mCADRE method overview. (A) After binarizing context-specific input data, we quantify how often a gene is expressed across samples of the same tissue; this is the ubiquity score U(g) for each gene g. (B) From ubiquity scores, we calculate the expression-based evidence E
for each gene-associated reaction. Reactions with sufficiently high E
are defined as the core reaction set and are included in the tissue-specific model. To deterministically rank non-core reactions with low to moderate expression-based evidence, we introduce a network topology metric to calculate connectivity-based evidence Ec.