Supervised machine learning for nutrient and contaminant profiling and benefit-risk assessment of Arctic seafood
Environment International · 2025
DOI: 10.1016/j.envint.2025.109699 ↗Kurz gefasst
Fachartikel in der Zeitschrift Environment International, 2025. Verfasst von Quang Tri Ho und 8 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Marine Research, Bergen, Norway. Electronic address: Quang.Tri.Ho@hi.no. Hauptthemen laut MEDLINE: Seafood, Machine Learning, Water Pollutants, Chemical, Nutrients. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Havforskningsinstituttet, Fiskeri - og havbruksnæringens forskningsfond, Mattilsynet.
Abstract (Originalsprache)
Arctic seafood is a crucial source of nutrients due to its richness in proteins, omega-3 polyunsaturated fatty acids, vitamins and minerals. However, contaminants including Hg and persistent organic pollutants (POPs) in fish may pose a health concern for consumers. This study aimed to profile concentration patterns including interspecies and regional variations between Barents Sea and Norwegian Sea, and to explore correlations of nutrients and contaminants in five commercially important Arctic marine fish species using supervised machine learning applied to analytical data. These multivariate patterns were subsequently used as input in Monte Carlo simulations to assess the benefit and risk of fish consumption. Except for Atlantic mackerel, fish from the Barents Sea generally had lower concentrations of Hg and POPs than those from the Norwegian Sea, while nutrient concentrations were largely comparable across both regions. Fatty fish, especially Atlantic mackerel and herring had higher nutrient scores than lean fish. Consuming lean fish from 300 to 450 g per week did not entail risk of excessive Hg, dioxins and DL-PCBs exposure. Although weekly consumption of more than 190 g of mackerel or 130 g of herring from the Norwegian Sea, may exceed the tolerable weekly intake for dioxins and DL-PCBs (P > 0.05), both species are excellent sources of EPA + DHA, Se and vitamin D. The nutritional benefits may outweigh the potential risk of dioxins and DL-PCBs exposure. Arctic fish rich in many essential nutrients are important for food security.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Institute of Marine Research, Bergen, Norway. Electronic address: Quang.Tri.Ho@hi.no.
- Institute of Marine Research, Bergen, Norway.
Förderung
- Havforskningsinstituttet
- Fiskeri - og havbruksnæringens forskningsfond
- Mattilsynet
- Nærings- og Fiskeridepartementet
Zitieren
APA
Ho, Q. T., Nøstbakken, O. J., Sanden, M., Myrmel, L. S., Wiech, M., Bernhard, A., Nilsen, B. M., Maage, A., & Dahl, L. (2025). Supervised machine learning for nutrient and contaminant profiling and benefit-risk assessment of Arctic seafood. Environment International. https://doi.org/10.1016/j.envint.2025.109699
Vancouver
Ho QT, Nøstbakken OJ, Sanden M, Myrmel LS, Wiech M, Bernhard A, et al. Supervised machine learning for nutrient and contaminant profiling and benefit-risk assessment of Arctic seafood. Environment International. 2025. doi:10.1016/j.envint.2025.109699
BibTeX
@article{Ho2025,
author = {Ho, Quang Tri and Nøstbakken, Ole Jakob and Sanden, Monica and Myrmel, Lene Secher and Wiech, Martin and Bernhard, Annette and Nilsen, Bente M and Maage, Amund and Dahl, Lisbeth},
title = {Supervised machine learning for nutrient and contaminant profiling and benefit-risk assessment of Arctic seafood},
journal = {Environment International},
year = {2025},
doi = {10.1016/j.envint.2025.109699}
}
EUVIMED