CoChat
Sign Up
PreprintCited by 7Genomics and Phylogenetic StudiesMachine Learning in BioinformaticsGut microbiota and health

The DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.

Protein Language Models Uncover Carbohydrate-Active Enzyme Function in Metagenomics

Open access
Thurimella, Kumar, Mohamed, Ahmed M. T., Graham, Daniel B., Owens, Róisín M., La Rosa, Sabina Leanti, Plichta, Damian R., Bacallado, Sergio, Xavier, Ramnik J.·2023·bioRxiv (Cold Spring Harbor Laboratory)
DOI 10.1101/2023.10.23.563620PMC PMC10634757

Preprint — published as 10.1186/s12859-025-06286-y · Paper page

Open this paper in CoChat — free accountOpen PDF

Abstract

In metagenomics, the pool of uncharacterized microbial enzymes presents a challenge for functional annotation. Among these, carbohydrate-active enzymes (CAZymes) stand out due to their pivotal roles in various biological processes related to host health and nutrition. Here, we present CAZyLingua, the first tool that harnesses protein language model embeddings to build a deep learning framework that facilitates the annotation of CAZymes in metagenomic datasets. Our benchmarking results showed on average a higher F1 score (reflecting an average of precision and recall) on the annotated genomes of Bacteroides thetaiotaomicron , Eggerthella lenta and Ruminococcus gnavus compared to the traditional sequence homology-based method in dbCAN2. We applied our tool to a paired mother/infant longitudinal dataset and revealed unannotated CAZymes linked to microbial development during infancy. When applied to metagenomic datasets derived from patients affected by fibrosis-prone diseases such as Crohn’s disease and IgG4-related disease, CAZyLingua uncovered CAZymes associated with disease and healthy states. In each of these metagenomic catalogs, CAZyLingua discovered new annotations that were previously overlooked by traditional sequence homology tools. Overall, the deep learning model CAZyLingua can be applied in combination with existing tools to unravel intricate CAZyme evolutionary profiles and patterns, contributing to a more comprehensive understanding of microbial metabolic dynamics.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/2023.10.23.563620

