Fig. 3: Analysis and validation of core genes in the Test Dataset.

A Heat map showing the expression of six core genes in different clinical stages and tumor statuses. B Circos plot depicting the expression correlations among the six core genes. C Data correction plot for the test dataset (Control, N = 6; OC, N = 392) after integrating the TCGA-OV dataset (OC, N = 376) and GSE54388 dataset (Control, N = 6; OC, N = 16), with the upper and lower graphs illustrating data distribution before and after correction. D Volcano plot of 774 DEGs between normal and OC samples, where red dots represent significantly upregulated genes, blue dots signify significantly downregulated genes, and black dots indicate genes with no expression differences. E Visualization of the neural network model, with the input layer on the left consisting of core genes connected to the middle hidden layers (H1 to H5) for feature extraction and non-linear transformations, further linked to the right output layer for making final disease status predictions. F ROC curve analysis is based on a deep learning model to evaluate the predictive performance of the test group model. G Comparison of the expression levels of four core genes in the test dataset between the normal and OC groups. H ROC curve analysis of the four core genes in the test dataset. Between-group comparisons were evaluated, where *P < 0.05 and ***P < 0.001.