Mapping the Metabolic Profiles of Long-Term Vegetable, Fruit, and Fruit Juice Consumption. (29th May 2020)

• Record Type: Journal Article
• Authors: Wang, Fenglei; Wan, Yi; Zeleznik, Oana; Baden, Megu; Wu, You; Clish, Clary; Willett, Walter; Liang, Liming; Eliassen, Heather
• Is Part Of: Current developments in nutrition. Volume 4(2020)Supplement 2
• Page Start: 787
• Page End: 787
• Publication Date: 2020-05-29
• Subjects: Nutrition -- Periodicals; Nutritional Physiological Phenomena; Nutrition; Periodicals; Periodicals; Fulltext; Internet Resources; Periodicals; 612.3
• DOI: 10.1093/cdn/nzaa052_056
• Languages: English
• ISSNs: 2475-2991

Abstract

Abstract: Objectives: The aim of this study was to explore the metabolic profiles, including both named metabolites and unknown peaks detected in the metabolomics measurement, that are related to long-term vegetable, fruit, and fruit juice consumption. Methods: The study population for exploration included 5270 participants and for replication included 4216 participants in the Nurses' Health Study (NHS), NHSII, and Health Professionals Follow-Up Study. Plasma metabolic profiling was conducted by liquid chromatography-tandem mass spectrometry. Long-term vegetable, fruit, and fruit juice consumption was estimated from food-frequency questionnaires. We included 332 named metabolites and 2561 unknown peaks in the exploration analysis. Partial Spearman correlation analyses were used to assess the associations of vegetable, fruit, and fruit juice consumption with individual metabolites. We further identified metabolic signatures using machine learning models. Results: We observed a panel of named metabolites and unknown peaks that were significantly associated with long-term vegetable, fruit, and fruit juice consumption (P value < 0.05 after adjusting for multiple testing). Several unknown peaks exhibited a comparable correlation (partial Spearman rho > 0.4) with fruit juice, especially orange juice consumption, relative to the named metabolites. Metabolic signatures, comprised of 78, 104, and 41 named metabolites, were robustly correlated with total vegetable, total fruit, and total fruit juice consumption, respectively (Pearson r = 0.27–0.37 between the signature and dietary consumption in exploration samples, and 0.24–0.27 in replication samples). Adding unknown peaks into the metabolic signature strengthened the Pearson r by 17.7% for total vegetable, 9.5% for total fruit, and 5.3% for total fruit juice consumption. Conclusions: Using plasma metabolomics platform, we identified metabolic profiles, including named metabolites and unknown peaks, that reflect long-term vegetable, fruit, and fruit juice consumption, respectively. Funding Sources: None.