Back to search

Methodology / v1

Uncertainty is part of the answer.

Gate Brief is a directional portfolio-screening tool. Every material conclusion keeps its source, date, evidence class, interpretation, and caveat attached.

Confidence describes evidence quality—not the probability of technical, legal, regulatory, or commercial success.

01

Identity before intelligence

A valid CAS checksum confirms syntax only. The public resolver uses PubChem, then asks the user to confirm the intended substance before product-level evidence is shown.

Parent compounds, salts, solvates, stereoisomers, mixtures, marketed products, and reference applications remain separate records. Ambiguity blocks formal sharing.

02

Evidence states

Verified

Multiple appropriate sources agree, or a qualified reviewer confirms the authoritative record.

Supported

One authoritative source or several credible secondary sources support the claim.

Inferred

brick42 derived the conclusion from disclosed evidence and a versioned rule.

Conflicting

Credible sources disagree or apply to different substance or product contexts.

Unknown

Evidence is insufficient; the product does not substitute a favorable zero.

Stale

The evidence exceeds the defined freshness threshold for its domain.

03

Claim boundaries

Containment

A published OEL/OEB is displayed only with a source and named banding scheme. Therapeutic class or molecular complexity never becomes an occupational limit.

Supply filing density

FDA Type II DMF records are a filing signal. They do not prove qualified supply, quality, capacity, an available letter of access, or the number of market competitors.

Patents and launch timing

Public events create scenarios—not FTO. European SPC validity and expiry require jurisdiction-specific verification, and process patents may not appear in product-linked lists.

CMO and NRE

Capacity evidence is dated and project-specific. NRE outputs remain ranges with assumptions until brick42 issues a scoped, human-reviewed proposal.

04

Recommendation guardrails

  1. Unknown evidence never becomes a favorable score.
  2. Unresolved identity blocks a recommendation.
  3. Critical conflict or multiple material unknowns force a hold or insufficient-evidence result.
  4. Every decision records the methodology and rubric version that produced it.

Primary public inputs

Sources are visible, dated, and challengeable.