Explainer
What a DSS colitis mouse is actually a model of
Dextran sulfate sodium damages the epithelial barrier chemically, on a schedule the experimenter sets. That is useful, and it is not inflammatory bowel disease.
Littermate
What the animal literature actually says.
An independent review of preclinical peptide science.
The lead · Methods
One lab, many tendons: the replication problem under the BPC-157 literature
Most of what is published about this pentadecapeptide in rat tendon and ligament models traces back to a small number of connected groups. That is not evidence of misconduct. It is a structural fact about the literature, and it changes how much any single reported effect is worth.
Reference
The Library
Eight peptides that have been studied in rodents and in cell culture, each with a standing profile: what it is, what the animal work reports, and where the evidence runs out.
About
What this publication is
Editorial standards
SpeciesNamed in every claim
Evidence gapsStated, never smoothed over
Dosing and sourcingNever published
CorrectionsRun in public, appended
FundingNo advertising, no sponsorship
LITTERMATE reads published preclinical papers about eight peptides and writes up what they found, in which animal, under which conditions. Most of that literature is rodent work and cell culture. A result in a Sprague-Dawley rat is a result about Sprague-Dawley rats; whether it holds in a person is a separate question that almost none of these papers were designed to answer, and we say so every time rather than once.
We do not sell anything, we take no advertising and no sponsorship, and we have no arrangement with anyone who supplies these compounds. We publish nothing about dosing, sourcing, preparation, or use in humans. What we publish instead is the shape of the evidence: how many animals, whether the assessor was blinded, whether a second laboratory has ever reproduced the result, and which endpoint the authors chose to report. When a compound has essentially no rigorous human evidence, that sentence appears in its profile in those words. Corrections run in public and stay attached to the piece.
A mouse that heals is a fact about a mouse. Everything after that is an argument, and arguments have to be made out loud.
House method
How we read a paper
Sample size
We record how many animals were in each arm, because most of this literature runs on group sizes of six to twelve. Small groups can find large effects and routinely miss moderate ones, so a null result in an underpowered study tells you almost nothing.
Blinding and randomization
We check whether animals were allocated at random and whether whoever scored the outcome knew which group they were scoring. Histological and behavioural endpoints are graded by a human being, and unblinded grading of a subjective endpoint inflates effects in a predictable direction.
Independent replication
We ask whether any group unconnected to the original authors has run the same experiment and reported the same direction of effect. A dozen papers from one laboratory and its collaborators is one body of work, not a dozen confirmations.
Endpoint selection
We note which outcome was declared in advance and which appears only in the results. When a paper measures six things and discusses the two that moved, the honest reading of that paper is narrower than its abstract.
By email
New dispatches, roughly twice a month
One piece of reporting or criticism, plus a note on anything in the Library that changed since the last issue. Nothing else.