Skip to content

Releases: yangfeichang2012/hcc-senescence-construct-audit

v1.1.0 - corrected full-chain reproducibility release

Choose a tag to compare

@yangfeichang2012 yangfeichang2012 released this 30 Jul 16:41

What changed

  • Adds the complete GEO accession ? processed expression ? frozen score route for 15 public input files, 10 data sets, and 92 biological samples.
  • Deposits the missing frozen manifests, sample-level derived source data, preprocessing parameters, fold coefficients, and clean-run evidence.
  • Reproduces 60/60 v1.2 scores, 100/100 v1.3 scores, and 816/816 estimable v2 module scores; all 12 frozen non-estimable states are retained.
  • Repairs Fig 1 text overflow and corrects the Fig 4/S1 production dimensions.
  • Restricts the cross-run fixed hash registry to six substantive numerical outputs; timestamped run summaries remain auditable but intentionally have run-specific hashes.

No model was refit or retuned, no threshold or test was changed, and no scientific conclusion changed. The locked conclusion remains that directional responsiveness was observed but HCC-senescence specificity was not established.

Tagged commit: 5bbe6a207ed01a230a1c86adbddd4de6dd277361

Audited standalone archive SHA-256: 8cce9c2ffda0cf4203a7239d6f6e75f51647ea087f5d014a320594a4fa57eecf

HCC senescence construct audit v1.0.1

Choose a tag to compare

@yangfeichang2012 yangfeichang2012 released this 29 Jul 16:52

Zenodo archival trigger release. This tag points to the same commit as v1.0.0 and introduces no scientific, code, data, or documentation changes. v1.0.0 is retained as transparent release history.

HCC senescence construct audit v1.0.0

Choose a tag to compare

@yangfeichang2012 yangfeichang2012 released this 29 Jul 16:43

Initial manuscript-linked code archive.

Includes:

  • frozen Python and R analysis code;
  • prespecified protocols and contracts;
  • aggregate v1.2, v1.3, v2, and integrated evidence;
  • figure source tables and SVG/PNG/PDF assets;
  • complete SHA-256 file manifest.

Expression matrices, external model weights, sample-level scores, personal information, and submission files are intentionally excluded.

Scientific boundary: transcriptomic responsiveness and internal directional transport were supported, but specificity against hypoxia, damage, apoptosis, and hepatic-lineage states was not established; the developmental model was not promoted.