Predicting conformational changes induced by biomolecular interactions
- 1Department of Chemistry; Kim Jaechul Graduate School of AI; Department of AX; InnoCORE AI-CRED Institute, KAIST, Republic of Korea; HITS Inc., Seoul, Republic o
Abstract
Biomolecular interactions trigger structural changes of protein complexes essential for cellular signaling, regulation, and drug action. Modern co-folding models can now generate detailed structures of multi-component complexes directly from sequence. However, recent evaluations reveal that they frequently fail to capture specific conformational states induced by a ligand or partner. Here, we show that explicitly modeling the structural transition from an unbound to a bound state, rather than generating the bound complex alone, better resolves these interaction-induced states. Our generative learning framework, K-Fold, transforms independently predicted unbound component priors into an assembled complex through an ultrafast diffusion bridge, operating entirely without multiple sequence alignments. By eliminating co-evolutionary dependencies, K-Fold accurately recovers non-coevolved interfaces in antibody-antigen and targeted-degrader complexes, as well as ligand-directed regulatory states in GPCRs and kinases. Furthermore, by coupling this diffusion bridge with diverse monomeric equilibrium ensembles, K-Fold provides a computationally tractable foundation for sampling complex multi-state landscapes without retraining. These results demonstrate that formulating molecular assembly as an explicit structural transition moves structure prediction beyond static geometries toward capturing biomolecular dynamics.
Related
- Molecular dynamics analysis of T567 phosphorylation-induced conformational changes in ezrin.
- Navigating the Biomolecular Corona: Biological Interactions and Therapeutic Implications of RNA–Lipid Nanoparticles
- Mapping interactions between disordered regions reveals promiscuity in biomolecular condensate formation
- Molecular determinants of arginine versus lysine cation–π interactions in biomolecular condensates
- Measuring protein conformational changes with genetically encoded bipyridylalanine and transition-metal ion FRET
BibTeX
@misc{bae2026predicting,
title = {{Predicting conformational changes induced by biomolecular interactions}},
author = {Y. Bae and J. Baek and E. Choi and J. H. Choi and H. Han and H. Jang and W. Jeon and J. Jo and D. Kim and H. Kim and J. H. Kim and J. H. Kim and S. Kim and T. Kim and Y. Kim and J. Lee and J. Lee and S. Moon and M. Park and C. Samuel and H. Seo and J. Seo and S. Seo and M. Shin and Y. Shin and W. Zhung and S. Ahn and D. Han and S. J. Hwang and H. M. Kim and W. Y. Kim and G. R. Lee and B.-H. Oh},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.26.760299},
url = {https://www.biorxiv.org/content/10.64898/2026.10.26.760299v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Predicting conformational changes induced by biomolecular interactions
AU - Y. Bae
AU - J. Baek
AU - E. Choi
AU - J. H. Choi
AU - H. Han
AU - H. Jang
AU - W. Jeon
AU - J. Jo
AU - D. Kim
AU - H. Kim
AU - J. H. Kim
AU - J. H. Kim
AU - S. Kim
AU - T. Kim
AU - Y. Kim
AU - J. Lee
AU - J. Lee
AU - S. Moon
AU - M. Park
AU - C. Samuel
AU - H. Seo
AU - J. Seo
AU - S. Seo
AU - M. Shin
AU - Y. Shin
AU - W. Zhung
AU - S. Ahn
AU - D. Han
AU - S. J. Hwang
AU - H. M. Kim
AU - W. Y. Kim
AU - G. R. Lee
AU - B.-H. Oh
PY - 2026
DA - 2026/10/26
PB - bioRxiv
DO - 10.64898/2026.10.26.760299
UR - https://www.biorxiv.org/content/10.64898/2026.10.26.760299v1
AB - Biomolecular interactions trigger structural changes of protein complexes essential for cellular signaling, regulation, and drug action. Modern co-folding models can now generate detailed structures of multi-component complexes directly from sequence. However, recent evaluations reveal that they frequently fail to capture specific conformational states induced by a ligand or partner. Here, we show that explicitly modeling the structural transition from an unbound to a bound state, rather than generating the bound complex alone, better resolves these interaction-induced states. Our generative learning framework, K-Fold, transforms independently predicted unbound component priors into an assembled complex through an ultrafast diffusion bridge, operating entirely without multiple sequence alignments. By eliminating co-evolutionary dependencies, K-Fold accurately recovers non-coevolved interfaces in antibody-antigen and targeted-degrader complexes, as well as ligand-directed regulatory states in GPCRs and kinases. Furthermore, by coupling this diffusion bridge with diverse monomeric equilibrium ensembles, K-Fold provides a computationally tractable foundation for sampling complex multi-state landscapes without retraining. These results demonstrate that formulating molecular assembly as an explicit structural transition moves structure prediction beyond static geometries toward capturing biomolecular dynamics.
ER -