The single-lab dominance problem

Open a PubMed page on a research peptide, scroll the results, and ask a simple question: how many independent research groups have published on this molecule? The answer is often much smaller than the paper count suggests. Fifty papers split across two authors working at the same institution for three decades is not fifty independent results — it is one research program, reported fifty times.

Science earns its credibility through replication. A finding that only one group has ever observed is a hypothesis with an audience, not a settled fact. That is why single-lab dominance is a yellow flag — not because the work is wrong, but because nobody outside the originating lab has yet stood up and said "we saw it too, with our own hands, our own reagents, our own animals."

A compound shows single-lab dominance when one research group — usually one principal investigator and their trainees, sometimes an extended family of two or three collaborating labs at the same institution — accounts for the majority of primary-literature records on the compound. The threshold is not bright-line; a reasonable working definition is more than 60 percent of indexed publications share a first or senior author from a single institution.

The pattern is distinct from being "cited by one group." Review articles, textbook chapters, and derivative commentaries can multiply across the literature without ever adding a new data point. A peptide with two hundred citations and three original laboratories behind them all is still three laboratories.

Three things tend to happen when a compound lives inside one lab for a long time:

• Protocol drift becomes invisible. Animal strain, solvent, vehicle, route of administration, and assay conditions become house standards. External reviewers accept them because they match earlier papers from the same group — which is circular.
• Negative results stay unpublished. A single group has no structural incentive to publish experiments that failed to replicate its own earlier positive findings.
• The mechanism story stabilizes before it has been stress-tested. The first plausible mechanism proposed becomes the reference mechanism, and later papers extend it rather than testing alternatives.

None of this makes the underlying data wrong. It means the uncertainty on the data is larger than it would be if three unrelated groups had independently repeated the work.

Replication is not a formality; it is how science distinguishes real effects from artifacts of a particular bench. The replication-crisis literature in psychology DOI10.1126/science.aac4716 and in cancer biology DOI10.1038/483531a showed that an alarming fraction of published findings could not be reproduced by independent teams — and that the findings most likely to fail replication were those which had never been replicated in the first place.

The rodent-behavior and tissue-repair literature has its own version of this problem. Effect sizes in rodent studies are consistently larger in the originating lab than in independent replication attempts; this is the reason the NIH rewrote its preclinical rigor guidance in 2014 to require stricter randomization, blinding, and cross-lab validation before advancing compounds into Phase 1.

A peptide with striking effects that only one group has ever observed is not disqualified. It is unconfirmed. The correct posture is interested skepticism, not dismissal — and not confident therapeutic framing in consumer content.

Anyone with a browser can audit the independence of a peptide's evidence base in ten minutes. The workflow:

1. PubMed author-facet search. Search the compound name. Click the "Authors" facet in the left sidebar. If the top author accounts for more than half the records, and the next three authors are co-authors at the same institution, you are looking at single-lab dominance.
2. Google Scholar author-affiliation filter. On a Scholar result, hover the author name. Scholar reveals the listed affiliation. Cross-check across the top twenty hits. A single institution appearing on more than half of them is the same signal.
3. PubMed "Similar articles" clustering. On any paper's PubMed record, the right rail shows similar articles. If the cluster is dominated by the same author list in permuted order, the citation graph is intramural, not independent.
4. Funding-source check. Read the funding declarations of the top ten papers. If more than half cite the same national-science grant number or the same private sponsor, the independence is further narrowed.
5. Reagent-source check. Look at the Materials section of several papers. If they all obtained the peptide from the same custom-synthesis lab or the same corresponding author, the supposed replications share an input channel — meaning a bad batch would affect every paper.

This audit is not a literature review. It is a quality-control check that takes less time than reading a single abstract in detail, and it should precede any confident claim about a compound's generalizability.

A sixth step is worth doing when the compound has a commercial champion: search author AND supplier to see whether the principal investigator, a co-author, or a past trainee of the originating lab sits on a supplier's scientific advisory board. This is not disqualifying — founder-scientists advising companies working on their compounds is a normal pattern in biotech — but it is a disclosure readers deserve, and it is information that consumer content about the peptide almost never includes.

Understanding why some compounds stay intramural for decades clarifies what the pattern does and does not mean about the underlying biology.

