PLMFit: benchmarking transfer learning with protein language models for protein engineering
Briefings in Bioinformatics · 02.07.2025 · 26 (4)
DOI: 10.1093/bib/bbaf381 ↗Kurz gefasst
Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 30.07.2025. Verfasst von Thomas Bikias und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biosystems Science and Engineering, ETH Zurich, Basel, Switzerland. Hauptthemen laut MEDLINE: Protein Engineering, Machine Learning, Proteins, Software. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Protein language models (PLMs) have emerged as a useful resource for protein engineering applications. Transfer learning (TL) leverages pre-trained parameters to extract features to train machine learning models or adjust the weights of PLMs for novel tasks via fine-tuning (FT) through back-propagation. TL methods have shown potential for enhancing protein predictions performance when paired with PLMs, however there is a notable lack of comparative analyses that benchmark TL methods applied to state-of-the-art PLMs, identify optimal strategies for transferring knowledge and determine the most suitable approach for specific tasks. Here, we report PLMFit, a benchmarking study that combines, three state-of-the-art PLMs (ESM2, ProGen2, ProteinBert), with three TL methods (feature extraction, low-rank adaptation, bottleneck adapters) for five protein engineering datasets. We conducted over >3150 in silico experiments, altering PLM sizes and layers, TL hyperparameters and different training procedures. Our experiments reveal three key findings: (i) utilizing a partial fraction of PLM for TL does not detrimentally impact performance, (ii) the choice between feature extraction (FE) and fine-tuning is primarily dictated by the amount and diversity of data, and (iii) FT is most effective when generalization is necessary and only limited data is available. We provide PLMFit as an open-source software package, serving as a valuable resource for the scientific community to facilitate the FE and FT of PLMs for various applications.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Biosystems Science and Engineering, ETH Zurich, Basel, Switzerland.
- Botnar Institute of Immune Engineering, Basel, Switzerland.
ORCID-Profile: Thomas Bikias, Sai T Reddy
Zitieren
APA
Bikias, T., Stamkopoulos, E., & Reddy, S. T. (2025). PLMFit: benchmarking transfer learning with protein language models for protein engineering. Briefings in Bioinformatics, 26(4). https://doi.org/10.1093/bib/bbaf381
Vancouver
Bikias T, Stamkopoulos E, Reddy ST. PLMFit: benchmarking transfer learning with protein language models for protein engineering. Briefings in Bioinformatics. 2025;26(4). doi:10.1093/bib/bbaf381
BibTeX
@article{Bikias2025,
author = {Bikias, Thomas and Stamkopoulos, Evangelos and Reddy, Sai T},
title = {PLMFit: benchmarking transfer learning with protein language models for protein engineering},
journal = {Briefings in Bioinformatics},
year = {2025},
volume = {26},
number = {4},
doi = {10.1093/bib/bbaf381}
}
EUVIMED