Methodology v0.4
How we grade what we publish
Every profile in the Peptides Pub encyclopedia is assigned an evidence tier and a risk band. This document explains what those mean and how we assign them.
Evidence scale
We use a four-tier scale based on the quality and quantity of human clinical data available for each peptide and each indication.
At least one well-powered randomised controlled trial (RCT) in humans, or a meta-analysis of multiple RCTs. Effect sizes are reproducible across sites and populations.
Multiple human trials exist, but results are mixed, sample sizes are modest, or methodological quality is inconsistent. Signals are plausible but not definitive.
Small pilot studies, case series, or a single human trial. Mechanistic plausibility may be high but large controlled trials are absent.
Evidence is confined to animal models, cell cultures, or in-vitro studies. No meaningful human data. Included for completeness where interest is high.
How the score is computed
Every evidence rating is built from two factors: the strength of the human clinical data behind a peptide, and the volume of primary research supporting it. Both are evaluated against the fixed criteria below and applied identically to every entry, so ratings remain consistent and comparable.
1 · The tier reflects the level of human clinical evidence. Each substance is placed on the four-tier scale above — Strong, Moderate, Limited, or Preliminary — by reviewing its published human-trial record: randomised controlled trials and meta-analyses, smaller or single human studies, or animal and in-vitro data alone. Popularity, marketing, and proposed mechanism do not raise a tier.
2 · The score places the substance within that tier's band. Within a tier, the rating reflects how strong and broad the human evidence is — a single small study sits near the floor of the band, an extensive programme of large randomised trials and meta-analyses near the ceiling:
- E
- the evidence rating (0–10) shown on the peptide's profile
- b
- the base — the lowest rating in that peptide's evidence tier
- c
- the ceiling — the highest rating that tier can reach
- q
- how strong the human evidence is within the tier, from 0 (only just meets it) to 1 (the strongest in the tier) — judged from the number, size, and quality of human studies
Tier bands
| Tier | b (base) | c (ceiling) |
|---|---|---|
| Strong | 8.0 | 9.8 |
| Moderate | 6.5 | 7.9 |
| Limited | 4.5 | 6.4 |
| Preliminary | 2.0 | 4.4 |
Stronger, broader human evidence — a higher q — lifts a rating toward c, but never past it. Because the tier, and therefore b and c, is fixed by the level of human evidence, a compound with no human trials cannot reach a Strong rating no matter how widely it is promoted. That is what keeps a popular but unproven peptide from being over-rated.
What "evidence strength" measures
- Best study design available — randomised controlled trials and meta-analyses, versus animal or in-vitro work; read from PubMed publication-type tags.
- Replication and scale — the number of independent human trials and their total enrolment, from ClinicalTrials.gov.
- The regulatory bar — whether a regulator (FDA / EMA) has approved it, an independent judgement that the evidence is sufficient.
These signals are read against the primary literature — a molecule studied widely as a disease target is not credited for trials of administering it as a drug — and every profile's Summary cites the specific studies behind its rating.
Risk bands
Risk bands reflect the reported adverse event (AE) profile, regulatory history, and potential for serious harm at typical research doses.
Well-characterised safety profile with minor, reversible adverse events reported. No serious signals in human populations at research doses.
Known adverse events that warrant monitoring. Dose-dependent risks. Not recommended without understanding the profile.
Serious adverse events reported or theoretically plausible. Meaningful uncertainty about long-term safety.
Withdrawn from clinical use, banned by regulatory agencies, or carries major safety signals. Included for reference; use is strongly discouraged.
Source selection
We draw exclusively from primary literature indexed in PubMed, ClinicalTrials.gov, and EMA/FDA approval documents. Secondary sources, community forums, and vendor-funded white papers are excluded. Every citation is linked to its DOI or registry entry.
Update cadence
Profiles are reviewed weekly. New publications are assessed within 14 days of indexing. Evidence tiers and risk bands are revised whenever new trials change the weight of evidence. All revisions are logged in the profile change history.