Generalized biological foundation model with unified nucleic acid and protein language

Yong He; Pan Fang; Yongtao Shan; Yuanfei Pan; Yanhong Wei; Yichang Chen; Yihao Chen; Yi Liu; Zhenyu Zeng; Zhan Zhou; Feng Zhu; Edward C. Holmes; Jieping Ye; Jun Li; Yuelong Shu; Mang Shi; Zhaorong Li

doi:10.1038/s42256-025-01044-4

Generalized biological foundation model with unified nucleic acid and protein language

Yong He^*, Pan Fang, Yongtao Shan, Yuanfei Pan, Yanhong Wei, Yichang Chen, Yihao Chen, Yi Liu, Zhenyu Zeng, Zhan Zhou, Feng Zhu, Edward C. Holmes, Jieping Ye, Jun Li, Yuelong Shu, Mang Shi^*, Zhaorong Li^*

^*Corresponding author for this work

Department of Infectious Diseases and Public Health

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

1 Downloads (CityUHK Scholars)

Abstract

The language of biology, encoded in DNA, RNA and proteins, forms the foundation of life but remains challenging to decode owing to its complexity. Traditional computational methods often struggle to integrate information across these molecules, limiting a comprehensive understanding of biological systems. Advances in natural language processing with pre-trained models offer possibilities for interpreting biological language. Here we introduce LucaOne, a pre-trained foundation model trained on nucleic acid and protein sequences from 169,861 species. Through large-scale data integration and semi-supervised learning, LucaOne shows an understanding of key biological principles, such as DNA–protein translation. Using few-shot learning, it effectively comprehends the central dogma of molecular biology and performs competitively on tasks involving DNA, RNA or protein inputs. Our results highlight the potential of unified foundation models to address complex biological questions, providing an adaptable framework for bioinformatics research and enhancing the interpretation of life’s complexity. © The Author(s) 2025

Original language	English
Pages (from-to)	942-953
Journal	Nature Machine Intelligence
Volume	7
Issue number	6
DOIs	https://doi.org/10.1038/s42256-025-01044-4
Publication status	Published - 18 Jun 2025

Funding

This work was supported by the National Natural Science Foundation of China (82341118). M.S. is funded by the Shenzhen Science and Technology Program (KQTD20200820145822023), the Guangdong Province ‘Pearl River Talent Plan’ Innovation and Entrepreneurship Team project (2019ZT08Y464), and the Guangzhou National Laboratory Major Project (GZNL2023A01001). Y.P. is funded by the National Natural Science Foundation of China (NSFC) Basic Research Project for Doctoral Students (grant number 323B2018).

Publisher's Copyright Statement

This full text is made available under CC-BY 4.0. https://creativecommons.org/licenses/by/4.0/

Access Output

10.1038/s42256-025-01044-4

291252429.pdfFinal Published version, 3.2 MBLicence: CC BY

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{b936ec0f3a484be2adf9c0cdb844ac8c,

title = "Generalized biological foundation model with unified nucleic acid and protein language",

abstract = "The language of biology, encoded in DNA, RNA and proteins, forms the foundation of life but remains challenging to decode owing to its complexity. Traditional computational methods often struggle to integrate information across these molecules, limiting a comprehensive understanding of biological systems. Advances in natural language processing with pre-trained models offer possibilities for interpreting biological language. Here we introduce LucaOne, a pre-trained foundation model trained on nucleic acid and protein sequences from 169,861 species. Through large-scale data integration and semi-supervised learning, LucaOne shows an understanding of key biological principles, such as DNA–protein translation. Using few-shot learning, it effectively comprehends the central dogma of molecular biology and performs competitively on tasks involving DNA, RNA or protein inputs. Our results highlight the potential of unified foundation models to address complex biological questions, providing an adaptable framework for bioinformatics research and enhancing the interpretation of life{\textquoteright}s complexity. {\textcopyright} The Author(s) 2025",

author = "Yong He and Pan Fang and Yongtao Shan and Yuanfei Pan and Yanhong Wei and Yichang Chen and Yihao Chen and Yi Liu and Zhenyu Zeng and Zhan Zhou and Feng Zhu and Holmes, {Edward C.} and Jieping Ye and Jun Li and Yuelong Shu and Mang Shi and Zhaorong Li",

year = "2025",

month = jun,

day = "18",

doi = "10.1038/s42256-025-01044-4",

language = "English",

volume = "7",

pages = "942--953",

journal = "Nature Machine Intelligence",

issn = "2522-5839",

publisher = "Springer Nature Switzerland AG",

number = "6",

}

TY - JOUR

T1 - Generalized biological foundation model with unified nucleic acid and protein language

AU - He, Yong

AU - Fang, Pan

AU - Shan, Yongtao

AU - Pan, Yuanfei

AU - Wei, Yanhong

AU - Chen, Yichang

AU - Chen, Yihao

AU - Liu, Yi

AU - Zeng, Zhenyu

AU - Zhou, Zhan

AU - Zhu, Feng

AU - Holmes, Edward C.

AU - Ye, Jieping

AU - Li, Jun

AU - Shu, Yuelong

AU - Shi, Mang

AU - Li, Zhaorong

PY - 2025/6/18

Y1 - 2025/6/18

N2 - The language of biology, encoded in DNA, RNA and proteins, forms the foundation of life but remains challenging to decode owing to its complexity. Traditional computational methods often struggle to integrate information across these molecules, limiting a comprehensive understanding of biological systems. Advances in natural language processing with pre-trained models offer possibilities for interpreting biological language. Here we introduce LucaOne, a pre-trained foundation model trained on nucleic acid and protein sequences from 169,861 species. Through large-scale data integration and semi-supervised learning, LucaOne shows an understanding of key biological principles, such as DNA–protein translation. Using few-shot learning, it effectively comprehends the central dogma of molecular biology and performs competitively on tasks involving DNA, RNA or protein inputs. Our results highlight the potential of unified foundation models to address complex biological questions, providing an adaptable framework for bioinformatics research and enhancing the interpretation of life’s complexity. © The Author(s) 2025

AB - The language of biology, encoded in DNA, RNA and proteins, forms the foundation of life but remains challenging to decode owing to its complexity. Traditional computational methods often struggle to integrate information across these molecules, limiting a comprehensive understanding of biological systems. Advances in natural language processing with pre-trained models offer possibilities for interpreting biological language. Here we introduce LucaOne, a pre-trained foundation model trained on nucleic acid and protein sequences from 169,861 species. Through large-scale data integration and semi-supervised learning, LucaOne shows an understanding of key biological principles, such as DNA–protein translation. Using few-shot learning, it effectively comprehends the central dogma of molecular biology and performs competitively on tasks involving DNA, RNA or protein inputs. Our results highlight the potential of unified foundation models to address complex biological questions, providing an adaptable framework for bioinformatics research and enhancing the interpretation of life’s complexity. © The Author(s) 2025

UR - http://www.scopus.com/inward/record.url?scp=105008314635&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-105008314635&origin=recordpage

U2 - 10.1038/s42256-025-01044-4

DO - 10.1038/s42256-025-01044-4

M3 - RGC 21 - Publication in refereed journal

SN - 2522-5839

VL - 7

SP - 942

EP - 953

JO - Nature Machine Intelligence

JF - Nature Machine Intelligence

IS - 6

ER -

Generalized biological foundation model with unified nucleic acid and protein language

Abstract

Funding

Publisher's Copyright Statement

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this