Deep-Learning Predictions of Biomolecular Structures: Persistent Limitations and New Horizons Extended by Explicit Ion Addition
Journal of Chemical Information and Modeling · 10.11.2026
DOI: 10.1021/acs.jcim.6c02555 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Chemical Information and Modeling, online erschienen am 10.11.2026. Verfasst von Jules Marien und 16 weiteren Autor:innen. Die Arbeit nennt 91 Literaturangaben. Gefördert von: Agence Nationale de la Recherche.
Abstract (Originalsprache)
The advent of deep-learning-driven tools such as AlphaFold has revolutionized the prediction of biomolecular structures, offering unprecedented accuracy and accessibility for proteins, RNA, and their complexes. While these tools have demonstrated remarkable success in benchmarking competitions and enabled experimentalists to generate models with ease, their widespread use has also highlighted persistent challenges. These include difficulties in assessing model confidence, limitations in predicting transmembrane domains, nucleic acids, conformational diversity, and interactions with ions or ligands, as well as the tendency to misfold intrinsically disordered regions (IDRs). In this perspective, we critically evaluate the strengths and limitations of current AI-based structure prediction tools through illustrative examples with a particular emphasis on the impact of explicit ion modeling. We notably report that the explicit addition of a few potassium cations to the prediction of IDRs or G-quadruplexes can substantially shift AF3’s predicted conformation compared to “dry” predictions. On this basis, we suggest modeling sequences both “dry” and in the presence of explicit potassium cations as a simple, practical way to sample alternative conformations and to expose disordered regions that current predictors tend to overfold. We discuss the importance of reporting confidence metrics in publications to avoid overinterpretation. Furthermore, we address the unique challenges of RNA structure prediction, where data scarcity and structural complexity limit the performance of both classical and deep learning methods. Our analysis underscores the need for continued methodological advancements, integration of complementary computational tools, and expansion of high-quality experimental data sets.
Förderung
- Agence Nationale de la Recherche
Zitieren
APA
Marien, J., Sritharan, S., Caviglia, B., Versini, R., Auclair, L., Barraud, P., Tisné, C., Reguei, A., Murail, S., Leclerc, F., Dijon, E. D., Tao, J., Basdevant, N., Baaden, M., Prévost, C., Sacquin-Mora, S., & Taly, A. (2026). Deep-Learning Predictions of Biomolecular Structures: Persistent Limitations and New Horizons Extended by Explicit Ion Addition. Journal of Chemical Information and Modeling. https://doi.org/10.1021/acs.jcim.6c02555
Vancouver
Marien J, Sritharan S, Caviglia B, Versini R, Auclair L, Barraud P, et al. Deep-Learning Predictions of Biomolecular Structures: Persistent Limitations and New Horizons Extended by Explicit Ion Addition. Journal of Chemical Information and Modeling. 2026. doi:10.1021/acs.jcim.6c02555
BibTeX
@article{Marien2026,
author = {Marien, Jules and Sritharan, Sujith and Caviglia, Beatrice and Versini, Raphaelle and Auclair, Lucas and Barraud, Pierre and Tisné, Carine and Reguei, Alaa and Murail, Samuel and Leclerc, Fabrice and Dijon, Elise Duboué and Tao, JianJun and Basdevant, Nathalie and Baaden, Marc and Prévost, Chantal and Sacquin-Mora, Sophie and Taly, Antoine},
title = {Deep-Learning Predictions of Biomolecular Structures: Persistent Limitations and New Horizons Extended by Explicit Ion Addition},
journal = {Journal of Chemical Information and Modeling},
year = {2026},
doi = {10.1021/acs.jcim.6c02555}
}
EUVIMED