skills/open-science/preprints

stars:0
forks:0
watches:0
last updated:N/A

Preprints

A preprint is a complete, draft manuscript posted to a public server before peer review. Preprints accelerate dissemination, establish priority ("scoop protection"), invite community feedback, and increasingly count as formal scholarly outputs (NIH allows preprints in grant submissions, Plan S funders require immediate open access, and PubMed Central indexes NIH-funded preprints since 2020). This skill covers server selection, licensing, versioning, and the journal transfer process — the four decisions a researcher must make before the first click.

When to use

  • Choosing which server to post a manuscript (bioRxiv vs medRxiv vs ChemRxiv vs SSRN vs arXiv vs EarthArXiv vs PsyArXiv).
  • Selecting a license for the preprint (CC-BY-4.0 vs CC-BY-NC-ND-4.0 vs CC-BY-NC-4.0 vs no license / "all rights reserved" until acceptance).
  • Planning versioning: how to handle v1, v2 (after reviewer feedback), v3 (post-acceptance).
  • Understanding scoop protection: what a preprint does and does not protect.
  • Navigating journal transfer: In Review, Reviewed Preprint, and direct-to-journal workflows.
  • Building a preprint-first lab culture: lab norms, timing, and policy.
  • Grant or job applications that ask for a list of preprints.

When NOT to use

  • For preregistration of a study design (no results yet) — see ors-open-science-preregistration.
  • For data deposits without a manuscript — see ors-open-science-fair-data.
  • For clinical trials registration (which has its own regulatory regime) — see ors-open-science-preregistration (ClinicalTrials.gov section).
  • For code release (no paper) — see ors-open-science-code-release.
  • If the manuscript reports findings on dual-use research of concern (DURC) or pathogens of pandemic potential — pre-publication review with NIH/NSABB may be required before posting.

Prerequisites

  • A complete draft manuscript (not an outline; not a partial study).
  • ORCID iD for all authors.
  • Conflict-of-interest disclosure statement (most servers require one).
  • Approval of all co-authors for posting.
  • Institutional preprint policy check (some institutions have rules about timing, embargoes, or which server).

Core workflow

  1. Decide the server. Use the decision tree in "Document patterns" below.
  2. Choose a preprint license. Default: CC-BY-4.0 (most permissive, most funder-aligned). Some journals mandate CC-BY-NC-ND-4.0 if you intend to transfer copyright later — check the journal's preprint policy.
  3. Prepare the manuscript file. Single PDF (or Word for ChemRxiv), including all figures, tables, and supplementary material inline. Most servers allow supplementary files as separate uploads.
  4. Submit metadata. Title, abstract, author list with ORCID iDs, conflict of interest, funding statement.
  5. Screen on submission. Servers do a basic screen (~24-48 h): scope fit, plagiarism check, dual-use research of concern, ethics statement for human/animal work. bioRxiv/medRxiv famously do not perform peer review.
  6. Post v1. Once accepted by the screen, the preprint is public with a DOI. The DOI never changes; version increments.
  7. Version up as the work evolves. v2 after peer-review feedback, v3 after journal acceptance, v4 with the published-version link.
  8. Use the server's journal-transfer or reviewed-preprint service (In Review, Review Commons, biOverlay) to forward the manuscript to a journal — preserving the preprint DOI.

Document patterns

Pattern 1: Server selection decision tree

What is the field?
├── Life sciences (biology, bioinformatics, ecology, neuroscience, etc.)
│   └── bioRxiv — life-sciences preprint server (Cold Spring Harbor Laboratory)
│       ├── Clinical / patient-level health science?
│       │   └── medRxiv — clinical preprint server (also CSH)
│       └── Chemistry subfield?
│           └── ChemRxiv — chemistry preprint server (ACS, RSC, DECHEMA, GDCh)
├── Physical sciences / math / CS / quantitative biology
│   └── arXiv — oldest preprint server (Cornell, 1991)
├── Social sciences / economics / law / finance
│   └── SSRN — Elsevier-owned social-sciences network
├── Earth sciences
│   └── EarthArXiv
├── Psychology
│   └── PsyArXiv
├── Engineering
│   └── engrXiv
├── Education
│   └── EdArXiv
├── Medical / health-services research (alternative to medRxiv)
│   └── medRxiv OR preprints.org (multi-disciplinary, ASAPbio-listed)
└── Truly cross-disciplinary or unsure
    └── preprints.org (multidisciplinary, accepts any field) OR Research Square

