amirabozaid68/ferroptosis-luminal-breast-cancer: v1.1.0 — code and result tables accompanying the submitted manuscript

Analysis code, deposited result tables and supporting analyses for the manuscript "GCH1 predicts survival within luminal A but not luminal B breast cancer" (GSE96058 discovery, METABRIC replication), as submitted.

Added since v1.0.1

• results_tables/ — T01 to T24, every table behind a reported number, so a reader can check any value without running the pipeline.
• revision/ and revision_survival_checks.Rmd — cause-of-death analyses within luminal A: cause-specific hazards with the Lunn-McNeil test, Fine-Gray fitted within luminal A rather than across both subtypes, age correlation, overall survival under 70, and relapse-free survival. The plan is in PRESPECIFIED_revision_analyses.md, committed before the models were fitted.
• 04_coupling_prior_TCGA_GTEx.Rmd and coupling_prior_output/ — a post hoc test of the GCH1 and SLC40A1 correlation in GTEx normal breast, TCGA adjacent normal and TCGA luminal tumours. The result is negative and is deposited because the manuscript Limitations cites it. Not preregistered.

Changed

• Figure numbering throughout 03_analysis_lean.Rmd: the covariate figure is now Figure 1, the survival figure is now main Figure 6 with an added hazard-ratio forest panel, and the replication scatter and the hepcidin panel are Supplementary Figures 1 and 4.
• Figure 3 now shows all nine concordant genes rather than the first seven.
• Numbers at risk in the survival figure are taken from the survfit object rather than counted directly, correcting a misaligned at-risk table.
• Effect-size axes read "LumB relative to LumA" throughout.
• README rewritten: deposited output is documented separately from the paths the code writes to locally, and the two post hoc analyses are explained.

Not included

The two expression datasets. GSE96058 comes from GEO and METABRIC from cBioPortal; see Obtaining the raw data in the README.