Not all evidence carries the same weight. Where a finding sits on the evidence ladder decides how much it supports, and why a result in animals does not automatically say anything about humans.
MVMichel van der VeenRegistered Nurse · Science Editor, Peptalis Sources checked via PubMed · 4 referencesReviewed 15 JULY 2026 · 8 MIN READResearch on a peptide rarely yields a single kind of evidence. It yields a stack of findings sitting at different levels: a test tube, an animal model, a small human study. Those levels are not interchangeable. Knowing where a result sits on the ladder is the core of honest reading, and precisely where Peptalis draws the distinction rather than folding everything into one label.
The evidence ladder: in-vitro, animal, human
The evidence ladder orders research by increasing relevance to the human situation. At the bottom is in-vitro work: experiments in cells or isolated systems, outside a living organism. It shows that a mechanism can in principle exist under controlled conditions. It is valuable as a first indication, but far removed from an intact body.
A rung higher are animal studies, in which a substance is examined in a living organism with all the complexity that entails: absorption, distribution, breakdown. At the top is human research, the only kind that speaks directly about humans. Each rung adds realism and at the same time loses controllability. A finding lower on the ladder is not worthless; it simply carries less far than a finding higher up. This is also the logic behind the dot classification Peptalis shows per compound: it weighs how much is known and at what level, not how desirable an outcome would be.
What dose-response shows, and what a small n means
Within a study, the dose-response relationship is one of the strongest indications that an effect is real. It describes how the response changes as the dose changes. If the response rises systematically with the dose and levels off at higher doses, it often follows a sigmoidal curve described by the Emax model or the Hill equation (Lees et al., 2004). Such an ordered relationship is harder to attribute to chance than a single, isolated result.
Just as important is the size of the study. The number of animals or participants, the n, determines how reliable a finding is. Studies with a small n have a lower chance of detecting a true effect and, less widely appreciated, a greater chance that a detected effect is overestimated. An analysis of the neurosciences showed that average statistical power there is low, with overestimated effect sizes and lower reproducibility as a result (Button et al., 2013). A spectacular result from a small study is therefore not a strong result; it is a preliminary result that needs replication.
Why species differences limit translation
The jump from animal to human is the hardest on the ladder, and historically the most disappointing. In biomedical development a large share of the substances that looked promising in animals still fail in humans; a narrative review places that failure rate at more than 92%, largely due to unexpected toxicity or lack of efficacy (Marshall et al., 2023). That is not a detail. It is the pattern.
The cause lies in the differences between species. An animal model usually reproduces only part of the human situation. In Alzheimer's disease, for example, commonly used transgenic mouse models mirror only part of the human pathology, and premature translation of successes in those models to humans contributes to the high number of failed trials (Drummond & Wisniewski, 2016). Species also differ pharmacokinetically: absorption and breakdown do not scale simply with body weight, so a dose that works in one species does not translate straightforwardly to another (Lees et al., 2004). An effect in animals is therefore a hypothesis about humans: a reason for human research, not a replacement for it.
How Peptalis weighs this, and stays honest about what has not yet been shown
Peptalis translates this evidence ladder into a visible classification per compound. The dot classification indicates the level the evidence sits at: from early, exploratory work to research in which multiple independent human studies exist. The accompanying evidence meters (for animal studies, human trials and mechanism clarity) are assigned conservatively: when in doubt, the lower estimate. In this way the platform shows how much is known, not how desirable an outcome would be.
That honesty works in both directions. It names what has been shown, and it names just as clearly what is still missing: where the evidence is limited to small cohorts, to a single research group, or to animal studies without human confirmation. Precisely for patent-free, under-researched peptides that gap is not a suspicious silence but a description of where the work lies: room for randomised studies with larger cohorts, for independent replication, for human confirmation of a mechanism so far seen only in animals or cells. Anyone who reads the ladder as it is meant sees what a study shows and exactly how far that statement reaches.
About the authorMVMichel van der VeenRegistered Nurse · Science Editor, PeptalisRegistered nurse with eleven years in psychiatry and founder of Peptalis. Writes the platform's knowledge layer: compound profiles, evidence reviews and the quality methodology. Works from primary literature (PubChem for chemistry, PubMed for studies) and states where evidence is absent.
Sources checked via PubMed · 4 references
Compound profilesEvidence reviewsQualityIN THIS ARTICLE- The evidence ladder: in-vitro, animal, human
- What dose-response shows, and what a small n means
- Why species differences limit translation
- How Peptalis weighs this, and stays honest about what has not yet been shown
- BasicsReading the evidence: the dots and the levels
- BasicsWhy most peptide research is still preclinical
- In DepthWhat a Certificate of Analysis does and does not say
Compounds in this article
- Epitalon · In stock · Batch PEP-EPI-2026-01 · HPLC 99.11% · released 2 Jun 2026 · Research profile · View in catalogue
For laboratory research use only. Not for human use.
SCIENTIFIC REFERENCES
These references are provided for informational and research purposes only. They do not constitute medical advice.
- 01ReviewPoor Translatability of Biomedical Research Using Animals — A Narrative Review.Marshall LJ, Bailey J, Cassotta M, Herrmann K, Pistollato F. Altern Lab Anim. 2023;51(2):102–135.Narrative review documenting the high failure rate (reported at more than 92%) of substances that looked promising in animal studies but did not translate to humans, mostly due to unexpected toxicity or lack of efficacy.View on PubMed →
- 02ReviewAlzheimer's disease: experimental models and reality.Drummond E, Wisniewski T. Acta Neuropathol. 2016;133(2):155–175.Review showing that commonly used transgenic mouse models of Alzheimer's disease mirror only part of the human pathology, and that premature translation of successes in those models contributes to failed clinical trials.View on PubMed →
- 03Review / MethodologyPower failure: why small sample size undermines the reliability of neuroscience.Button KS, Ioannidis JPA, Mokrysz C, Nosek BA, Flint J, Robinson ESJ, Munafò MR. Nat Rev Neurosci. 2013;14(5):365–376.Analysis of the neurosciences showing that average statistical power is low, so real effects are detected less often and detected effects tend to be overestimated, with lower reproducibility as a consequence.View on PubMed →
- 04Method / ReviewPK-PD integration and PK-PD modelling of NSAIDs.Lees P, Giraudel J, Landoni MF, Toutain PL. J Vet Pharmacol Ther. 2004;27(6):491–502.Methods review of dose-response and PK-PD modelling, including the Emax model and the Hill equation as descriptions of how a response changes with dose, and why pharmacokinetic differences between species limit direct translation.View on PubMed →