Identification and Validation of Candidate Biomarkers Co-expressed with Creatine Metabolism-Related Genes in Ischemic Stroke Based on Transcriptomics Data.

Ischemic stroke (IS) is a major cause of death and disability, and creatine metabolism (CM) is potentially involved in its development. This study examined IS candidate biomarkers co-expressed with CM-related genes (CMRGs). IS data and CMRGs were sourced from public databases and published research. We identified candidate genes by intersecting genes from weighted co-expression network analysis and differential expression analysis. Candidate biomarkers were selected via machine learning. A diagnostic nomogram was constructed, and functional roles were explored via enrichment and immune infiltration analyses. Candidate biomarker expression was preliminarily validated by reverse transcription quantitative polymerase chain reaction in a small clinical cohort. Intersection analysis of the 336 differentially expressed genes and 3914 module genes yielded 111 candidate genes. Subsequently, F12 and PLXDC2 were identified as candidate biomarkers, and these genes were upregulated in IS samples. The nomogram based on these candidate biomarkers showed promising capacity for differentiating sample types in the training set, warranting further evaluation in independent cohorts. In addition to enrichment in pathways such as oxidative phosphorylation, VEGF-VEGFR2 signaling, and interleukin signaling, F12 and PLXDC2 were also strongly and positively correlated with neutrophils (r > 0.30, P < 0.001). Further screening highlighted cyclosporin A and trichostatin A as drugs that could simultaneously target both candidate biomarkers. This study identified F12 and PLXDC2 as candidate biomarkers co-expressed with CMRGs, offering preliminary insights that warrant further investigation in larger cohorts.