Skip to content
DraconiticResearch
Search Sources API
201,683 studies31 sourcesUpdated 26 min ago
Peer-reviewedCC BY-NC 4.0Nucleic Acids Research

CLEAN Database: providing enzyme function coverage for over 55 million sequences

  • Le Yuan1,2,3
  • Jianan Canal Li4
  • Miaoyuan Wen5
  • David M Bianchi3,6,7
  • Katherine Arneson3,6,7
  • Bingji Guo3,6
  • Sara Lambert3,6
  • Chris Stephens3,6
  • Tianhao Yu1,2,3
  • Lijiang Fu3,6
  • Christopher Pond3,6
  • Matthew J Berry3,6,7
1 more authors
  • Huimin Zhao1,2,3,7,8
8 affiliations
  1. 1Department of Chemical and Biomolecular Engineering, University of Illinois Urbana-Champaign , Urbana, IL 61801 ,
  2. 2Carl R. Woese Institute for Genomic Biology, University of Illinois Urbana-Champaign , Urbana, IL 61801 ,
  3. 3NSF Molecule Maker Lab Institute, University of Illinois Urbana-Champaign , Urbana, IL 61801 ,
  4. 4Department of Electrical Engineering and Computer Sciences, University of California , Berkeley, CA 94720 ,
  5. 5School of Information Sciences, University of Illinois Urbana-Champaign , Champaign, IL 61820 ,
  6. 6National Center for Supercomputing Applications, University of Illinois Urbana-Champaign , Urbana, IL 61801 ,
  7. 7NSF iBioFoundry, University of Illinois Urbana-Champaign , Urbana, IL 61801 ,
  8. 8NSF Global Center for Biofoundry Applications, University of Illinois Urbana-Champaign , Urbana, IL 61801 ,
Open at sourcePDFDOICite

Abstract

Enzyme function annotation is critical for decoding biological systems and enabling rational enzyme design. However, experimentally characterized enzyme functions remain available for only a small fraction of known protein sequences, resulting in a widening gap between rapidly expanding sequence space and reliable functional assignments. Recent advances in artificial intelligence (AI) have accelerated enzyme function annotation by facilitating accurate and high-throughput prediction from protein sequences. Here, we present CLEAN Database (CLEAN DB), a comprehensive resource encompassing >55 million enzyme function annotations across 1 098 305 strains, predicted by the CLEAN AI model. The database features an intuitive web interface that enables users to efficiently search, explore, and download enzyme function data through flexible query options. Additionally, powered by a protein language model, CLEAN DB supports mutation effect prediction through interactive heatmaps, structural overlays, and tabular data visualizations. We anticipate that CLEAN DB will serve as a valuable platform for advancing enzyme discovery and applications in synthetic biology, metabolic engineering, enzyme engineering, retrosynthesis, and drug discovery. The database is freely available at https://cleandb.platform.moleculemaker.org/.

Related

  1. THEOBROMA: an aggregated open database of 1.13 million natural products with per-compound license auditing, three-tier classification, and stereochemistry-aware deduplication
    Thor Klamt, Anne Jaczkowski, Jakob Franke +1
    Nucleic Acids ResearchPeer-reviewedCC BY 4.0Full textMolecular biology & gene regulationCrossref
  2. G4All: a database of experimentally confirmed G-quadruplex-forming sequences
    Parth Rajput, Anne Cucchiarini, Dmitrii Trubetskoy +6
    Nucleic Acids ResearchPeer-reviewedCC BY 4.0Full textMolecular biology & gene regulationCrossref
  3. From sequences and structures to fitness, folding and function: challenges in the age of AI
    Ezequiel A. Galpern, Charles W. J. Pugh, Mafalda Dias +1
    ChemRxivPreprintCC BYFull textMolecular biology & gene regulationChemRxiv
  4. OrganiZymeDB: a comprehensive database of enzyme stability and function in the presence of organic solvents
    Romain Debruyne, Fabrizio Pucci
    Nucleic Acids ResearchPeer-reviewedCC BY-NC 4.0Full textMolecular biology & gene regulationCrossref
  5. Calibration, accuracy, and domain variation across one million collective forecasts from prediction platforms
    Justin Dickerson, Hilla Shinitzky, Maayan Peer +2
    Proceedings of the National Academy of Sciences of the United States of AmericaPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationPMC

Cite

BibTeXRIS
BibTeX
@article{yuan2026clean,
  title = {{CLEAN Database: providing enzyme function coverage for over 55 million sequences}},
  author = {Le Yuan and Jianan Canal Li and Miaoyuan Wen and David M Bianchi and Katherine Arneson and Bingji Guo and Sara Lambert and Chris Stephens and Tianhao Yu and Lijiang Fu and Christopher Pond and Matthew J Berry and Huimin Zhao},
  journal = {Nucleic Acids Research},
  year = {2026},
  month = oct,
  doi = {10.1093/nar/gkag1003},
  url = {https://academic.oup.com/nar/advance-article/doi/10.1093/nar/gkag1003/8916836},
  note = {CC BY-NC 4.0},
}
RIS
TY  - JOUR
TI  - CLEAN Database: providing enzyme function coverage for over 55 million sequences
AU  - Le Yuan
AU  - Jianan Canal Li
AU  - Miaoyuan Wen
AU  - David M Bianchi
AU  - Katherine Arneson
AU  - Bingji Guo
AU  - Sara Lambert
AU  - Chris Stephens
AU  - Tianhao Yu
AU  - Lijiang Fu
AU  - Christopher Pond
AU  - Matthew J Berry
AU  - Huimin Zhao
PY  - 2026
DA  - 2026/10/15
JO  - Nucleic Acids Research
DO  - 10.1093/nar/gkag1003
UR  - https://academic.oup.com/nar/advance-article/doi/10.1093/nar/gkag1003/8916836
AB  - Enzyme function annotation is critical for decoding biological systems and enabling rational enzyme design. However, experimentally characterized enzyme functions remain available for only a small fraction of known protein sequences, resulting in a widening gap between rapidly expanding sequence space and reliable functional assignments. Recent advances in artificial intelligence (AI) have accelerated enzyme function annotation by facilitating accurate and high-throughput prediction from protein sequences. Here, we present CLEAN Database (CLEAN DB), a comprehensive resource encompassing >55 million enzyme function annotations across 1 098 305 strains, predicted by the CLEAN AI model. The database features an intuitive web interface that enables users to efficiently search, explore, and download enzyme function data through flexible query options. Additionally, powered by a protein language model, CLEAN DB supports mutation effect prediction through interactive heatmaps, structural overlays, and tabular data visualizations. We anticipate that CLEAN DB will serve as a valuable platform for advancing enzyme discovery and applications in synthetic biology, metabolic engineering, enzyme engineering, retrosynthesis, and drug discovery. The database is freely available at https://cleandb.platform.moleculemaker.org/.
SN  - 0305-1048
SN  - 1362-4962
ER  -