There is a newer version of the record available.

Published August 22, 2026 | Version 2.0.3

NeuroMetBench 2.0.3 derived-results verification and provenance package

Authors/Creators

  • 1. Independent Researcher, Amman, Jordan

Description

Derived-results verification and provenance package for the NeuroMetBench retrospective case study. Frozen biological endpoints, cohorts, thresholds, statistical results, and adverse results are unchanged from v2.0.0. Version 2.0.3 binds the exact article source and Online Resource 2, corrects LSCC cohort-flow semantics, adds a prespecified-versus-retrospective analysis boundary, exposes current prior-art boundaries, adds deterministic BCa sensitivity verification, and preserves 23 byte-identified recovered historical stage packages or successful execution receipts. The chain covers method identity, comparator execution, public PDC/GDC acquisition, specificity, transport, missingness, multiplicity, simulation, ablation, numerical compatibility, CNS analysis, and the final scientific result lock. Raw third-party matrices are not redistributed. Contemporary raw-to-final execution requires source reacquisition, reconstruction of recorded platform-specific environments, and applicable external licenses. The reusable neurometbench-ref software component remains version 1.0.1.

Notes

Version 2.0.3 is a scientific-integrity, provenance, and reproducibility correction. It does not change any frozen biological endpoint, cohort, threshold, statistical result, adverse result, non-admission, or claim boundary from version 2.0.0. It adds the recovered historical computation chain while retaining all raw-data, environment, and license limitations.

Files

reproducibility_release.zip

Files (144.5 MB)

Name Size
md5:619e1ea70416c06f9e68a4e5cdb0cf05
144.5 MB Preview Download

Additional details

Related works

Is supplement to
Software: https://github.com/build33Mercury/NeuroMetBench (URL)