Identifying critical States of complex diseases by local network Wasserstein distance

Complex diseases often undergo abrupt transitions from pre-disease to disease states, with the pre-disease state is typically unstable but potentially reversible through timely intervention. Detecting these critical transitions is crucial. We propose a model-free method, Local Network Wasserstein Distance (LNWD), for identifying critical transitions/pre-disease states in complex diseases using single sample analysis. LNWD measures statistical perturbations in normal samples caused by diseased samples using the Wasserstein distance, and identifies critical states by observing LNWD score changes. Applied to KIRP, KIRC, LUAD, ESCA (TCGA datasets) and GSE2565, GSE13268 (GEO datasets), the method successfully identified critical states in six disease datasets. This single-sample, local network-based approach provides early warning signals for medical diagnosis and holds great potential for personalized disease diagnosis.