EnsemblePAE: A PAE-Guided Structure-Based Model for Efficient Sampling of Protein Conformational Ensembles from AlphaFold Predictions

AlphaFold has revolutionized protein structure prediction, yet extracting conformational dynamics from a single predicted structure remains a major challenge. Here, we present EnsemblePAE, a PAE-guided structure-based model that transforms AlphaFold’s predicted aligned error (PAE) into heterogeneous native interactions for efficient coarse-grained simulations. By systematically evaluating different PAE rescaling schemes, we identify an optimal formulation that balances structural stability and conformational flexibility. Benchmarking on a diverse set of allosteric proteins and 106 proteins from the DynDom database demonstrates that EnsemblePAE substantially improves the prediction of alternative conformations over conventional structure-based models and achieves performance comparable to BioEmu while requiring only minutes of simulation on a single GPU. Beyond conformational transitions, EnsemblePAE also captures complex protein folding dynamics and reveals metastable folding intermediates absent in conventional structure-based models. These results demonstrate that AlphaFold PAE encodes physically meaningful information about protein energy landscapes and establish EnsemblePAE as a simple, interpretable, and computationally efficient framework for exploring protein conformational ensembles.

Authors

Institutions

Publication Details

Journal
Journal of Chemical Theory and Computation
Published
2026-11-21
DOI
https://doi.org/10.1021/acs.jctc.6c01861
Primary Topic
Protein Structure and Dynamics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

EnsemblePAE: A PAE-Guided Structure-Based Model for Efficient Sampling of Protein Conformational Ensembles from AlphaFold Predictions

Qian-Yuan Tang, Weitong Ren, Wenfei Li, Zhiqiang Yan et al.
Journal of Chemical Theory and Computation
Protein Structure and Dynamics
article

EnsemblePAE: A PAE-Guided Structure-Based Model for Efficient Sampling of Protein Conformational Ensembles from AlphaFold Predictions

Qian-Yuan Tang, Weitong Ren, Wenfei Li, Zhiqiang Yan, Wei Wang, Aoru Liu
article en

Abstract

AlphaFold has revolutionized protein structure prediction, yet extracting conformational dynamics from a single predicted structure remains a major challenge. Here, we present EnsemblePAE, a PAE-guided structure-based model that transforms AlphaFold’s predicted aligned error (PAE) into heterogeneous native interactions for efficient coarse-grained simulations. By systematically evaluating different PAE rescaling schemes, we identify an optimal formulation that balances structural stability and conformational flexibility. Benchmarking on a diverse set of allosteric proteins and 106 proteins from the DynDom database demonstrates that EnsemblePAE substantially improves the prediction of alternative conformations over conventional structure-based models and achieves performance comparable to BioEmu while requiring only minutes of simulation on a single GPU. Beyond conformational transitions, EnsemblePAE also captures complex protein folding dynamics and reveals metastable folding intermediates absent in conventional structure-based models. These results demonstrate that AlphaFold PAE encodes physically meaningful information about protein energy landscapes and establish EnsemblePAE as a simple, interpretable, and computationally efficient framework for exploring protein conformational ensembles.

Journal of Chemical Theory and Computation
Hong Kong Baptist University (HK), University of Chinese Academy of Sciences (CN), Nanjing University (CN)
Protein Structure and Dynamics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon