Explained
What kind of study is behind each entry?
Built from index records dated 2026-10-13 Compiled by an automated pipeline, checked by The Peptide Observer editorial desk How this site works
Each entry carries a short label worked out by counting the kinds of record behind it, and that label describes what this index holds rather than saying whether anything works.
The label is derived by counting. It is not a grade, not a score, and not anyone's assessment.
What we actually know
Across the entries this site describes, the labels fall out like this. Each bar is a count of entries, not a ranking of them:
| Label, in plain words | Entries | What was counted, for one of them |
|---|---|---|
| The records kept here are animal and laboratory studies only. | 21 | no interventional human trial record; 4 rodent and 0 in vitro records among 11 |
| No public chemistry record was found under this name, so the entry describes a preparation, a protein or a mixture rather than one defined molecule. | 5 | PubChem returned no compound record for this name, so the subject is a preparation, protein or mixture rather than a defined small molecule; identity fields are unsourced |
| None of the records kept here state what kind of study they are. | 1 | 12 indexed records, none of whose assigned subject headings state a study model |
A label is derived metadata. It is not a judgement on whether a compound does anything, not a measure of study quality, and not a summary of expert opinion. It is a description of what this index happens to hold.
Every label on this site is printed together with the sentence explaining exactly what was counted to produce it. Where the two disagree in a reader's mind, the counted sentence is the one to believe.
What is not known
- Whether a study was well conducted. No study behind any label here has been appraised, and the label would be identical if every one of them were poor.
- Whether the records agree with one another. A separate check reads record titles only, looking for wording that suggests an absent or negative result; it cannot establish disagreement and does not try to.
- Whether a label would survive someone reading the underlying papers. Nobody has.
How to read this
- Read the counted sentence under a label before reading the label.
- A label describing animal and laboratory records is a statement about where the research has got to, not a verdict on the compound.
Read next
Where this comes from
Labels are derived at build time from the study kinds counted on each entry's evidence profile. The derivation sentence is carried with every label.