CBG Atlas · The Mother CannabinoidEvidence-graded · Provenance-tracked
Atlas/How we grade
Gap-MapResearch Library
How we know

How we grade the evidence.

CBG is among the most marketed and least studied cannabinoids. So every claim here carries a grade, and the grade can only rise when the evidence does. Where we hold no evidence, we say catalogue in progress rather than fill the gap. That is the whole promise: honesty is the authority.

The ladder

What do the six evidence grades mean?

A grade describes how strong the evidence is — never how promising something sounds. A claim moves up a rung only when new evidence earns it, and a retraction or a contradiction can move it down automatically.

EstablishedSettled science — replicated human data, or biochemistry that is not in dispute. e.g. CBGA is the precursor the plant branches into the other cannabinoid acids.
DemonstratedShown reproducibly in controlled experiments — typically laboratory (in-vitro) work. e.g. CBG's characterised receptor interactions in binding and functional assays.
EmergingAn early signal — preclinical (animal) work, or human evidence too limited to settle the question. e.g. An anti-inflammatory signal for CBG in a mouse model of colitis.
LowLimited and low-confidence — sparse, indirect, or single-source data. e.g. Overall human safety, resting on one small acute trial.
HypothesizedProposed and plausible — but not yet shown. e.g. The “entourage effect” as a synergy in people.
Not supportedTested, and the evidence does NOT support it. This is a finding against the claim — not an absence of evidence. e.g. A null result where a proposed effect was looked for and not found.
Figure F · how we grade
The six-rung evidence ladder
The six-rung evidence-grade ladder, strongest at top: Established, Demonstrated, Emerging, Low, Hypothesized, Not-supported.Each rung has a plain-language definition and a real example from the registry — from CBGA biosynthesis (Established) down to a mouse null result on fear memory, Zhou 2022 (Not-supported).STRONGERWEAKER · UNPROVENEstablishedSettled science — replicated human data, or biochemistry that is not in dispute.e.g. CBGA → the acids · Taura 1996DemonstratedShown reproducibly in controlled experiments — typically laboratory (in-vitro) work.e.g. CBG receptor targets · Cascio 2010EmergingAn early signal — preclinical (animal) work, or human evidence too limited to settle the question.e.g. IBD, mouse · Borrelli 2013LowLimited and low-confidence — sparse, indirect, or single-source data.e.g. overall safety · one acute human trialHypothesizedProposed and plausible — but not yet shown.e.g. the entourage effect · synergy in peopleNot-supportedTested, and the evidence does NOT support it. This is a finding against the claim — not an absence of evidence.e.g. fear-memory, mouse null · Zhou 2022
Every claim on the platform wears one of these six grades — and it can only go up a rung when the evidence does. The ladder is the promise: we tell you exactly how strong each thing is, from settled biochemistry to a result the data contradicts.
Not everything is a rung

Two states that are not grades.

Some things we publish are true but are not claims about what a cannabinoid does in people, and some things we simply do not have. Both get their own label, because filing them on the ladder would misrepresent them.

Validated in vitro — a mechanism, not an outcome

A receptor interaction measured in a dish is a real, reproducible finding. It describes how a molecule behaves at a target — it is not evidence that anything happens in a person. We label it as mechanism and stop there.

No primary evidence — catalogue in progress

Our registry holds no study either way. This is not the same as Not-supported: “Not-supported” means the evidence weighs against a claim; “no primary evidence” means we have none. Restating the second as the first would be claiming a negative finding we do not have, so we never do. Where a compound’s catalogue is thin, the page says so plainly instead of borrowing a better-studied compound’s data.

Figure G · human vs. preclinical
Where CBG’s evidence actually sits
An evidence pyramid.The base — preclinical in-vitro and animal studies — holds most of CBG's evidence. Above it, observational human evidence is roughly none. Human RCTs contain a single small acute trial, Cuttler 2024. The apex — systematic reviews and meta-analyses of human trials — is empty for CBG. Most CBG evidence is preclinical.HIGHER-QUALITY ↑MORE STUDIES ↓META-ANALYSESHUMAN RCTsOBSERVATIONALPRECLINICALin vitro & animalSystematic reviews / meta-analysesnone yet for CBGHuman RCTsone small acute trial (Cuttler 2024)Observational human≈ nonePreclinical — in vitro & animalmost CBG evidenceCascio · De Petrocellis · Borrelli · Farha · Valdeolivas…Most CBG evidence is preclinical — that’s the honest map, and it’s why every claim here is graded and hedged.
A mouse or a cell line is a lead, not a conclusion. CBG’s base is full and its apex is empty — so we present it that way, rather than promoting a preclinical result into a human promise. Cuttler 2024 · 10.1038/s41598-024-66879-0 · Zhou 2022 · PMID 34182770
Why you can trust a grade

How a claim earns — and keeps — its grade.

Grades are not editorial opinion. Each one is produced by a gated pipeline and pinned to a verifiable snapshot.

  • A gated publish. Nothing ships unless every registry gate passes — including a concordance gate that refuses to let an authored grade sit above the ceiling its evidence supports. A red gate blocks the publish outright.
  • Hash-chained provenance. The evidence set carries a registryHash and an append-only audit chain, so what you read traces to an exact, tamper-evident snapshot rather than to “our latest draft”.
  • Bound citations. Every graded claim links to a real registry study. We never invent a citation; where no qualifying study exists, the claim is marked catalogue in progress instead of guessed.
  • Only up on evidence. A grade rises only when the evidence does, and a retraction or a peer-or-stronger contradiction can lower it automatically.
Independence & limits

What we do not claim.

We report, we do not affirm

Statements about effects, benefits, or uses reflect claims made by third parties — manufacturers, media, traditional use, or preliminary research. CBG Atlas does not make, adopt, endorse, or affirm them, and does not present them as fact. Nothing here is medical advice, and nothing here is intended to diagnose, treat, cure, or prevent any disease. Binding constants and other researcher-tier detail are kept off consumer pages by design. Report a safety event.

Funding. Peregrine Biopharma funds the platform and holds no editorial authority. Grades, citations, and the gates that enforce them are not reviewable by the funder. Where the evidence is thin, that is what the page will say — regardless of what it would be convenient to claim.

This page describes the method; the Research Library is the evidence itself, and the Gap-Map is the honest picture of where it runs out.