EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

FGeneBERT: function-driven pre-trained gene language model for metagenomics

Chenrui Duan iD, Zelin Zang, Yongjie Xu iD, Hang He, Siyuan Li, Zihan Liu iD, Zhen Lei, Ju-Sheng Zheng, Stan Z Li iD

Briefings in Bioinformatics · 01.11.2025 · 26 (6)

DOI: 10.1093/bib/bbaf592 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 10.11.2025. Verfasst von Chenrui Duan und 8 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer Science and Technology, Zhejiang University, No. 866, Yuhangtang Road, 310058 Zhejiang, P. R. China. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 75 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: National Science and Technology Major Project, National Natural Science Foundation of China, Center of Synthetic Biology and Integrated Bioengineering of Westlake University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Metagenomic data, comprising mixed multi-species genomes, are prevalent in diverse environments like oceans and soils, significantly impacting human health and ecological functions. However, current research relies on K-mer, which limits the capture of structurally and functionally relevant gene contexts. Moreover, these approaches struggle with encoding biologically meaningful genes and fail to address the one-to-many and many-to-one relationships inherent in metagenomic data. To overcome these challenges, we introduce FGeneBERT, a novel metagenomic pre-trained model that employs a protein-based gene representation as a context-aware and structure-relevant tokenizer. FGeneBERT incorporates masked gene modeling to enhance the understanding of inter-gene contextual relationships and triplet enhanced metagenomic contrastive learning to elucidate gene sequence–function relationships. Pre-trained on over 100 million metagenomic sequences, FGeneBERT demonstrates superior performance on metagenomic datasets at four levels, spanning gene, functional, bacterial, and environmental levels and ranging from 1 to 213 k input sequences. Case studies of ATP synthase and gene operons highlight FGeneBERT’s capability for functional recognition and its biological relevance in metagenomic research.

Schlagwörter

Stichwörter der Autor:innen

DNAmetagenomicspre-trained language modeltransformer

Autor:innen und Institutionen

  • College of Computer Science and Technology, Zhejiang University, No. 866, Yuhangtang Road, 310058 Zhejiang, P. R. China.
  • School of Engineering, Westlake University, No. 600 Dunyu Road, 310030 Zhejiang, P. R. China.
  • Centre for Artificial Intelligence and Robotics (CAIR), HKISI-CAS Hong Kong Institute of Science & Innovation, Chinese Academy of Sciences, Hong Kong 310000, China.
  • School of Medicine and School of Life Sciences, Westlake University, No. 600 Dunyu Road, 310030 Zhejiang, P. R. China.
  • State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing 100190, China.
  • School of Artificial Intelligence, University of Chinese Academy of Sciences (UCAS), Beijing 100049, China.

ORCID-Profile: Chenrui Duan, Yongjie Xu, Zihan Liu, Stan Z Li

Förderung

  • National Science and Technology Major Project
  • National Natural Science Foundation of China
  • Center of Synthetic Biology and Integrated Bioengineering of Westlake University
  • InnoHK program and Ant Group through CAAI-Ant Research Fund

Zitieren

APA

Duan, C., Zang, Z., Xu, Y., He, H., Li, S., Liu, Z., Lei, Z., Zheng, J. S., & Li, S. Z. (2025). FGeneBERT: function-driven pre-trained gene language model for metagenomics. Briefings in Bioinformatics, 26(6). https://doi.org/10.1093/bib/bbaf592

Vancouver

Duan C, Zang Z, Xu Y, He H, Li S, Liu Z, et al. FGeneBERT: function-driven pre-trained gene language model for metagenomics. Briefings in Bioinformatics. 2025;26(6). doi:10.1093/bib/bbaf592

BibTeX
@article{Duan2025,
  author = {Duan, Chenrui and Zang, Zelin and Xu, Yongjie and He, Hang and Li, Siyuan and Liu, Zihan and Lei, Zhen and Zheng, Ju-Sheng and Li, Stan Z},
  title = {FGeneBERT: function-driven pre-trained gene language model for metagenomics},
  journal = {Briefings in Bioinformatics},
  year = {2025},
  volume = {26},
  number = {6},
  doi = {10.1093/bib/bbaf592}
}
Kostenlos starten →