Crotonylation-Associated Transcriptional States Mark a Fibroblast-Enriched Tumor Ecosystem in Ovarian Cancer: Single-Cell, Spatial, and Perturbation-Informed Evidence

Background: Lysine crotonylation (Kcr) is a metabolically responsive modification, but the cellular ecosystems associated with crotonylation-related transcripts in ovarian cancer remain incompletely characterized. Methods: Three ovarian cancer single-cell cohorts were integrated to define a crotonylation-related transcriptional score (CRTS). We examined cell composition, malignant-cell programs, and inferred ligand–receptor communication and evaluated a 12-gene survival-associated readout across bulk cohorts. Spatial transcriptomics, SHAP, cancer dependency mapping, and virtual gene knockout were used for exploratory triangulation. Results: High-CRTS cells included relatively more fibroblasts and malignant cells and fewer CD8+ T cells than low-CRTS cells. Among malignant cells, EMT, invasion, and metastasis module scores were higher (all nominal cell-level p < 0.0001), and CRTS correlated with invasion (rho = 0.269) and metastasis (rho = 0.252). CellChat inferred 1915 versus 519 interactions, with stronger ECM-related information flow in the high-CRTS group. A 12-gene model stratified survival across the evaluated patient cohorts and was associated with higher predicted immune exclusion. WTAP, SIRT2, and ALDH5A1 were prioritized by model interpretation and in silico perturbation. Nine model genes correlated with inferred fibroblast abundance in one spatial section (rho approximately 0.37–0.62; reported nominal spot-level p < 0.05). Conclusions: CRTS marks a fibroblast-enriched, ECM-active ovarian cancer transcriptional ecosystem associated with invasive malignant-cell programs and predicted immune exclusion. The results support testable hypotheses, but they do not establish direct Kcr modification, patient-level ecosystem remodeling, or molecular causality.