Protein Language Models Identify Early Folding Residues as Critical Hotspots for Pathogenic Variants and Cancer Prognosis
5 affiliations
- 1Department of Biostatistics, School of Public Health, Cheeloo College of Medicine, Shandong University, Jinan 250100, China
- 2National Institute of Health Data Science of China, Shandong University, Jinan 250100, China
- 3Key Laboratory of Computing Power Network and Information Security, Ministry of Education, Shandong Computer Science Center (National Supercomputer Center in Jinan), Qilu University of Technology (Shandong Academy of Sciences), Jinan 250300, China
- 4School of Data and Computer Science, Shandong Women’s University, Jinan 250300, China
- 5IBIVU–Center for Integrative Bioinformatics, Department of Computer Science, Vrije Universiteit Amsterdam, 1081HV Amsterdam, The Netherlands
Abstract
Identifying the specific residues that initiate protein folding is critical for understanding protein stability and the mechanisms of misfolding-related diseases. However, the systematic, proteome-wide identification of these early folding residues (EFRs) has been hampered by both experimental and computational limitations. Here, we present LMEFold, a deep learning framework using protein language models (PLMs) to accurately predict EFRs from amino acid sequence alone. LMEFold outperforms existing sequence-based predictors on the Start2Fold benchmark (ROC AUC = 0.831), and its predicted scores show consistent associations with independent HDX-NMR protection-factor measurements, providing complementary biophysical evidence. Applying LMEFold to annotate over 6.6 million variants across diverse genomic datasets, we find that pathogenic mutations curated in ClinVar are significantly enriched within EFRs (~17%) compared to millions of population variants from large-scale cohorts including the UKB and gnomAD databases (~10%, P < 0.001). Critically, in the MSK-MET pan-cancer clinical cohort, EFR mutations are associated with severe functional disruption and significantly reduced overall patient survival. Structural analysis of hotspot cancer mutations, such as TP53 p.Y163C, suggests a mechanism whereby these variants destabilize protein structure by disrupting the early folding core. Our work establishes EFRs as critical sites of pathogenic vulnerability and provides a powerful framework for the rapid identification of folding-centric prognostic markers using sequence only.
Related
- Computational Studies of the Dynamics of the Human Iron Exporter Ferroportin and Insight into the Role of Crucial Residues Involved in Pathogenic Mutations
- Loss of residues required for β-arrestin recruitment in the calcium-sensing receptor C-terminal tail may contribute to hypocalcemia
- Sequence Engineering at Non-motif Modulator Residues Yields a Peptide That Effectively Targets a Single PDZ Protein in a Disease-relevant Cellular Context
- Probing Catalytic Residues Responsible for Primer Substrate Reactivity in a Versatile cis -Geranylgeranyl Diphosphate Synthase
- Mutational screening reveals a cluster of residues within the SARS-CoV-2 nsp1 N-terminus that confers RNA-targeting selectivity
BibTeX
@article{guo2026protein,
title = {{Protein Language Models Identify Early Folding Residues as Critical Hotspots for Pathogenic Variants and Cancer Prognosis}},
author = {Lisha Guo and Fuchong Yuan and Na Zhou and Minghui Zhao and Ruotong Liu and Ming Jing and K. Anton Feenstra and Qingzhen Hou},
journal = {International journal of molecular sciences},
year = {2026},
month = oct,
doi = {10.3390/ijms27198914},
url = {https://europepmc.org/article/PMC/PMC13666245},
note = {CC BY},
}
RIS
TY - JOUR
TI - Protein Language Models Identify Early Folding Residues as Critical Hotspots for Pathogenic Variants and Cancer Prognosis
AU - Lisha Guo
AU - Fuchong Yuan
AU - Na Zhou
AU - Minghui Zhao
AU - Ruotong Liu
AU - Ming Jing
AU - K. Anton Feenstra
AU - Qingzhen Hou
PY - 2026
DA - 2026/10/01
JO - International journal of molecular sciences
DO - 10.3390/ijms27198914
UR - https://europepmc.org/article/PMC/PMC13666245
AB - Identifying the specific residues that initiate protein folding is critical for understanding protein stability and the mechanisms of misfolding-related diseases. However, the systematic, proteome-wide identification of these early folding residues (EFRs) has been hampered by both experimental and computational limitations. Here, we present LMEFold, a deep learning framework using protein language models (PLMs) to accurately predict EFRs from amino acid sequence alone. LMEFold outperforms existing sequence-based predictors on the Start2Fold benchmark (ROC AUC = 0.831), and its predicted scores show consistent associations with independent HDX-NMR protection-factor measurements, providing complementary biophysical evidence. Applying LMEFold to annotate over 6.6 million variants across diverse genomic datasets, we find that pathogenic mutations curated in ClinVar are significantly enriched within EFRs (~17%) compared to millions of population variants from large-scale cohorts including the UKB and gnomAD databases (~10%, P < 0.001). Critically, in the MSK-MET pan-cancer clinical cohort, EFR mutations are associated with severe functional disruption and significantly reduced overall patient survival. Structural analysis of hotspot cancer mutations, such as TP53 p.Y163C, suggests a mechanism whereby these variants destabilize protein structure by disrupting the early folding core. Our work establishes EFRs as critical sites of pathogenic vulnerability and provides a powerful framework for the rapid identification of folding-centric prognostic markers using sequence only.
SN - 1422-0067
ER -