NExON-Bayes: a Bayesian approach to network estimation informed by ordinal covariates.

NExON-Bayes improves network estimation in heterogeneous biomedical data by integrating ordinal covariates, achieving higher accuracy than existing methods in high-dimensional settings.

• Why it matters: Accurately modeling sample variability is essential for reliable interpretation of omic networks in diseases like cancer, where heterogeneity can mislead traditional analyses.
• What they did: The approach extends the graphical spike-and-slab framework with a variational inference algorithm, jointly estimating covariate relevance and network structure using data from breast carcinoma patients.
• The result: NExON-Bayes outperforms existing methods in simulations and reveals how proteomic networks and key protein group dependencies evolve across cancer stages, enhancing biological understanding.