Fig. 1
From: Identification and validation of a refined CAF-Associated diagnostic signature in breast cancer
Feature selection by machine learning. (A) Integration of CAFs-associated gene lists from 3 studies with TCGA breast cancer DEGs revealed 28 candidate CAFs markers. (B) Among six evaluated feature selection methods, random forest emerged as the optimal choice, achieving high accuracy and stability; it also determined the optimal number of variables to be 3. (C) The Sankey diagram presents the optimal variables selected by different algorithms during the feature selection process. (D-E) Boxplots display expressions of feature genes in GSE65194 and GSE233242, respectively. Both datasets underwent PCA and tSNE analyses, effectively differentiating cancer from normal breast tissues.
