Chemical foundation models reveal molecular principles underlying glycobiology
- 1IBM Research Europe - Dublin, IBM Research, Ireland; School of Biochemistry & Immunology, Trinity College Dublin, Dublin 2, Ireland
Abstract
Glycans encode biological specificity through branching, linkage and stereochemical diversity, and have traditionally been modelled using specialised computational representations distinct from those developed for general chemistry. It remains unclear to what extent representations learned from general molecular chemistry capture the molecular principles of glycobiology. Here, we systematically evaluate the transfer of chemical foundation models to glycobiology using standardized benchmarks spanning taxonomy, immunogenicity, glycosylation type and protein-glycan interactions. We show that atom-level representations learned from general chemistry encode biologically meaningful molecular organization, providing particularly strong performance for molecular recognition, including immunogenicity and protein-glycan interaction prediction. Fine-tuning reorganizes these representations into latent spaces aligned with glycan architecture, glycosylation programmes and molecular recognition. Mechanistic analyses identify stereochemistry as an important determinant of successful transfer, while tokenization, context length and conformer generation influence the representation of large, branched glycans. These representations generalize from natural glycans to synthetic glycomimetics and recover experimentally validated stereochemistry-dependent glycan-lectin recognition. Together, our findings demonstrate that molecular principles learned from general chemistry are sufficient to capture fundamental features of glycobiology. They position glycans as a specialised region of chemical space and chemical foundation models as a broadly applicable framework for glycan function prediction, molecular recognition and glycomimetic discovery.
Related
- Decoding enzyme-substrate interaction topology reveals principles underlying catalytic efficiency and mutational outcomes
- Molecular Mechanisms and Signaling Pathways Underlying the Synergistic Effect of Metformin and Paclitaxel in Preclinical Cancer Models: A Systematic Review.
- AlvaQSAR: A Web-Based Platform for Molecular Prediction of Environmental and Toxicological Endpoints Aligned with OECD Principles
- How 'Foundational' Are Current Molecular Foundation Models?
- Near-atomistic simulations reveal the molecular principles that control chromatin structure and phase separation
BibTeX
@misc{ravasqueira2026chemical,
title = {{Chemical foundation models reveal molecular principles underlying glycobiology}},
author = {M. Sokolov Ravasqueira and J. Bettencourt-Silva and V. Lopez Garcia and R. Fernandez-Diaz and N. Mulligan and G. P. Davey},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.13.758502},
url = {https://www.biorxiv.org/content/10.64898/2026.10.13.758502v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Chemical foundation models reveal molecular principles underlying glycobiology
AU - M. Sokolov Ravasqueira
AU - J. Bettencourt-Silva
AU - V. Lopez Garcia
AU - R. Fernandez-Diaz
AU - N. Mulligan
AU - G. P. Davey
PY - 2026
DA - 2026/10/13
PB - bioRxiv
DO - 10.64898/2026.10.13.758502
UR - https://www.biorxiv.org/content/10.64898/2026.10.13.758502v1
AB - Glycans encode biological specificity through branching, linkage and stereochemical diversity, and have traditionally been modelled using specialised computational representations distinct from those developed for general chemistry. It remains unclear to what extent representations learned from general molecular chemistry capture the molecular principles of glycobiology. Here, we systematically evaluate the transfer of chemical foundation models to glycobiology using standardized benchmarks spanning taxonomy, immunogenicity, glycosylation type and protein-glycan interactions. We show that atom-level representations learned from general chemistry encode biologically meaningful molecular organization, providing particularly strong performance for molecular recognition, including immunogenicity and protein-glycan interaction prediction. Fine-tuning reorganizes these representations into latent spaces aligned with glycan architecture, glycosylation programmes and molecular recognition. Mechanistic analyses identify stereochemistry as an important determinant of successful transfer, while tokenization, context length and conformer generation influence the representation of large, branched glycans. These representations generalize from natural glycans to synthetic glycomimetics and recover experimentally validated stereochemistry-dependent glycan-lectin recognition. Together, our findings demonstrate that molecular principles learned from general chemistry are sufficient to capture fundamental features of glycobiology. They position glycans as a specialised region of chemical space and chemical foundation models as a broadly applicable framework for glycan function prediction, molecular recognition and glycomimetic discovery.
ER -