Direct links to canonical server homes (path-only, server names are well-established):

Pattern 2: License comparison

LicenseShareAdaptCommercial useJournal friendly?Notes
CC0 1.0YesYesYesUnusual for manuscriptsMost permissive; used for datasets more than papers.
CC-BY-4.0YesYesYesMost permissive; some journals require transfer of non-exclusive rights on topFunder-preferred (Plan S, NIH, Wellcome).
CC-BY-SA-4.0YesYesYes (with same license)Sometimes accepted; some journals reject"Copyleft" — derivatives must use the same license.
CC-BY-NC-4.0YesYesNoCommon compromise"Non-commercial" clause can complicate reuse.
CC-BY-NC-ND-4.0YesNoNoMost restrictive; many journals accept thisNo derivatives; e.g., cannot use the figure in a derivative review.
No license / "all rights reserved"Default under copyrightNoNoYes (most journals assume this until acceptance)Limits reuse, but some journals require it pre-acceptance.

Practical default: post as CC-BY-4.0. If the target journal forbids this, switch to CC-BY-NC-ND-4.0 before posting. Check the journal's preprint policy on its "Instructions for Authors" page (Sherpa Romeo lists them: https://www.sherpa.ac.uk/romeo/).

Pattern 3: Scoop-protection timeline

Day 0          Manuscript complete. Post to preprint server. DOI minted.
Day 1-2        Screen (~48 h). Authors notified.
Day 3          Preprint public. Author priority established.
               Note: priority is social/convention-based, not legal.
               No formal "patent" or "first-to-publish" right is granted.
Day 30-180     Submit to journal. Referee reports received.
Day 180-365    Revise; version v2 of preprint with reviewer response.
Day 365-540    Journal accepts; v3 with accepted manuscript.
               Add DOI of the published version.
Day 540+       v4 (optional) with the version of record link.

What scooping means in practice: if a competing group publishes the same finding before you submit, the preprint timestamp is evidence that you had the finding first. It does not guarantee authorship priority — the convention is "whoever's paper comes out first" in many fields. Post early for protection, but don't expect legal rights.

Pattern 4: Version metadata

Each version of a preprint has the same DOI, with a version suffix. Use the README to record changes:

## Versions
- v1 (2025-09-12): Initial submission.
- v2 (2025-12-04): Revised after bioRxiv community comments; added
  Supplementary Figure 3; clarified methods section 2.4.
- v3 (2026-03-22): Revised after journal peer review at Nature
  Communications; added replication cohort; reviewer responses in
  supplementary material.
- v4 (2026-06-15): Added link to published version (DOI 10.1038/...).

Pattern 5: Journal-transfer flow with In Review / Reviewed Preprint

bioRxiv / medRxiv (v1)
        │
        │ (In Review partnership)        ← optional
        ▼
Review Commons (peer review at a journal-independent platform)
        │
        ▼
Journal transfer: eLife, PLOS, Nature Communications, etc.
        │
        ▼
Reviewed Preprint (peer-reviewed version with reviews + response)
        │
        ▼
Version of record at the journal
  • In Review (bioRxiv/medRxiv partnership): forwards the preprint + supporting info to a partner journal.
  • Review Commons (EMBO): journal-independent peer review; you choose the journal after reviews.
  • eLife's Reviewed Preprints: eLife reviews the preprint and publishes a Reviewed Preprint (DOI + reviews + author response) before any journal's version of record.

