Screening Biomarkers for Nerve Injury Using Weighted Gene Co‐Expression Network Analysis and Machine Learning
Brain and Behavior · 2026
DOI: 10.1002/brb3.71279 ↗At a glance
Journal article in Brain and Behavior, 2026. Written by Shuming Cao and 4 co-authors. Institution: Clinical School/College of Orthopedics, Tianjin Medical University, Tianjin, China. Major MEDLINE topics: Machine Learning, Peripheral Nerve Injuries, Gene Regulatory Networks. Cited 2 times so far (Crossref); 48 references. The full text is freely available (open access, CC BY 4.0).
Abstract (original language)
Background
Nerve injury triggers complex molecular responses involving immune activation and neuronal damage, yet the key regulatory genes and their mechanisms remain poorly understood. Here, we integrated multi‐transcriptomic datasets and machine learning to identify and validate novel biomarkers of nerve injury and elucidate their functional roles.
Methods
The RNA‐seq data and the single‐cell transcriptome data of nerve injury and sham‐surgery samples were sourced from the Gene Expression Omnibus (GEO) database. Weighted gene co‐expression network analysis (WGCNA), differential expression analysis, and three machine learning algorithms were used to identify hub genes associated with nerve injury. The expression patterns and diagnostic value of these hub genes were validated in independent datasets. The correlation between these genes and immune cell infiltration was analyzed using the CIBERSORT algorithm. Finally, single‐cell RNA sequencing (scRNA‐seq) data were used to investigate the cell‐specific expression patterns of the hub genes in neural cells.
Results
Seven nerve injury‐related genes were identified via WGCNA and three machine learning methods, of which Atf3 , Bin2 , Fcgr2b , and Ucn exhibited robust diagnostic performance (AUC > 0.7) across validation cohorts. Functional enrichment implicated these genes in neuroinflammation, neuronal fate commitment, and JAK‐STAT/NF‐κB signaling. Immune infiltration analysis correlated their expression with M2 macrophage polarization and CD4+ T cell depletion, while scRNA‐seq highlighted cell‐specific patterns: Atf3 and Ucn were neuron‐enriched, whereas Fcgr2b and Bin2 predominated in macrophages/NK cells. Moreover, Fcgr2b promoted the outgrowth of neurites in PC12 cells.
Conclusion
Our study unveils Atf3 , Bin2 , Fcgr2b , and Ucn as critical nerve injury biomarkers with dual roles in neuroimmune crosstalk, offering novel insights into therapeutic targeting for nerve repair. Moreover, Fcgr2b may be involved in neurite outgrowth after nerve injury.
Subjects
MeSH terms (MEDLINE)
Author keywords
Substances
Authors and institutions
- Clinical School/College of Orthopedics, Tianjin Medical University, Tianjin, China.
- Department of Hand Surgery, Tianjin Hospital, Tianjin, China.
- Tianjin University Tianjin Hospital, Tianjin, China.
- Department of General Internal Medicine, Tianjin Hospital, Tianjin, China.
ORCID profiles: Shuming Cao, Weiguo Xu
Cite
APA
Cao, S., Yu, C., Wang, N., Xu, J., & Xu, W. (2026). Screening Biomarkers for Nerve Injury Using Weighted Gene Co‐Expression Network Analysis and Machine Learning. Brain and Behavior. https://doi.org/10.1002/brb3.71279
Vancouver
Cao S, Yu C, Wang N, Xu J, Xu W. Screening Biomarkers for Nerve Injury Using Weighted Gene Co‐Expression Network Analysis and Machine Learning. Brain and Behavior. 2026. doi:10.1002/brb3.71279
BibTeX
@article{Cao2026,
author = {Cao, Shuming and Yu, Chengyue and Wang, Nana and Xu, Jianhua and Xu, Weiguo},
title = {Screening Biomarkers for Nerve Injury Using Weighted Gene Co‐Expression Network Analysis and Machine Learning},
journal = {Brain and Behavior},
year = {2026},
doi = {10.1002/brb3.71279}
}
EUVIMED