Page 39 - Read Online

Aydin et al. J Transl Genet Genom. 2025;9:406-26 https://dx.doi.org/10.20517/jtgg.2025.108 Page 408

diagnostic power of the proposed biomarkers and the drug repositioning based on these disease signatures
may provide a methodological illustration for addressing existing gaps in this field.

In this field, we identified 33 genes with the potential to serve as both network and diagnostic biomarkers
for DMD. Among these genes, we found that 17 (SQSTM1, PML, SPTAN1, SPTBN1, KIAA1429, SOX4, SP1,
SPP1, NFKB1, TP53, NKX3-1, CIITA, ARL6IP1, IGFBP5, OCIAD2, RAP2B, and NFIB) showed higher
binding affinities to candidate repositioned drugs, including celastrol, radicicol, withaferin-A, emetine
dihydrochloride hydrate, and apigenin triacetate. It is predicted that emetine dihydrochloride hydrate and
celastrol may act synergistically in the management of DMD, given their essential roles in apoptosis,
autophagy, anti-inflammatory responses, and infection-related pathways.

METHODS
Selection of datasets and identification of differentially expressed genes
The transcriptome-level data were searched through the Gene Expression Omnibus (GEO) database with
the keywords “DMD, Duchenne Muscular Dystrophy, microarray,” and the datasets with the accession
numbers of GSE109178, GSE70955, and GSE38417 were retrieved [15,16] . The data, including six controls and
17 test groups for the GSE109178 dataset, three controls and 3 test groups for the GSE70955 dataset, and six
controls and 16 tests for the GSE38317 dataset, were analyzed using GEO2R (https://www.ncbi.nlm.nih.gov/
geo/geo2r/). For normalization, the Linear Models for Microarray Data (LIMMA) package was employed.
Detailed information on the selected datasets is provided in Table 1. To determine the expression profiles of
these DMD datasets, the cut-off values were adjusted to 0.001 and |log FC| ≥ 1.0 for P-values and fold
2
changes, respectively. Benjamini-Hochberg’s correction was employed to control the false discovery rate.
Genes with log FC > 1.0 were considered upregulated, whereas those with log FC < -1.0 were classified as
2
2
downregulated. The common differentially expressed genes (DEGs) across these three datasets were
identified and used for further analysis.

Functional enrichment analysis
Common DEGs across 3 datasets (GSE38417, GSE70955, and GSE109178) were identified. These genes were
used to enhance the understanding of biological pathways and functions around DEGs using Metascape,
which assists in analyzing certain gene or protein lists to identify the biological processes, cellular
[17]
constituents, or molecular activities associated with the genes or proteins . The heatmap resulting from the
analysis may indicate enriched biological pathways or processes .
[18]
Biological network constructions around DEGs
Three-layered biological network constructions were carried out by protein-protein interactions [PPIs;
BioGrid (Biological General Repository for Interaction Datasets, thebiogrid.org)] , miRNA-target DEG
[19]
interactions (MiRTarBase, https://mirtarbase.cuhk.edu.cn/) , and transcription factor (TF)-target DEG
[20]
interactions (TRRUST, http://www.grnpedia.org/trrust) . Network constructions were conducted using
[21]
Cytoscape 3.10.2 . The topological properties of these networks were analyzed using the CytoHubba
[22]
plugin . Hubs with the highest local and global topological parameters, including degree and betweenness
[23]
centrality obtained from these three networks, were considered as DMD-specific network biomarkers and
potential targets for drug repositioning.

Principal component analysis to establish diagnostic power of network signatures
To further assess whether network biomarkers can discriminate between the healthy and diseased states of
the dataset samples, principal component analysis (PCA) was performed. R software version 4.3.2 and
[24]
RStudio were used for PCA. The “BiocManager” package was used to install the required Bioconductor
[25]
dependencies, while the “factoextra” package was employed for visualization and interpretation of PCA