Fig. 3

Machine learning screening of key molecules. A Heat map visualizing the differential expression pattern of genes in RNA-seq data obtained from ovarian cancer (OC) cell lines and fibroblast cell lines sourced from the Cancer Cell Line Encyclopedia (CCLE) database. The heat map highlights the distinct gene expression profiles between OC and fibroblast cells. B Venn diagram demonstrating the intersection between the identified 229 differential genes and the previously established set of 510 key genes. C Forest plot presenting the results of univariate Cox regression analysis, which screened 20 genes derived from CAFs for their association with prognosis. The plot quantifies the hazard ratios and confidence intervals for each molecule. D Heat map illustrating the C-index values of 99 combined algorithms in each cohort. The rightmost column represents the average C-index values of the six cohorts. E Ranking of the importance of six molecules in random forest tree models. F Determination of the optimal cutoff value to classify the entire patient population into high-risk and low-risk groups. G Kaplan–Meier survival analysis of the RNA-seq cohort. H Kaplan–Meier survival analysis of the GPL570 cohort