Figure 4
From: Bayesian probit regression model for the diagnosis of pulmonary fibrosis: proof-of-principle
The training sets. All IPF (brown), IPF biopsy (red), IPF explant (green) and normal lung (black). (A) Unsupervised hierarchical clustering of 11 IPF samples (6 biopsies and 5 explants) and 6 normals. (B) Unsupervised hierarchical clustering of 6 IPF biopsy samples and 6 normals. (C) Unsupervised hierarchical clustering of 5 IPF explants samples and 6 normals. (D) 11 IPF samples (biopsy and explant) and 6 normals are plotted according to expression of the first two Principal Components. The left panel shows the difference between IPF and normal lung; while the right panel reveals the difference between IPF biopsy and IPF explant. (E) 6 IPF biopsies and 6 normals are plotted according to expression of the first two Principal Components. (F) 5 IPF explants and 6 normals are plotted according to their expression of the first two Principal Components.