What a Murine Colitis Model Can and Cannot Tell You
Dextran sulphate sodium injures a mouse colon chemically, on a schedule, in an animal bred for uniformity. Human inflammatory bowel disease does none of those things.
Independent · Ann Arbor, MI · Est. 2021
The Control Group
What the baseline actually shows.
An independent review of preclinical peptide research.
Lead feature · Evidence review
One Laboratory, Thirty Years, and the Tendon Literature That Rests On It
Nearly everything published about BPC-157 and rat tendon healing traces back to a small number of affiliated investigators. That is not fraud. It is a structural problem, and it is the reason the field cannot yet say what it thinks it says.
Preclinical literature summary. Not medical advice.
The Library
Eight compounds, eight standing summaries
Each entry describes the published animal and cell-culture work, then says plainly where that work runs out.
- BPC-157 Gastric pentadecapeptide · Rat
- Thymosin β4 / TB-500 fragment Actin-sequestering · Mouse
- GHK-Cu Copper tripeptide · Rodent, cell culture
- KPV Tripeptide, MSH-derived · Mouse
- LL-37 Cathelicidin host-defence · Mouse
- Thymosin α1 Immunomodulatory · Rodent, cell culture
- MOTS-c Mitochondrial-derived · Mouse
- Humanin Mitochondrial-derived · Rodent
What this publication is
We read published preclinical papers about eight peptides and write up what they report. The subjects are rats, mice, and cells in dishes. We do not run experiments, we do not have a laboratory, and we have no stake in any result going one way or the other. When a finding is strong we say so; when it rests on one research group, twelve animals, and an unblinded histology score, we say that instead.
Our editorial standards are short enough to state here. We name the species and the model every time a result is described. We say out loud when a finding has never been independently replicated, and when no registered human trial exists we write that sentence rather than leaving a gap the reader will fill in optimistically. We publish no dosing, no preparation, no sourcing, and no supplier information of any kind. We accept no advertising, sponsorship, or affiliate arrangement. Corrections are appended to the article in public, dated, with the original text intact.
A mouse that heals is a fact about a mouse. Everything after that is inference — and inference is where this literature keeps losing its footing.
The Control Group editorial desk
Method
How we read a paper
Sample size
Most rodent work in this area runs six to twelve animals per arm, which is enough to detect a very large effect and almost nothing else. A small study that finds nothing has usually not shown absence; it has shown that the study could not have seen it either way.
Blinding and randomisation
We look for who assigned animals to groups and who scored the outcome, because histology grading and behavioural scoring are judgement calls made by humans who often know which cage is which. Papers that omit both details are not necessarily wrong, but they cannot rule out the most ordinary explanation.
Independent replication
A result reproduced by an unaffiliated laboratory, with different animals and different hands, is worth more than ten papers from the same corridor. Much of the peptide literature has never cleared this bar, and we mark those entries explicitly rather than counting citations as confirmation.
Endpoint selection
We ask what was actually measured, and whether the measure was chosen before the data arrived or after. Heterogeneous endpoints across a literature — tensile strength here, a histology index there, a cytokine panel somewhere else — make pooled claims about a compound very difficult to justify.
The dispatch
New dispatches, roughly twice a month.
One email when we publish. Summaries of preclinical literature, corrections, and the occasional note on how a paper was put together. No advertising, ever.
On submit you will see a confirmation notice and a single verification email. Unsubscribe from any issue; we keep the address and nothing else.