• Specialty reagents. Research peptides are synthesized to order. If the canonical supplier of a compound is a small peptide synthesis facility associated with the originating lab, other groups have to either source from that facility or invest in their own synthesis and purity characterization before they can begin experiments. The activation energy for a second laboratory is high enough that many never cross it.
• Negative-result publication bias. A different lab that tried the compound and saw nothing has a strong incentive not to publish. "We tried X and it did nothing" is a harder sell to a journal than a positive mechanism paper, and it invites conflict with the originating group on peer review. The absence of failed-replication papers is not evidence that nobody tried.
• Funding-line continuity. A successful grant in a lab tends to renew for the same lab. A compound that produced fundable data in one group produces more fundable data in the same group. Independent replication requires a second group to persuade a second funder to underwrite work that duplicates an existing line — which is exactly the kind of project funders are structurally reluctant to support.
• Therapeutic area niche. A compound studied at the intersection of two small fields (for example, gut motility and wound healing) will not attract attention from either field's mainstream researchers, so the originating lab can hold the territory unchallenged for a long time without anything nefarious happening.
• Regulatory framing. When a compound is not on a track toward formal clinical development, it does not benefit from the standardization of methods that Phase 1 and Phase 2 trials force. Every group that works on it ends up doing so with slightly different material and slightly different assays, which makes head-to-head comparison difficult even when it is attempted.

These forces explain why single-lab dominance can be long-lived without implying misconduct. They also explain why it takes so long for the pattern to resolve: the structural conditions that produced it often persist even after the compound becomes commercially interesting, because the commercial interest tends to flow to the originating group's spinout rather than into public replication.

BPC-157

The canonical example. More than 80 percent of PubMed records on BPC-157 list one research group — Sikiric and Seiwerth at the University of Zagreb — as first or senior author, spanning roughly three decades of publication. The group has published plausible mechanisms in rat tendon, gut, and vascular models, but the work has not been independently replicated at scale by Western academic laboratories. No completed, peer-reviewed Phase 2 or Phase 3 human trial has been published.

Two independent signals of interest exist: the Chang group at National Taiwan University reported rat tendon fibroblast effects PubMed21030672, and the Hsieh group reported VEGFR2-mediated angiogenesis PubMed28245525. Those are meaningful independent observations, but they sit beside a literature in which the originating lab dominates every substantive claim about therapeutic relevance.

SS-31 / elamipretide

The mechanistic work on SS-31 binding cardiolipin in the inner mitochondrial membrane traces almost entirely to the Szeto laboratory (originally Cornell) and to Stealth Biotherapeutics, the company founded to develop it. Independent mechanistic confirmation exists but the bulk of preclinical rationale is intramural to that program. The most important pieces of independent evidence are the company-run randomized trials themselves — notably MMPOWER-3 PubMed36754636, the Phase 3 trial in primary mitochondrial myopathy which failed its primary endpoints. The mechanism story and the pivotal-trial story are pointing in different directions, and the pivotal trial is the one that counts for therapeutic claims.

MOTS-c

The discovery and most downstream biology of MOTS-c come from the Cohen laboratory at USC, which first identified the peptide as a mitochondrial-derived regulator of AMPK PubMed25738459. Later independent work has extended the pattern — for example the Reynolds study on acute exercise induction PubMed33504802 — but the foundational biology still leans heavily on a single group. Human-therapeutic randomized evidence does not exist.

Not every peptide fits the single-lab pattern. The counterexamples are instructive because they show what distributed evidence actually looks like.

GLP-1 receptor agonists

Semaglutide, liraglutide, and the rest of the GLP-1 class have been published on by hundreds of independent laboratories across dozens of countries over roughly four decades. Industry pivotal trials sit beside independent academic mechanism work, independent meta-analyses, independent post-market surveillance, and regulatory review from multiple national agencies. The STEP program PubMed33567185 and SELECT PubMed37952131 are public, large, and have been scrutinized by reviewers who do not share institutional affiliation with the sponsor.

Retatrutide

The triple agonist at GLP-1, GIP, and glucagon receptors entered Phase 2 with results published in NEJM PubMed37366315 and is advancing through Phase 3 trials registered on ClinicalTrials.gov with tens of thousands of participants across multiple sites and investigators. The sponsor (Eli Lilly) is the same across the program, which is normal for a regulated development pathway — but the participating sites, principal investigators, and independent data-safety monitoring boards are distributed. That is the structural difference between a compound in regulated development and a compound in single-lab academic development.

Single-lab dominance is an explicit input to evidence-tier assignment on every peptide page. The rule we apply:

• A compound whose primary-literature base is more than 60 percent intramural to one laboratory cannot advance above Tier 3 regardless of how frequently it is cited in consumer content, regardless of how plausible the mechanism is, and regardless of affiliate relationships with suppliers who sell it.
• A compound with mechanism-only evidence from a single lab and no independent replication sits in Tier 4 (speculative) until independent work appears.
• Every peptide reference page lists a Research groups footer that names the principal investigators behind the top five cited papers and their institutions, so readers can see the distribution for themselves.
• When independent replication does appear, we update the page in the changelog rather than silently re-tiering, so the history of what we believed and when is auditable.

This is not an anti-promising-compound posture. It is the opposite: we want the evidence base for interesting peptides to get stronger, and we think the best way to encourage that is to describe the current evidence accurately rather than to let marketing language smooth over the replication gap.