Editorial standards

Last updated 2 July 2026 · Adopted November 2021 · Revised annually

1. The no-advice, no-commerce boundary

Sequence & Signal publishes explanatory journalism about peptide science, clinical evidence, and regulation. We do not practise medicine, and nothing we publish is individualised medical advice. Reading this site does not create a clinician–patient relationship with anyone, and we are not qualified to have one with you.

The boundary is drawn concretely, and it is not negotiable at any price:

• We publish no dosages, dose ranges, concentrations, reconstitution or preparation instructions, routes of administration, cycles, schedules, combinations, or protocols of any kind.
• We publish no vendor names, supplier names, marketplaces, purchase links, affiliate links, discount codes, or price information, and we do not provide them privately on request.
• We publish no synthesis or compounding instructions.
• We do not state or imply that any unapproved compound treats, prevents, cures, or improves any condition.
• We publish no personal-use anecdotes, testimonials, user reports, results claims, or physique imagery. A single experience is not evidence, and presenting it as evidence is a category error with a body count.
• We do not use "research use only" or any similar formula as a disclaimer wrapped around consumer-facing product talk. If a framing exists to make product promotion legally survivable, we do not use the framing, because we do not do the promotion.

An editor may decline any piece, section, or sentence on these grounds without further justification. In practice this most often means cutting a specific figure from an otherwise sound explanation. We would rather publish a slightly less complete article than a slightly more actionable one.

2. Sourcing hierarchy

Not all evidence carries the same weight, and we rank it explicitly rather than leaving readers to infer it from tone. In descending order:

1. Peer-reviewed primary literature. Randomised controlled trials, systematic reviews, and meta-analyses in indexed journals, read in full rather than by abstract. Trial reports are read alongside their registered protocol and statistical analysis plan where those are available.
2. Regulatory and standards documents. Approval packages, review memoranda, labelling, guidance documents, advisory committee materials, inspection findings, enforcement correspondence, pharmacopeial chapters, and international standards. We identify the version in force at the time we are describing, because guidance is revised and the current version is often not the relevant one.
3. Preprints and conference abstracts. Legitimate and useful, and always labelled as unreviewed in the sentence that relies on them, not in a note at the bottom. We do not build the central claim of a piece on a preprint alone.
4. Expert interviews. Named researchers and practitioners, used to interpret and contextualise the documentary record — not to substitute for it. An expert's assertion is reported as their assessment, not as established fact.

Below that line: press releases, sponsor communications, secondary reporting, marketing material, and forum content. We read these to know what is being claimed. We do not cite them as evidence for whether a claim is true.

Every factual claim is attributed in the text or in the Sources block at the end of the piece, by publication name, authorship style, and year, or by agency name and document title. We deliberately cite without links or identifiers. Two reasons. Link rot silently degrades a citation into a dead end that looks intact, and a reader who searches a named journal, author and year finds the current version, its corrections, and any retraction notice — which a stale link would have hidden. We never invent an identifier to make a citation look more precise than our reading was.

3. Compound classification

Every time we mention a compound or compound class, we state which of three categories it falls in. We repeat this in every piece rather than assuming a reader has seen it before.

4. Fact-checking process

Every piece is checked before publication by someone who did not write it. The process is unglamorous and it is followed:

1. The writer annotates the draft, sentence by sentence, with the source supporting each factual claim.
2. The fact-checker works from the primary documents, not the writer's summary of them. Where a claim rests on a study, the checker reads the study.
3. Chemistry, pharmacology, and statistical claims are additionally reviewed by our science editor or by an outside reader with relevant expertise, who is disclosed on the piece if they are named.
4. The checker flags any sentence that could be read as instruction, endorsement, or a route to obtaining something. Flagged sentences are rewritten or cut. This check is run separately from the accuracy check, because the two failure modes are different.
5. Nothing publishes on a deadline that would require skipping a step. If checking is not finished, the piece stays in edit and the library shows it as such.

5. Corrections policy

We get things wrong. When we do, we fix it in the open.

How to report an error

Email editors@sequenceandsignal.com or use the contact form with "Correction" selected. Quoting the sentence helps. You do not need to prove the error to us; a specific, good-faith question is enough to trigger a check. We acknowledge corrections reports within two business days.

What happens next

• Correction. A material factual error — a wrong year, misattributed finding, misstated endpoint, misdescribed regulatory status. We amend the text and append a dated correction note at the foot of the piece describing what was wrong and what it now says. The note is permanent and is never removed in a later revision.
• Clarification. The text was accurate but readably ambiguous. We rewrite for clarity and append a dated note.
• Update. New information has superseded an accurate piece. We add a dated update block at the top and leave the original text intact below it.
• Retraction. A piece whose central claim does not survive checking is retracted. We leave the URL live with a retraction notice explaining what failed and why, because silently deleting a page destroys the record for anyone who read or cited it.

We do not make silent edits to substantive claims after publication. Typographical and formatting fixes are made without notice; anything that changes meaning gets a note.

6. Conflict-of-interest disclosure

Staff and contributors disclose, before assignment, any financial or professional interest that a reasonable reader might consider relevant: equity holdings, consulting or advisory relationships, speaking fees, research funding, employment history within the past five years, and immediate-family equivalents of the same. Disclosures are held by the editor and published on the affected piece.

Staff may not hold equity in, consult for, or accept payment from any company that manufactures, compounds, distributes, prescribes, or markets any substance within our coverage area, or from any trade association representing them. Where a contributor has a relevant interest that cannot be avoided, the interest is disclosed prominently on the piece and a second reviewer without that interest reviews the work.

If a conflict comes to light after publication, we disclose it on the piece with a dated note and re-review the work. If the conflict is serious enough to undermine the piece, we retract it under the policy above.

7. Funding disclosure

Sequence & Signal is funded by voluntary reader contributions and a small number of institutional subscriptions held by libraries and pharmacy programmes. As of the date on this page we carry no advertising, accept no sponsored content, hold no affiliate relationships, and have received no funding — direct, in kind, or through an intermediary — from any pharmaceutical manufacturer, compounder, distributor, telehealth service, supplement company, or trade association.

No funder, contributor, or subscriber receives advance sight of, or influence over, editorial content. Institutional subscriptions are sold at a published rate and confer no editorial privileges whatsoever. Should any of this change, the change would be disclosed at the top of this page, dated, before it took effect anywhere on the site.

8. Use of automated tools

We do not publish machine-generated text as editorial content. Automated tools are used, where useful, for transcription, for searching document sets, and for spotting inconsistencies in a draft. Every sentence that publishes has been written or rewritten by a named human being, and every factual claim is verified against a primary source by a human being. Where a tool is used in reporting in a way a reader would want to know about, we say so in the piece.

9. Anonymous sources and right of reply

We grant anonymity sparingly, and only where a source faces a credible professional or legal risk and the information cannot be obtained otherwise. The editor knows the identity of every anonymous source. We describe such sources as specifically as safety allows, and we say why anonymity was granted. We do not grant anonymity to allow someone to attack a named party without accountability.

Any person or organisation whose conduct is the subject of critical reporting is given a fair and specific opportunity to respond before publication, with enough detail to answer and enough time to do so. Where a response is declined or not received, we say that in the piece.

---

Questions about this policy, or about how it was applied to a specific piece, go to editors@sequenceandsignal.com. See also our disclaimer, privacy policy, and terms of use.