Sharab A
Journals now require that Methods declare RRIDs for antibodies, authentication for cell lines, and animal-rigor reporting.
Core trick
Claude extracts typed resources from Methods prose, and deterministic gates check cell lines and antibodies against the ICLAC and Antibody Registry.
Description
Journals now require that Methods declare RRIDs for antibodies, authentication for cell lines, and animal-rigor reporting. Methods Verifier is the pre-submission check: paste your Methods + Key-Resources and get a per-resource verdict — PASS / FAIL / NEEDS-VERIFICATION — each cited to a public record. Claude does the two jobs a lookup can't. It extracts a typed resource inventory out of messy, hyphen-stripped prose, and it reasons about whether each antibody was validated with a genetic knockout control — the specificity gold standard. Deterministic gates then issue every identity verdict against ground truth a model cannot fabricate: cell lines against the ICLAC register, antibodies against the Antibody Registry, software against SciCrunch RRIDs. A missing datum → it abstains, cited. Every finding is tagged registry (deterministic) or model judgment (Claude), so the two are never blurred. Why it must exist: over the entire 594-line ICLAC register, a frontier model names a contaminated line from memory only 18% of the time — confidently wrong on 230. Give it the register and extraction catches 92%; on 47 real open-access papers, 89% with 0 false alarms on 14 clean controls, every FAIL cited to an ICLAC ID + CVCL.
README excerpt
# Methods Verifier — a pre-submission reproducibility check for your Methods **Built with Claude: Life Sciences · Builder track · July 7–13 2026** Repo: https://github.com/Moe1117/verdict · Demo video: https://youtu.be/HQW0DZOr84U · Live demo: https://magenta-tartufo-666742.netlify.app A bench scientist or PI pastes their manuscript's **Methods + Key-Resources** section and gets a per-resource report card — **PASS / FAIL / NEEDS-VERIFICATION**, each cited to a public record — before Reviewer 2, or a post-publication correction, finds the problem. Claude does the two things a lookup can't: it **extracts** a typed resource inventory out of messy reagent prose, and it **reasons** about whether an antibody was validated with a genetic control. Deterministic gates issue every *identity* verdict against ground truth a model cannot fabricate. - **Cell lines** → the **ICLAC Register of Misidentified Cell Lines** (594 lines, bundled offline; indexed under 672 name-spelling keys so a line's aliases all resolve). On the register → **FAIL**, cited with the ICLAC ID + CVCL + the line's *true* identity. - **Antibodies** → the **Antibody Registry** (catalog# → `RRID:AB_…`, live API). Catalog#/vendor must match, or it abstains — it never cites a different vendor's RRID. - **Software & tools** → the **SciCrunch registry** (`RRID:SCR_…`, live). A cited RRID that resolves names the exact tool (ImageJ, GraphPad Prism, FlowJo…) → **PASS**; a tool with no / an unresolvable RRID absta