Integrating Multi‐Group Microbial eDNA Data With Machine Learning for Large‐Scale Coastal Ecological Quality Assessment
Environmental DNA · 2026-11 · 8 (6)
DOI: 10.1002/edn3.70372 ↗Kurz gefasst
Fachartikel in der Zeitschrift Environmental DNA, online erschienen am 07.11.2026. Verfasst von Xin Gao und 7 weiteren Autor:innen. Die Arbeit nennt 90 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China.
Abstract (Originalsprache)
Among the many statistical methods for elucidating the complex relationships between microbial communities and specific environmental conditions, machine learning‐based methods are considered the most promising. The application of machine learning in the field of ecological quality assessment has been well documented; however, there is a paucity of research in the marine domain at a multi‐region scale. Here, a multi‐region investigation was conducted to assess the ecological quality of coastal area along China's coastline. Among the calculated abiotic indices, the water nutrient quality index and the sediment Nemerow pollution index had the strongest ability to discriminate the ecological status among the monitored regions. Data for different microbial groups (bacteria, fungi, and protists) of the collected samples were obtained by environmental DNA (eDNA) metabarcoding sequencing, and a total of nine core taxa for China's coasts were recognized. Using characteristics (alpha diversity, beta distance, and core taxa) of multi‐group microbial communities as input variables, machine learning models based on nine different algorithms were constructed to predict the ecological quality of sampling sites. Random forest models performed best in predicting ecological quality as classified by ecological indices as well as their actual values. Furthermore, the random forest models based on features from multi‐group microbial communities showed higher predictive accuracy for ecological index values than those using a single microbial group. Our study first explored the applicability of an eDNA‐based machine learning approach to replace the abiotic indices based on physicochemical parameters to assess the ecological quality of coastal ecosystems. Moreover, the findings of this study corroborate the value of machine learning and underscore its prospective applications in coastal ecological assessment.
Förderung
- National Key Research and Development Program of China
Zitieren
APA
Gao, X., Li, H., Sun, Y., Zhang, Y., Peng, M., Huang, B., Lan, W., & Yang, J. (2026). Integrating Multi‐Group Microbial eDNA Data With Machine Learning for Large‐Scale Coastal Ecological Quality Assessment. Environmental DNA, 8(6). https://doi.org/10.1002/edn3.70372
Vancouver
Gao X, Li H, Sun Y, Zhang Y, Peng M, Huang B, et al. Integrating Multi‐Group Microbial eDNA Data With Machine Learning for Large‐Scale Coastal Ecological Quality Assessment. Environmental DNA. 2026;8(6). doi:10.1002/edn3.70372
BibTeX
@article{Gao2026,
author = {Gao, Xin and Li, Hongjun and Sun, Yi and Zhang, Yunlei and Peng, Mo and Huang, Bei and Lan, Wenlu and Yang, Jun},
title = {Integrating Multi‐Group Microbial eDNA Data With Machine Learning for Large‐Scale Coastal Ecological Quality Assessment},
journal = {Environmental DNA},
year = {2026},
volume = {8},
number = {6},
doi = {10.1002/edn3.70372}
}
EUVIMED