Figure 1.

Workflow of the PSF-SOM method as applied to lung diseases. Expression data were analyzed in terms of pathway signal flows (PSF) in a series of KEGG canonical pathways. The PSF values of selected sink nodes in the pathways were then clustered using SOM machine learning. This method provides individual “portraits” of sink-node activities of each disease class. Similarities between them were studied using different clustering methods. For details please refer to the Methods section of the manuscript.