EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Transductive Learning From a Transformer Foundation Model Improves the Taxonomic Classification of eDNA Metabarcodes

Simon Bettinger iD, Auguste Verdier iD, Morgane Bruno, Stéphanie Manel iD, Alexis Joly iD, Maximilien Servajean iD, Alice Valentini iD, Rodolphe Devillers iD, David Mouillot iD

Molecular Ecology Resources · 2026-11 · 26 (8)

DOI: 10.1111/1755-0998.70204 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Molecular Ecology Resources, online erschienen am 11.11.2026. Verfasst von Simon Bettinger und 8 weiteren Autor:innen. Beteiligte Einrichtung: MARBEC, Univ Montpellier, CNRS, IFREMER, IRD, Montpellier, France. Hauptthemen laut MEDLINE: DNA, Environmental, Fishes, DNA Barcoding, Taxonomic, Deep Learning. Die Arbeit nennt 85 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: PPR Océan & Climat programme, BPI project SEAMAP, the IA-Biodiv Agence Nationale de la Recherche (ANR), ANR Chaire Industrielle ADNeIA.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Environmental DNA (eDNA) has emerged as a powerful and cost‐effective biodiversity monitoring tool, providing comprehensive and non‐invasive taxonomic inventories. However, eDNA‐based monitoring is strongly dependent on the completeness of genetic reference databases, which varies greatly among regions and taxa. Since completing gaps at the species level is unrealistic in the short term, an alternative is to infer coarser taxonomic assignments for sequences belonging to species not yet in reference databases. To address this challenge, we introduce BouillaBERT , a novel deep learning method tailored for the operational classification of eDNA sequences from unknown genera. BouillaBERT leverages data augmentation, transfer and transductive learning to overcome the small size and imbalance of eDNA reference databases for fishes. We implement and train the model across four commonly used fish eDNA markers and compare its performance with lowest common ancestor (LCA) alignment methods and existing deep learning approaches. BouillaBERT outperforms all tested LCA alignment and deep learning methods in recall for the assignment of unknown sequences across a range of commonly used fish eDNA primers, while achieving comparable or higher precision. BouillaBERT can be rapidly re‐trained to fit a custom sample set or used as‐is, enabling more consistent biodiversity assessments across undersampled regions or taxa.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

DNA, EnvironmentalFishesDNA Barcoding, TaxonomicDeep LearningAnimalsBiodiversityExtrachromosomal DNA

Stichwörter der Autor:innen

DNABERTdeep learningenvironmental DNAfish biodiversitymetabarcodingtransformer models

Substanzen

DNA, EnvironmentalExtrachromosomal DNA

Autor:innen und Institutionen

  • MARBEC, Univ Montpellier, CNRS, IFREMER, IRD, Montpellier, France.
  • CEFE, Univ Montpellier, CNRS, EPHE- PSL University, IRD, Montpellier, France.
  • Institut Universitaire de France, Paris, France.
  • Inria, Univ Montpellier, LIRMM, CNRS, Montpellier, France.
  • SPYGEN, Le Bourget-du-Lac, France.
  • Espace-Dev, IRD, Univ Montpellier, Univ Réunion, Montpellier, France.

ORCID-Profile: Simon Bettinger, Auguste Verdier, Stéphanie Manel, Alexis Joly, Maximilien Servajean, Alice Valentini, Rodolphe Devillers, David Mouillot

Förderung

  • PPR Océan & Climat programme
  • BPI project SEAMAP, the IA-Biodiv Agence Nationale de la Recherche (ANR)
  • ANR Chaire Industrielle ADNeIA

Zitieren

APA

Bettinger, S., Verdier, A., Bruno, M., Manel, S., Joly, A., Servajean, M., Valentini, A., Devillers, R., & Mouillot, D. (2026). Transductive Learning From a Transformer Foundation Model Improves the Taxonomic Classification of eDNA Metabarcodes. Molecular Ecology Resources, 26(8). https://doi.org/10.1111/1755-0998.70204

Vancouver

Bettinger S, Verdier A, Bruno M, Manel S, Joly A, Servajean M, et al. Transductive Learning From a Transformer Foundation Model Improves the Taxonomic Classification of eDNA Metabarcodes. Molecular Ecology Resources. 2026;26(8). doi:10.1111/1755-0998.70204

BibTeX
@article{Bettinger2026,
  author = {Bettinger, Simon and Verdier, Auguste and Bruno, Morgane and Manel, Stéphanie and Joly, Alexis and Servajean, Maximilien and Valentini, Alice and Devillers, Rodolphe and Mouillot, David},
  title = {Transductive Learning From a Transformer Foundation Model Improves the Taxonomic Classification of eDNA Metabarcodes},
  journal = {Molecular Ecology Resources},
  year = {2026},
  volume = {26},
  number = {8},
  doi = {10.1111/1755-0998.70204}
}
Kostenlos starten →