Cerebra: A computationally efficient framework for accurate protein structure prediction
Protein Science · 2026-12 · 35 (12)
DOI: 10.1002/pro.70846 ↗At a glance
Journal article in Protein Science, published online on 2026-11-14. Written by Jian Hu and 4 co-authors. Institution: MOE Key Laboratory of Bioinformatics, School of Life Sciences, Tsinghua University, Beijing, China. Major MEDLINE topics: Proteins, Computational Biology, Software. 45 references. The full text is freely available in PubMed Central. Funded by: National Natural Science Foundation of China, Ministry of Science and Technology of the People's Republic of China, Beijing Frontier Research Center for Biological Structure.
Abstract (original language)
Remarkable progress has been made in the field of protein structure prediction. Representative methods like AlphaFold and RoseTTAFold achieve prediction accuracy close to experimental structural determination, but at the cost of heavy computational consumption for model training. In this work, we propose a new framework, Cerebra, for improving the computational efficiency of protein structure prediction. In this network architecture, multiple sets of atomic coordinates are predicted parallelly and their mutual complementarity is leveraged to rapidly improve the quality of predicted structures through a new attention mechanism named path synthesis attention (PSA). Consequently, Cerebra markedly reduces the computational consumption, achieving evident acceleration on both model training and inference, in comparison to OpenFold, the academic version of AlphaFold2. When evaluated on the CAMEO 2025 dataset, the Cerebra model trained on a limited number of GPUs shows comparable performance to the published OpenFold model that has been thoroughly optimized using a plethora of computational resources. Moreover, ablation studies confirm that the PSA mechanism efficiently accelerates the network convergence, suggesting its potential in handling the many‐body effects of protein geometries in graph‐based models. Finally, the fast inference of Cerebra effectively assists hallucination‐based de novo protein backbone design, allowing the production of abundant backbone structures with sufficient designability and novelty.
Subjects
MeSH terms (MEDLINE)
Author keywords
Substances
Authors and institutions
- MOE Key Laboratory of Bioinformatics, School of Life Sciences, Tsinghua University, Beijing, China.
- State Key Laboratory of Green Manufacturing, Tsinghua University, Beijing, China.
- Beijing Frontier Research Center for Biological Structure, Tsinghua University, Beijing, China.
ORCID profiles: Weizhe Wang, Bo Zhang, Yike Cao, Haipeng Gong
Funding
- National Natural Science Foundation of China
- Ministry of Science and Technology of the People's Republic of China
- Beijing Frontier Research Center for Biological Structure
Cite
APA
Hu, J., Wang, W., Zhang, B., Cao, Y., & Gong, H. (2026). Cerebra: A computationally efficient framework for accurate protein structure prediction. Protein Science, 35(12). https://doi.org/10.1002/pro.70846
Vancouver
Hu J, Wang W, Zhang B, Cao Y, Gong H. Cerebra: A computationally efficient framework for accurate protein structure prediction. Protein Science. 2026;35(12). doi:10.1002/pro.70846
BibTeX
@article{Hu2026,
author = {Hu, Jian and Wang, Weizhe and Zhang, Bo and Cao, Yike and Gong, Haipeng},
title = {Cerebra: A computationally efficient framework for accurate protein structure prediction},
journal = {Protein Science},
year = {2026},
volume = {35},
number = {12},
doi = {10.1002/pro.70846}
}
EUVIMED