Explainability of Protein Deep Learning Models
International Journal of Molecular Sciences · 2025
DOI: 10.3390/ijms26115255 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Molecular Sciences, 2025. Verfasst von Zahra Fazel und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, University of Western Ontario, London, ON N6A 5B7, Canada. Hauptthemen laut MEDLINE: Deep Learning, Proteins. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Sciences and Engineering Research Council, NSERC Discovery.
Abstract (Originalsprache)
Protein embeddings are the new main source of information about proteins, producing state-of-the-art solutions to many problems, including protein interaction prediction, a fundamental issue in proteomics. Understanding the embeddings and what causes the interactions is very important, as these models lack transparency due to their black-box nature. In the first study of its kind, we investigate the inner workings of these models using XAI (explainable AI) approaches. We perform extensive testing (3.3 TB of total data) involving nine of the best-known XAI methods on two problems: (i) the prediction of protein interaction sites using the current top method, Seq-InSite, and (ii) the production of protein embedding vectors using three methods, ProtBERT, ProtT5, and Ankh. The results are evaluated in terms of their ability to correlate with six basic amino acid properties—aromaticity, acidity/basicity, hydrophobicity, molecular mass, van der Waals volume, and dipole moment—as well as the propensity for interaction with other proteins, the impact of distant residues, and the infidelity scores of the XAI methods. The results are unexpected. Some XAI methods are much better than others at discovering essential information. Simple methods can be as good as advanced ones. Different protein embedding vectors can capture distinct properties, indicating significant room for improvement in embedding quality.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Computer Science, University of Western Ontario, London, ON N6A 5B7, Canada.
- Department of Statistical and Actuarial Sciences, University of Western Ontario, London, ON N6A 5B7, Canada.
- Department of Biology, McMaster University, Hamilton, ON L6S 4K1, Canada.
ORCID-Profile: Camila P E de Souza, G Brian Golding, Lucian Ilie
Förderung
- Natural Sciences and Engineering Research Council
- NSERC Discovery
Zitieren
APA
Fazel, Z., de Souza, C. P. E., Golding, G. B., & Ilie, L. (2025). Explainability of Protein Deep Learning Models. International Journal of Molecular Sciences. https://doi.org/10.3390/ijms26115255
Vancouver
Fazel Z, de Souza CPE, Golding GB, Ilie L. Explainability of Protein Deep Learning Models. International Journal of Molecular Sciences. 2025. doi:10.3390/ijms26115255
BibTeX
@article{Fazel2025,
author = {Fazel, Zahra and de Souza, Camila P E and Golding, G Brian and Ilie, Lucian},
title = {Explainability of Protein Deep Learning Models},
journal = {International Journal of Molecular Sciences},
year = {2025},
doi = {10.3390/ijms26115255}
}
EUVIMED