The conditional mutual information-based single-sample network biomarker approach reveals the critical transition moments of disease progression.

During the progression of complex diseases, it is common to observe that the deterioration of the condition does not follow a smooth trajectory, and there is often a critical transition from one state to another. Finding these critical transitions is of significant importance in the clinical treatment of cancer. In this study, we propose a novel computational approach, called the conditional mutual information-based single-sample network biomarker (CMISNB), which can reveal the critical transition moments of disease progression using only a single sample (https://github.com/ZLTSKY/CMISNB). By analyzing disease data from mouse acute lung injury, colon cancer, hepatocellular liver cancer, lung adenocarcinoma, and endometrial cancer, we validated the effectiveness of the CMISNB method during identifying tipping points in disease development. In particular, the CMISNB approach helps identify new markers that can predict patient outcomes and can provide personalized diagnoses for individuals.