Fig. 3

Results of WGCNA analysis in TCGA-PRAD. A According to the cutting height = 120,000, 495 samples were included, and the tree diagram of 265 samples was obtained. B We built scale-free co-expression networks. The best soft threshold was 5, R-squared was 0.9. C A hierarchical cluster tree was constructed, with each leaf representing a gene and each branch representing a co-expression module, and a total of eight co-expression modules were generated. D Correlation coefficients between different factors and co-expression modules, where the Brown module (COR = 0.6; P = 4E−41) and grey module (COR = 0.48; P = 1E−24) had a high correlation with TACR2. The correlation between Brown module membership and gene significance for TACR2 was COR = 0.69; P = 9.5e−74