Transformer‐Based Prediction of Sec‐ and Tat‐Type Signal Peptides for Enhanced Bacterial Protein Secretion
Biotechnology Journal · 2026
DOI: 10.1002/biot.70204 ↗Kurz gefasst
Fachartikel in der Zeitschrift Biotechnology Journal, 2026. Verfasst von Seongmo Kang und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Chemical and Biomolecular Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea. Hauptthemen laut MEDLINE: Protein Sorting Signals, Bacterial Proteins. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 41 Literaturangaben. Gefördert von: National Research Foundation, Korean government (MSIT).
Abstract (Originalsprache)
Designing signal peptides (SPs) for efficient recombinant protein secretion remains challenging, as current approaches depend largely on labor‐intensive screening. We developed a transformer‐based model trained on 158,768 SP‐protein pairs from Gram‐positive bacteria to generate type‐specific Sec‐ or Tat‐type SPs for given mature proteins. The model uses tailored tokenization strategies, including region delimiter tokens, to enable region‐aware sequence design. In silico evaluation showed accurate SP classification and mean pairwise sequence identities above 60% compared with native SPs. Training exclusively on Gram‐positive data outperformed training on a universal dataset, highlighting mechanistic differences in SP architectures. Beam‐search decoding and additional sampling methods further improved sequence diversity and ensured robust SP generation. Experimental validation in Corynebacterium glutamicum demonstrated successful secretion for 15 of 16 designed SPs across two target proteins (M18 and XynA). This study establishes a practical, data‐driven framework for rational SP design, supporting more efficient protein biomanufacturing in Gram‐positive hosts.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Department of Chemical and Biomolecular Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea.
- Graduate School of Engineering Biology, KAIST, Daejeon, Republic of Korea.
- KI For the BioCentury, KAIST, Daejeon, Republic of Korea.
- BioProcess Engineering Research Center, KAIST, Daejeon, Republic of Korea.
Förderung
- National Research Foundation
- Korean government (MSIT)
Zitieren
APA
Kang, S., Lee, S. M., Park, R. H., Lee, G., Lee, J. H., Jeong, K. J., & Kim, H. U. (2026). Transformer‐Based Prediction of Sec‐ and Tat‐Type Signal Peptides for Enhanced Bacterial Protein Secretion. Biotechnology Journal. https://doi.org/10.1002/biot.70204
Vancouver
Kang S, Lee SM, Park RH, Lee G, Lee JH, Jeong KJ, et al. Transformer‐Based Prediction of Sec‐ and Tat‐Type Signal Peptides for Enhanced Bacterial Protein Secretion. Biotechnology Journal. 2026. doi:10.1002/biot.70204
BibTeX
@article{Kang2026,
author = {Kang, Seongmo and Lee, Seong Min and Park, Ryu Hong and Lee, Gunhyeong and Lee, Je Hyeon and Jeong, Ki Jun and Kim, Hyun Uk},
title = {Transformer‐Based Prediction of Sec‐ and Tat‐Type Signal Peptides for Enhanced Bacterial Protein Secretion},
journal = {Biotechnology Journal},
year = {2026},
doi = {10.1002/biot.70204}
}
EUVIMED