Identification and validation of feature genes using a ferroptosis-informed analytical framework in medulloblastoma.

BACKGROUND

This study aimed to identify potential biomarker signatures for medulloblastoma using a ferroptosis-informed analytical framework.

METHODS

Transcriptomic data from GEO datasets were used to identify differentially expressed ferroptosis-associated genes (DEFeGs) in medulloblastoma by intersecting differentially expressed genes with the FerrDb V2 database. A ferroptosis-related gene (FeRG) score was calculated using ssGSEA. Hub genes correlated with the FeRG score were identified using WGCNA. Candidate feature genes were selected using machine learning algorithms. A nomogram was constructed to predict the probability of medulloblastoma diagnosis. Immune infiltration analysis, drug sensitivity prediction, and qRT-PCR validation of feature gene expression in human samples were performed.

RESULTS

A total of 36 DEFeGs associated with medulloblastoma were identified. WGCNA identified 114 hub genes, of which 58 were differentially expressed. Machine learning algorithms identified CEND1, LRP1B, and FEZ1 as candidate feature genes. A nomogram based on these genes showed high diagnostic performance, which was supported by ROC analysis (all AUCs > 0.85). Nineteen immune cell types differed significantly between the medulloblastoma and control groups. Additionally, FEZ1, CEND1, and LRP1B showed positive correlations with sensitivity to SGX-523, sabutoclax, and AZD-1208, respectively, whereas all three genes were negatively correlated with sensitivity to allopurinol. The three feature genes were consistently downregulated in medulloblastoma across public datasets and the validation cohort.

CONCLUSION

This study identified FEZ1, CEND1, and LRP1B as potential candidate biomarker genes for medulloblastoma through a ferroptosis-informed analytical framework.