Ubiquitination-anchored signature defines neuroendocrine prostate cancer: hub genes and single-cell ecosystem insights from integrated bioinformatics analysis of public transcriptomic datasets

BACKGROUND: Neuroendocrine prostate cancer (NEPC) is an aggressive, treatment-refractory state that often emerges under androgen-receptor pathway inhibition. We hypothesized that dysregulated ubiquitination underpins NEPC lineage plasticity and that integrating bulk and single-cell transcriptomes would define a ubiquitination-centered signature and tumor microenvironment (TME) circuits of diagnostic and therapeutic relevance. METHODS: Public bulk transcriptomic datasets were combined to compare NEPC with prostate adenocarcinoma, including a GEO discovery cohort of 49 tissue samples from GSE32967 and GSE104786, and intersected with a curated ubiquitination-related gene universe to derive ubiquitination-related differentially expressed genes (URDEGs). Co-expression networks, functional enrichment, and protein-protein interaction (PPI) network topology analyses were used to identify and prioritize candidate hub genes. An independent scRNA-seq cohort was integrated for biological contextualization to map cellular lineages, hub gene localization, and intercellular communication. RESULTS: We identified 317 URDEGs that clearly segregated NEPC from conventional prostate cancer and clustered into NEPC-associated modules enriched for cell-cycle and mitotic programs. Network integration yielded an 11-gene hub panel, including AURKA, CCNA2, EZH2, FGFR1, and TTK. In the single-cell dataset, 25,325 cells were retained after quality control, and UMAP resolved 17 clusters annotated into 8 major lineages. Single-cell mapping further revealed lineage-biased hub gene expression and highlighted a prominent stromal-vascular as well as immune MIF-CD74/CXCR4 signaling axis. CONCLUSIONS: This integrative analysis identifies a ubiquitination‑anchored transcriptomic signature and associated hub‑gene network that are consistently associated with NEPC in public bulk datasets and supported by cell‑type-resolved patterns in an independent scRNA‑seq cohort; these findings nominate candidate biomarkers and therapeutic hypotheses that warrant external validation in prospectively collected, clinically annotated cohorts with protein‑level assessment.