Exploratory multi‑omics analysis for candidate biomarker identification in Parkinson's disease

Parkinson's disease (PD) is a progressive neurodegenerative disorder lacking established clinical biomarkers. We performed an exploratory multi‑omics analysis of transcriptome and proteome data from small PD and control cohorts at Huaihe Hospital of Henan University. Using two local datasets, we identified 124 differentially expressed genes (DEGs) and 28 proteins, then performed protein‑protein interaction (PPI) and gene set variation analysis (GSVA). After intersecting with 2,964 DEGs from Gene Expression Omnibus (GEO) Dataset 3, we obtained a 28‑gene panel. Potential markers were prioritized via random forest (RF), support vector machine (SVM) and principal component analysis (PCA), and evaluated in external GEO data, clinical serum samples and a 1‑methyl‑4‑phenyl‑1,2,3,6‑tetrahydropyridine (MPTP)‑induced PD mouse model. PPI highlighted C‑X‑C motif chemokine ligand 1 (CXCL1), C‑X‑C chemokine receptor type 4 (CXCR4), S100 calcium‑binding protein A12 (S100A12) and C‑X‑C motif chemokine ligand 11 (CXCL11) as hub genes. GSVA indicated upregulated cardiac muscle contraction and oxidative phosphorylation (OXPHOS). C‑C motif chemokine ligand 4 (CCL4), CKLF like MARVEL transmembrane domain containing 2 (CMTM2), tudor domain containing 6 (TDRD6), potassium voltage‑gated channel subfamily S member 1 (KCNS1) and member RAS oncogene family (RAB15) were identified as candidate markers. They showed consistent expression changes in clinical samples and mice. Immune infiltration differed in B cells, CD8 + T cells and natural killer (NK) cells. LY‑303511 and HC‑toxin were predicted as candidate compounds. These five genes represent preliminary candidate biomarkers for PD; however, further validation in larger independent cohorts is warranted for clinical translation.