Multi-omics and explainable machine learning identify key enzymes associated with flavonoid biosynthesis in grape seedsDownload PDF Download PDF ArticleOpen accessPublished: 17 November 2026Jianxiang Liang ORCID: orcid.org/0000-0001-7628-79071,Yaru Luo ORCID: orcid.org/0009-0006-7437-04441,Lirui Liu2,Chuni Zhou1,Junnan Ding1,Shiren Song ORCID: orcid.org/0000-0002-1343-67811 &…Bohan Yang1 npj Science of Food (2026) Cite this articleView saved research We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.AbstractGrape seeds are abundant byproducts of grape juice and wine processing and contain diverse polyphenols associated with antioxidant potential, matrix interactions, and the quality attributes of grape-derived products. Although wild East Asian grape species represent valuable reservoirs of polyphenol diversity, the phytochemical potential of their seeds and the molecular mechanisms linking genetic variation to divergent polyphenol profiles remain poorly understood. Here, we integrated untargeted and widely targeted metabolomics, graph pangenome-guided transcriptomics, quantitative proteomics, explainable machine learning, and comparative genomics to characterize seed polyphenol variation during ripening in Vitis vinifera ‘Cabernet Sauvignon’ and Vitis heyneana ‘Maoputao’. Multi-omics profiling revealed that the V. heyneana seeds exhibited a polyphenol profile enriched in flavonoids, hydrolysable tannins, methylated flavonoids, and galloylated derivatives. An explainable machine learning framework with SHAP attribution prioritized 10 polyphenols that contributed most strongly to discrimination between the two grapevines. Transcriptomic differences were concentrated in flavonoid and isoflavonoid biosynthesis, whereas proteomic differences were enriched in glycolysis, the pentose phosphate pathway and pyruvate metabolism, revealing complementary pathway-level signatures. Haplotype-resolved analysis across 12 V. heyneana haplotypes identified extensive CYP copy-number variation and expansion driven mainly by tandem and proximal duplication, particularly within Clan 71. Expression-based ranking of 368 VvCYPs prioritized 6 seed-associated candidates, among which VvCYP82D13 showed marked developmental expression and conserved microsynteny. Apigenin was selected as a pathway-informed candidate ligand, and molecular docking and dynamics simulations supported its computationally plausible binding within the predicted VvCYP82D13 pocket. These findings provide comprehensive information on the grapevine CYP superfamily and its association with developmental polyphenol variation in seeds, facilitating further functional characterization of the prioritized CYP candidates.AcknowledgementsThis research was supported by the Shanghai Agricultural Science and Technology Innovation Program (Grant Nos. K2023017 and T2023208) and the Cyrus Tang Foundation Zhixing Research Program. The funders provided financial support for this study but were not involved in the study design, data collection and analysis, interpretation of the results, preparation of the manuscript, or the decision to submit the manuscript for publication. The schematic diagram of the graphical abstract was drawn by Figdraw and bioRender.FundingThe authors received no specific funding for this work.Author informationAuthors and AffiliationsSchool of Agriculture and Biology, Shanghai Jiao Tong University, Shanghai, ChinaJianxiang Liang, Yaru Luo, Chuni Zhou, Junnan Ding, Shiren Song & Bohan YangCollege of Food Science, Hebei Normal University of Science & Technology, Changli, ChinaLirui LiuAuthorsJianxiang LiangView author publicationsSearch author on:PubMed Google ScholarYaru LuoView author publicationsSearch author on:PubMed Google ScholarLirui LiuView author publicationsSearch author on:PubMed Google ScholarChuni ZhouView author publicationsSearch author on:PubMed Google ScholarJunnan DingView author publicationsSearch author on:PubMed Google ScholarShiren SongView author publicationsSearch author on:PubMed Google ScholarBohan YangView author publicationsSearch author on:PubMed Google ScholarCorresponding authorCorrespondence to Bohan Yang.Ethics declarationsCompeting interestsThe authors declare no competing interests.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationSupplementary Figures. (download PDF )Supplementary Tables. (download XLSX )Rights and permissionsOpen Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.Reprints and permissionsAbout this article