Liver hepatocellular carcinoma — ranked gene evidence

This is a reproducible evidence graph, not just a gene list: each column is one FFL facet computing one score from one open public dataset, and every score traces back to it. Over-expression = TCGA TCGA-LIHC primary tumors (371 samples) vs TCGA matched solid-tissue normal (50 samples), GDC STAR-Counts TPM — same pipeline, no batch effect. Tumor-specificity compares tumor TPM to GTEx healthy Liver median — this is a cross-cohort signal (different pipeline) shown as corroboration, not the primary fold-change. Survival = log-rank of high vs low expressors against GDC clinical OS (↑worse = high expression predicts worse survival). Mutation freq = fraction of mutated cases (GDC open SSM). Known driver = intOGen Compendium of Cancer Genes (CC0). Composite = weighted z-combination (overexpression=1.0, specificity=1.0, survival=1.0, mutation=0.8, driver=1.2). Candidates are filtered to genes expressed in tumor, over-expressed, and significant. Generated 2026-06-30 by an FFL workflow on Facetwork.