CoChat
Sign Up
Cited by 5Genomics and Phylogenetic StudiesMachine Learning in BioinformaticsFractal and DNA sequence analysis

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

Protein embeddings and local alignments

Open access
Malec, Julia, Golding, G.Brian, Ilie, Lucian·2026·Computational and Structural Biotechnology Journal
DOI 10.1016/j.csbj.2025.12.002PMC PMC12769827

Published version of a preprint

Open this paper in CoChat — free account

Abstract

Background: The advent of protein embeddings has revolutionized bioinformatics by providing contextual representations that capture functional and evolutionary patterns. They have become, alongside sequence alignments, the cornerstone of bioinformatics. Embeddings cannot replace alignments but they can greatly help improve their quality. While embedding-based improvements have been considered for global alignments, the more important counterpart, local alignments, has not been studied thoroughly. Our goal is to identify the most accurate local alignment algorithm for protein sequences. Results: We introduce a new scoring function into our previous E-score algorithm by using Ankh embeddings. We prove that the resulting algorithm produces the most accurate local alignments of protein sequences using a new comprehensive framework that enables thorough evaluation of local alignment quality. We design a new algorithm for local alignment extraction, localization and quality evaluation and employ five distance metrics to evaluate the similarity with the true alignment. We also build multiple datasets, using both natural and inserted sequences, from the Conserved Domain Database, BAliBASE, and GPCRdb. We perform over two and a half million tests to compare the new algorithm with the best BLOSUM matrices, specialized GPCRtm matrices, and top programs, such as PEbA, DEDAL, vcMSA and pLM-BLAST. Our testing also reveals interesting insights into the behaviour of various protein language models as some of them perform much better on natural sequences compared to artificial ones obtained by inserting domains into random protein sequences. Also, while some models combine to produce better results, Ankh does not combine well with other embeddings. Conclusions: The new, Ankh-score-based, program is clearly superior to all existing methods. New light shed on the protein embeddings can guide future improvements. In order to facilitate the use of the new method and protocol, they are freely available as a web server at e-score.csd.uwo.ca and as source code at github.com/lucian-ilie/E-score .

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 · 10.1016/j.csbj.2025.12.002

