Burns. 2025 Nov 19;52(1):107788. doi: 10.1016/j.burns.2025.107788. Online ahead of print.

BACKGROUND: The pathogenesis of keloids is still unclear and effective biomarkers are lacking. Therefore, it is urgent to find clinically effective biomarkers and study their regulatory mechanisms.

METHODS: Candidate genes were screened by amplifying RNA methylation-related gene modules by weighted correlation network analysis and intersecting them with differentially expressed genes in the GSE145725 dataset. After Mendelian randomisation, machine learning and gene expression validation, keloid prediction models were constructed. Subsequently, genomic enrichment analysis, immune infiltration, GeneMANIA analysis, molecular network and drug prediction were performed, and biomarker function was explored using single-cell datasets. Finally, biomarker expression in clinical samples was validated by reverse transcription quantitative polymerase chain reaction (RT-qPCR).

RESULTS: A total of two biomarkers (PEAR1 and MAPKAPK3) were identified and the disease prediction model performed well. They were mainly involved in the proteasomal and ribosomal pathways and were associated with myeloid dendritic cells and resting T CD4 + memory cells. In addition, fibroblasts were identified as key cells expressing PEAR1 and MAPKAPK3. RT-qPCR confirmed that the expression of both biomarkers was downregulated in the keloid group, consistent with the results of differential expression analysis.

CONCLUSION: This study suggests that PEAR1 and MAPKAPK3 may contribute to keloid formation and provides insights for subsequent functional studies.

PubMed:41297235 | DOI:10.1016/j.burns.2025.107788