Common pitfalls

PitfallWhy it failsFix
Posting to the wrong serverScope-mismatch → desk rejection by the serverUse the decision tree; ASAPbio has a unified submission portal.
No ORCID iDsAuthorship ambiguity; some servers require themAdd ORCID iDs for all authors at submission.
Posting before all co-authors approveAuthorship dispute (COPE case database is full of these)Get email approval from every co-author; keep a record.
Posting clinical advice or "treats COVID"medRxiv/ChemRxiv/bioRxiv explicitly forbid clinical recommendations; legal riskRestrict to the scientific finding; refer clinical readers to clinical guidelines.
Choosing CC-BY-NC-ND when a funder requires CC-BYPlan S / Wellcome / Horizon Europe compliance failureDefault to CC-BY-4.0.
Claiming "priority" in cover letterReviewers find it off-putting; "scooping" is a social convention, not a legal claimDon't lean on the preprint in cover letters; submit normally.
Posting a paper under review at Nature/Science without checkingMost Nature/Science family journals now allow preprints; Cell family allows; Elsevier/Wiley/ACS/Springer allow; some holdouts (some IEEE, some medical) prohibitCheck the journal's preprint policy; Sherpa Romeo aggregates them.
No version annotationReaders don't know if v1 differs from the published versionAdd a "v1 vs published" note; update v1's metadata with the published DOI.
Screen rejection because of ethicsHuman-subject research needs an ethics statement; animal research needs IACUC approval numberAdd to methods before posting.
Dual-use research posted without DURC reviewFederal DURC/PEPP review is required for certain pathogensPre-screen with institutional review board; do not post gain-of-function or enhanced-pandemic pathogen work without explicit approval.
Preprint cited as "peer-reviewed"It is not; reviewers will catch thisIn the paper, write "preprint (not peer-reviewed)"; in the journal version, cite the published version.
Self-archiving a journal-published PDF on a preprint serverMost journals require you to use the AAM (author accepted manuscript), not the VoR; some allow a 6/12/24-month embargoRead the journal's self-archiving policy; Sherpa Romeo lists it.

Validation

A preprint is "well-posted" when:

  • All co-authors have approved the version.
  • ORCID iDs are linked for every author.
  • License is explicitly declared (default CC-BY-4.0).
  • Ethics statement and conflict-of-interest statement are present.
  • Funding statement and grant numbers are present.
  • A "Comments" link (bioRxiv, medRxiv) is monitored and replied to.
  • The DOI is included in the lab's ORCID, Google Scholar, and the journal submission cover letter.
  • The preprint is listed on the lab's website and the authors' Twitter/X/Bluesky.
  • A v2 is posted when the journal reviews come back.
  • The published-version DOI is added to the preprint record.

Open alternatives

"Premium" / commercial serviceOpen alternativeTrade-off
Research Square (Elsevier, paid editing + DOI)bioRxiv / medRxiv (free)Research Square offers paid editing; bioRxiv is free with no editing.
SSRN (Elsevier)SocArXiv, MetaArXivSSRN is dominant in social sciences; SocArXiv is smaller but open.
Authorea (Wiley-owned)Manubot, QuartoAuthorea has nicer UX; Manubot is git-native, fully open.
Peerus (paid preprint commenting)bioRxiv's native comment thread; PubPeerPubPeer allows anonymous posting; bioRxiv's comments are public and tied to ORCID.
Journal "open access" fee ($11,690 for Nature)bioRxiv + self-archive (free)Journal OA is fee-based; preprint + self-archive is free but has no version of record.

References

Related skills

  • ors-open-science-preregistration — for studies without results yet.
  • ors-open-science-licensing — deep dive on license choice.
  • ors-scientific-writing-imrad-drafting — writing the manuscript before posting.
  • ors-ethics-compliance- (separate skills) — IRB / IACUC / DURC for clinical or pathogen work.

Changelog

  • 1.0.0 (2026-06-10): Initial adaptation by Pradyumna Jayaram. Synthesised bioRxiv/medRxiv posting guides; ChemRxiv author guidelines; arXiv help; ASAPbio server directory; Creative Commons license deeds; NIH Preprint Policy; Plan S; eLife Reviewed Preprints; Review Commons. Server decision tree, license comparison, scoop-protection timeline, and journal-transfer flow are original compositions.
    Good AI Tools