Abstract
Prostate cancer, one of the most prevalent malignancies among men worldwide, is intricately linked with androgen signaling, a key driver of its pathogenesis and progression. Understanding the diverse expression patterns of androgen-responsive genes holds paramount importance in unraveling the biological intricacies of this disease and prognosticating patient outcomes. In this study, utilizing consensus clustering analysis based on the expression profiles of androgen-responsive genes, prostate cancer patients from the TCGA database were stratified into two distinct subtypes, denoted as C1 and C2. Notably, the C1 subtype demonstrates a significant upregulation of certain genes, such as CGA and HSD17B12, along with a shorter progression-free survival duration, indicating a potentially unfavorable prognosis. Further analyses elucidated the immune infiltration disparities, mutation landscapes, and gene functional pathways characteristic of each subtype. Through integrated bioinformatics approaches and machine learning techniques, key genes such as BIRC5, CENPA, and MMP11 were identified as potential therapeutic targets, providing novel insights into tailored treatment strategies. Additionally, single-cell transcriptome analysis shed light on the heterogeneous expression patterns of these genes across different cell types within the tumor microenvironment. Furthermore, virtual screening identified candidate drugs targeting the BIRC5 receptor, offering promising avenues for drug development. Collectively, these findings deepen our understanding of prostate cancer biology, paving the way for personalized therapeutic interventions and advancing the quest for more effective treatments in prostate cancer management.