A Generative Neuro‐Symbolic AI for Protein Sequence Design
Advanced Science · 30.07.2026
DOI: 10.1002/advs.76464 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advanced Science, online erschienen am 30.07.2026. Verfasst von Marianne Defresne und 12 weiteren Autor:innen. Beteiligte Einrichtung: TBI, Université de Toulouse, CNRS, INRAE, INSA, ANITI, Toulouse, France. Die Arbeit nennt 133 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: French National Research Agency (ANR), EUR BioEco, ANITI AI cluster.
Abstract (Originalsprache)
Deep learning has revolutionized computational protein design, enabling the generation of sequences that fold onto target backbones with unprecedented accuracy. However, state‐of‐the‐art inverse folding tools largely rely on auto‐regressive sampling. While powerful, this paradigm is increasingly recognized for its inability to “think ahead”, a crucial capacity to reliably create the complex, long‐range inter‐residue dependencies essential for most biological functions. To overcome these fundamental limitations, we introduced EffieDes , a generative neuro‐symbolic AI framework that synergizes the predictive capabilities of deep learning with the logical precision of automated reasoning. EffieDes leverages deep learning to encode the target backbone's fitness landscape into Effie —a fully decomposable probabilistic graphical model (Potts model). This landscape can then be rigorously explored by an automated reasoning prover to identify sequences that simultaneously satisfy complex design constraints and optimize backbone fitness. We validated this neuro‐symbolic approach through the design of orthogonal sequence pairs that adopt identical folds but exhibit selective self‐assembly, as well as the design of a de novo selective nanobody with nanomolar affinity for an immune‐evasive SARS‐CoV‐2 variant. EffieDes provides a robust architecture for precisely dissecting learned fitness landscapes, offering a new path toward proteins with highly optimized performances and sophisticated functional objectives.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- TBI, Université de Toulouse, CNRS, INRAE, INSA, ANITI, Toulouse, France.
- MIAT, Université de Toulouse, ANITI, INRAE, Toulouse, France.
- LISM, CNRS, Université d'Aix-Marseille, Marseille, France.
ORCID-Profile: Thomas Schiex, Sophie Barbe
Förderung
- French National Research Agency (ANR)
- EUR BioEco
- ANITI AI cluster
- CALMIP
- Jean-Zay GENCI-IDRIS
Zitieren
APA
Defresne, M., Dessaux, D., Buchet, S., Barthe, L., Ammar-Khodja, L., Azizi, B., Durante, V., Cioci, G., de Givry, S., Roussel, A., Garcia-Alles, L. F., Schiex, T., & Barbe, S. (2026). A Generative Neuro‐Symbolic AI for Protein Sequence Design. Advanced Science. https://doi.org/10.1002/advs.76464
Vancouver
Defresne M, Dessaux D, Buchet S, Barthe L, Ammar-Khodja L, Azizi B, et al. A Generative Neuro‐Symbolic AI for Protein Sequence Design. Advanced Science. 2026. doi:10.1002/advs.76464
BibTeX
@article{Defresne2026,
author = {Defresne, Marianne and Dessaux, Delphine and Buchet, Samuel and Barthe, Lucie and Ammar-Khodja, Liza and Azizi, Bessam and Durante, Valentin and Cioci, Gianluca and de Givry, Simon and Roussel, Alain and Garcia-Alles, Luis F and Schiex, Thomas and Barbe, Sophie},
title = {A Generative Neuro‐Symbolic AI for Protein Sequence Design},
journal = {Advanced Science},
year = {2026},
doi = {10.1002/advs.76464}
}
EUVIMED