hydroxyls yet differ by 0.436 log units, which is the argument for a representation that encodes adjacency. And protocols cannot be pooled, since three compounds measured in a second laboratory differ systematically by 1.152 log units, more than the 1.477 log unit range of the training corpus. For training, 25 phenolics measured in one laboratory under one protocol reduced to 16 usable records once non-numeric entries were removed and dose-response fit quality was filtered on, a filter that published curation protocols do not apply and that removed a compound whose tabulated IC50 came from a regression explaining 19 percent of its variance. On those 16 compounds a message-passing network reached a leave-one-out R2 of 0.744 against 0.138 for a mean predictor, with a training fit of 0.696, and a label-scrambling control confirmed that no structure-activity signal was recovered at this sample size. Predictions for the five targets span 0.322 log units and all ten pairwise comparisons overlap at 95 percent. The pipeline runs correctly, and closing the gap between 16 compounds and the 1911 used by published benchmarks is the work that remains.},
      keywords = {Grape pomace, Polyphenols, Antioxidant activity, DPPH},
      month = {September},
  }