Cites

9 of 9 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 68

  1. 10.1038/nature08821 · DOI · Paper page
  2. 2021, A unified catalog of 204,938 reference genomes from the human gut microbiome, Nat. Biotechnol, 39, 105, 10.1038/s41587-020-0603-3 · DOI · Paper page
  3. 2019, Extensive Unexplored Human Microbiome Diversity Revealed by Over 150,000 Genomes from Metagenomes Spanning Age, Geography, and Lifestyle, Cell, 176, 649, 10.1016/j.cell.2019.01.001 · DOI · Paper page
  4. 10.1038/nature12352 · DOI · Paper page
  5. 10.1073/pnas.1508380112 · DOI
  6. 2022, Discovery of bioactive microbial gene products in inflammatory bowel disease, Nature, 606, 754, 10.1038/s41586-022-04648-7 · DOI
  7. 10.1038/nrg3182 · DOI · Paper page
  8. 2014, Identification of host-microbe interaction factors in the genomes of soft rot-associated pathogens Dickeya dadantii 3937 and Pectobacterium carotovorum WPP14 with supervised machine learning, BMC Genomics, 15, 508, 10.1186/1471-2164-15-508 · DOI
  9. Lozupone, C. A . Unraveling Interactions between the Microbiome and the Host Immune System To Decipher Mechanisms of Disease. mSystems 3, (2018). · DOI
  10. 10.1186/s40168-020-00875-0 · DOI · Paper page
  11. 10.1126/sci-ence.adf2465 · DOI
  12. 10.1038/s41586-023-05696-3 · DOI
  13. 2020, The Classification of Enzymes by Deep Learning, IEEE Access, 8, 89802, 10.1109/ACCESS.2020.2992468 · DOI
  14. 10.1093/bioinformatics/btx680 · DOI
  15. 10.1371/journal.pone.0028742 · DOI · Paper page
  16. 10.1093/nar/gkt1178 · DOI
  17. 10.1093/nar/gkn663 · DOI
  18. 10.1016/j.cell.2016.01.024 · DOI · Paper page
  19. 2022, Inflammatory Bowel Disease-Associated Gut Commensals Degrade Components of the Extracellular Matrix, MBio, 13, e0220122, 10.1128/mbio.02201-22 · DOI
  20. 10.1186/s13073-021-00853-7 · DOI
  21. 2004, Protein homology detection by HMM–HMM comparison, Bioinformatics, 21, 951
  22. 10.1093/nar/gkr367 · DOI · Paper page
  23. 10.1006/jmbi.1990.9999 · DOI · Paper page
  24. 10.1093/nar/25.17.3389 · DOI · Paper page
  25. 10.1093/nar/gky418 · DOI
  26. 10.1038/s41586-018-0124-0 · DOI · Paper page
  27. 10.1101/2022.11.24.517862 · DOI
  28. Madani, A. et al. Large language models generate functional protein sequences across diverse families. Nat. Biotechnol. 1–8 (2023). · DOI · Paper page
  29. 2023, Sequence-structure-function relationships in the microbial protein universe, Nat. Commun, 14, 2351, 10.1038/s41467-023-37896-w · DOI
  30. Rives, A. et al. Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. Proc. Natl. Acad. Sci. U. S. A. 118, (2021). · DOI · Paper page
  31. 2022, Contrastive learning on protein embeddings enlightens midnight zone, NAR Genom Bioinform, 4, lqac043, 10.1093/nargab/lqac043 · DOI
  32. 10.1126/science.ade2574 · DOI · Paper page
  33. Elnaggar, A. et al. ProtTrans: Towards Cracking the Language of Lifes Code Through Self-Supervised Deep Learning and High Performance Computing. IEEE Trans. Pattern Anal. Mach. Intell. PP, (2021).
  34. 2022, Mobile genetic elements from the maternal microbiome shape infant gut microbial assembly and metabolism, Cell, 185, 4921, 10.1016/j.cell.2022.11.023 · DOI
  35. 2021, Infant gut strain persistence is associated with maternal origin, phylogeny, and traits including surface adhesion and iron acquisition, Cell Rep Med, 2, 100393, 10.1016/j.xcrm.2021.100393 · DOI
  36. 2011, Scikit-learn: Machine learning in Python, the Journal of machine Learning research, 12, 2825
  37. 10.1093/nar/gkx894 · DOI
  38. 2022, Genetic manipulation of the human gut bacterium Eggerthella lenta reveals a widespread family of transcriptional regulators, Nat. Commun, 13, 7624, 10.1038/s41467-022-33576-3 · DOI
  39. 10.1073/pnas.1904099116 · DOI
  40. 10.1016/j.celrep.2021.108789 · DOI
  41. Liaw, R. et al. Tune: A Research Platform for Distributed Model Selection and Training. arXiv [cs.LG] (2018).
  42. Paszke, A. , et al. PyTorch: An Imperative Style, High-Performance Deep Learning Library. arXiv [cs.LG] (2019).
  43. 10.1038/S41592-022-01488-1 · DOI · Paper page
  44. 10.1038/s41587-023-01773-0 · DOI · Paper page
  45. 2022, US-align: universal structure alignments of proteins, nucleic acids, and macromolecular complexes, Nat. Methods, 19, 1109, 10.1038/s41592-022-01585-1 · DOI · Paper page
  46. 2022, SignalP 6.0 predicts all five types of signal peptides using protein language models, Nat. Biotechnol, 40, 1023, 10.1038/s41587-021-01156-3 · DOI
  47. 10.1093/nar/gkac387 · DOI
  48. 2019, Gut microbiome structure and metabolic activity in inflammatory bowel disease, Nat Microbiol, 4, 293
  49. 10.25080/majora-92bf1922-011 · DOI
  50. 10.1371/journal.pcbi.1009442 · DOI · Paper page
  51. 2017, Structural diversity of carbohydrate esterases, Biotechnology Research and Innovation, 1, 35, 10.1016/j.biori.2017.02.001 · DOI
  52. 2022, The SGNH hydrolase family: a template for carbohydrate diversity, Glycobiology, 32, 826
  53. 10.1016/S0969-2126(01)00684-0 · DOI
  54. 10.1101/2023.08.08.552477 · DOI
  55. 10.1136/gut.2005.073817 · DOI · Paper page
  56. 10.1016/j.chom.2014.02.005 · DOI · Paper page
  57. 10.1038/s41586-019-1237-9 · DOI · Paper page
  58. 10.1093/bioinformatics/bts565 · DOI · Paper page
  59. 10.1093/nar/26.1.320 · DOI
  60. 10.1093/nar/gkv1070 · DOI · Paper page
  61. 10.1038/s41592-021-01101-x · DOI · Paper page
  62. Abadi, M. , et al. TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems. arXiv [cs.DC] (2016).
  63. 10.1089/cmb.2017.0096 · DOI
  64. 10.1093/bioinformatics/btv033 · DOI
  65. 10.1186/1471-2105-11-119 · DOI
  66. 10.1038/nbt.3988 · DOI · Paper page
  67. 2007, Hunter. Matplotlib: A 2D Graphics Environment, 9, 90
  68. 10.1038/s41592-019-0686-2 · DOI · Paper page

Metadata

Authors
Thurimella, K., Mohamed, A. M. T., Graham, D. B., Owens, R. M., La Rosa, S. L., Plichta, D. R., Bacallado, S., & Xavier, R. J.
Affiliations
  • Broad Institute (US)
  • University of Cambridge (GB)
Corresponding author
  • Plichta, Damian R.
  • Bacallado, Sergio
  • Xavier, Ramnik J.
ORCID
Venue
bioRxiv (Cold Spring Harbor Laboratory)
Publisher
openRxiv
Published
2023-10-25
DOI
10.1101/2023.10.23.563620
PMC
PMC10634757
OpenAlex
W4388263226
Cited by
7
References
9
Citations by year
2026: 3 · 2025: 2 · 2024: 2
Type
posted content
Language
en
Open access
Licence
cc-by-nc-nd
Locations
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion, cc-by-nc-nd)
  • Apollo (University of Cambridge) (submittedVersion, cc-by)
Funders
  • Cambridge Trust
  • Gates Cambridge Trust
  • Norges Miljø- og Biovitenskapelige Universitet
  • National Institutes of Health (DK 043351)
  • National Heart, Lung, and Blood Institute (R01HL157717)
  • National Institute of Diabetes and Digestive and Kidney Diseases (P30DK043351, R01DK127171)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gut microbiota and health
Keywords
carbohydrate-active enzymes, metagenomics, protein language models, protein embeddings, functional annotation, Bacteroides thetaiotaomicron, Eggerthella lenta, Ruminococcus gnavus, infant gut microbiota, Crohn's disease, IgG4-related disease, fibrosis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review