Manuscripts, figures and review objects—separate from the research workflow.
Publication preparation can inspect and cite the Evidence Archive without interrupting a researcher who is framing a disease, cell-state or perturbation question.
Each packet retains its own status and evidence ceiling.
Cards are resolved from the same package authority that generates the v2 registry. “Current”, “working”, “frozen” and “superseded” are explicit identities, not filename conventions.
Falsification-first evaluation separates failure ranking from utility in perturbation transfer
Current Science Advances manuscript identity. Public scope is status only because the chemical replay produced no target score.
Object
PACKAGE:SCIENCE_ADVANCES:P59
Package state
HOLD PRIMARY SCORER FAILURE
Claim ceiling
The chemical replay produced no target score. The retained scorer failure is an engineering outcome, not evidence for or against perturbation-transfer performance or DMD efficacy.