The scenario
A researcher needs a reproducible pipeline to derive a prognostic multigene signature for hepatocellular carcinoma using public transcriptomic cohorts.
What is in this drawing
Read the decisions behind it.
01
Use TCGA as the discovery cohort and GEO datasets for independent validation
02
Apply univariate Cox screening before LASSO to reduce dimensionality
03
Use Kaplan-Meier separation and independent prognostic evaluation as validation gates
Reusable for similar cancer gene signature discovery workflows where public expression and survival data are available.
More like this