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: 40

  1. 10.1016/S0022-2836(05)80360-2 · DOI · Paper page
  2. 10.1093/nar/25.17.3389 · DOI · Paper page
  3. 10.1371/journal.pone.0141287 · DOI
  4. 10.1093/bioinformatics/btr701 · DOI
  5. Dayhoff M., Schwartz R. and Orcutt B.: 22 a model of evolutionary change in proteins. Atlas Protein Seq Struct 1978; 5: 345.
  6. Devlin J. Chang M.W. Lee K. and Toutanova K.B.: Pre-training of deep bidirectional transformers for language understanding. 2018. arXiv preprint arXiv:181004805.
  7. Elnaggar A. Essam H. Salah-Eldin W. Moustafa W. Elkerdawy M. Rochereau C. et al: Ankh: Optimized protein language model unlocks general-purpose modelling. 2023. arXiv preprint arXiv:230106568. · DOI
  8. 10.1038/s41467-021-23303-9 · DOI
  9. 10.1016/0022-2836(82)90398-9 · DOI
  10. Heinzinger M. Weissenow K. Sanchez J. Henkel A. Steinegger M. and Rost B.: ProstT5: bilingual language model for protein sequence and structure. bioRxiv 2023; 23. https://www.biorxiv.org/content/early/2023/07/25/202307. · DOI · Paper page
  11. Henikoff S. and Henikoff J.G.: Amino acid substitution matrices from protein blocks. In: Proceedings of the national academy of sciences. : 1992: 10915.
  12. Hosseini S., Golding G.B. and Ilie L.: Seq-InSite: sequence supersedes structure for protein interaction site prediction. Bioinformatics 01 2024; 40: btad738. · DOI · Paper page
  13. Iovino B.G. and Ye Y.: Protein embedding based alignment. BMC Bioinformatics 2024; 25: 85. · DOI · Paper page
  14. Johnson S.R., Peshwa M. and Sun Z.: Sensitive remote homology search by local alignment of small positional embeddings from protein language models. Elife 2024; 12: RP91415. · DOI · Paper page
  15. 10.1093/bioinformatics/btad579 · DOI
  16. Kulmanov M. and Hoehndorf R.: DeepGOPlus: improved protein function prediction from sequence. Bioinformatics 2020; 36: 422. · DOI
  17. Lai B. and Xu J.: Accurate protein function prediction via graph attention networks with predicted structure information. Brief Bioinform 2022; 23: bbab502. · DOI
  18. Lesk A.M. and Konagurthu A.S.: Protein structure prediction improves the quality of amino-acid sequence alignment. Proteins Struct Funct Bioinform 2022; 90: 2144. · DOI
  19. Liu Y. Ott M. Goyal N. Du J. Joshi M. Chen D. et al: RoBERTa: a robustly optimized bert pretraining approach. 2019. arXiv preprint arXiv:190711692.
  20. 10.1038/s41592-022-01700-2 · DOI · Paper page
  21. Malec J., Rusen K., Golding G.B. and Ilie L.: Ankh-score produces better sequence alignments than AlphaFold3. bioRxiv 2025; . 2025–09.
  22. Manfredi M., Savojardo C., Martelli P.L. and Casadio R.: ISPRED-SEQ: deep neural networks and embeddings for predicting interaction sites in protein sequences. J Mol Biol 2023; 435: 167963. · DOI
  23. 10.1093/nar/gku1221 · DOI
  24. McWhite C.D., Armour-Garb I. and Singh M.: Leveraging protein language models for accurate multiple sequence alignments. Genome Res 2023; 33: 1145.
  25. Mikolov T. Chen K. Corrado G. and Dean J.: Efficient estimation of word representations in vector space. 2013. arXiv preprint arXiv:13013781.
  26. Mizuguchi K., Deane C.M., Blundell T.L. and Overington J.P.: HOMSTRAD: a database of protein structure alignments for homologous families. Protein science 1998; 7: 2469. · DOI
  27. Nambiar A., Heflin M., Liu S., Maslov S., Hopkins M. and Ritz A.: Transforming the language of life: transformer neural networks for protein prediction tasks. In: Proceedings of the 11th ACM international conference on bioinformatics, computational biology and health informatics. : 2020: 1.
  28. 10.1093/nar/gkab852 · DOI
  29. Pantolini L., Studer G., Pereira J., Durairaj J., Tauriello G. and Schwede T.: Embedding-based alignment: combining protein language models with dynamic programming alignment to detect structural similarities in the twilight-zone. Bioinformatics 2024; 40: btad786. · DOI · Paper page
  30. Pennington J., Socher R. and Manning C.D.: GloVe: global vectors for word representation. In: Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP). : 2014: 1532. · DOI · Paper page
  31. Raffel C. Shazeer N. Roberts A. Lee K. Narang S. Matena M. et al: Exploring the limits of transfer learning with a unified text-to-text transformer. 2020. arXiv preprint arXiv:191010683.
  32. 10.1016/j.sbi.2023.102539 · DOI
  33. Rives A., Meier J., Sercu T., Goyal S., Lin Z., Liu J. et al: Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. In: Proceedings of the national academy of sciences. : 2021: e2016239118.
  34. Schütze K., Heinzinger M., Steinegger M. and Rost B.: Nearest neighbor search on embeddings rapidly identifies distant protein relations. Frontiers Bioinformatics 2022; 2: 1033775. · DOI
  35. 10.1038/s41586-019-1923-7 · DOI · Paper page
  36. 10.1016/0022-2836(81)90087-5 · DOI
  37. Thompson J.D., Koehl P., Ripp R. and Poch O.: BAliBASE 3.0: latest developments of the multiple sequence alignment benchmark. Proteins Struct Funct Bioinform 2005; 61: 127. · DOI
  38. 10.1093/nar/gky092 · DOI
  39. Yang J., Anishchenko I., Park H., Peng Z., Ovchinnikov S. and Baker D.: Improved protein structure prediction using predicted interresidue orientations. In: Proceedings of the national academy of sciences. : 2020: 1496.
  40. Yang Z., Dai Z., Yang Y., Carbonell J., Salakhutdinov R.R. and Le Q.V.: XLNet: generalized autoregressive pretraining for language understanding. Adv Neural Inf Process Syst 2019; : 32.

Metadata

Authors
Malec, J., Golding, G. B., & Ilie, L.
Affiliations
  • Western University (CA)
  • McMaster University (CA)
Corresponding author
Ilie, Lucian
ORCID
Venue
Computational and Structural Biotechnology Journal
Publisher
American Association for the Advancement of Science (AAAS)
Published
2025-12-15
Volume
31
Pages
24-37
ISSN
2001-0370 (ISSN-L)
DOI
10.1016/j.csbj.2025.12.002
PMC
PMC12769827
OpenAlex
W4417105621
Cited by
5
References
9
Citations by year
2026: 5
FWCI
1.50
Type
journal article
Language
en
Open access
Licence
Locations
  • Computational and Structural Biotechnology Journal (publishedVersion, cc-by-nc-nd)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Funders
Natural Sciences and Engineering Research Council of Canada (RGPIN 2021\u201303978, RGPIN 2020\u201305733)
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 › Fractal and DNA sequence analysis
Keywords
protein embeddings, local alignment, protein language models, BAliBASE, GPCRdb

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