Immune perturbation network identifies an EMT subtype with chromosomal instability and tumor immune-desert microenvironment.

Most gastric cancer (GC) subtypes are identified through transcriptional profiling overlooking dynamic changes and interactions in gene expression. Based on the background network of global immune genes, we constructed sample-specific edge-perturbation matrices and identified four molecular network subtypes of GC (MNG). MNG-1 displayed the best prognosis and vigorous cell cycle activity.