A research group reads together long before it writes together. A new doctoral student is handed the papers the lab is built on, each grant gathers its own literature, the journal club adds a paper a fortnight, and every co-authored manuscript draws on all of it. This guide shows how to keep a research group's shared sources as one body: a core reading list for new members, sources organised by project, a journal club that leaves a record, and citations the co-authors agree on.
EdCitation publishes it. The documents below and two papers on journal clubs were read at source on 24 September 2026, and the tool runs quoted date from the same day. Find sources gathers the reading, Cite a source makes each entry from the publisher's record, Verify references checks a manuscript, and a free EdCitation account keeps the group's list and shares it by email. A student group sharing one list for one assignment is covered in share a reference list with your group; a lab's list must outlast every student and grant.
What do funders and universities expect of a research group's sources?
They expect honest, accurate citing and good records, and they make whoever leads the group responsible for the culture around both. None of the documents we read prescribes a reading list, a reference manager or a citation style.
| Document | What it says | For the group's list |
|---|---|---|
| UKRI policy on good research practice (2026) | Grant holders develop "a positive and learning culture" in their teams; good practice includes record-keeping and training new staff and students | Onboarding reading and tidy records are part of the job |
| Concordat to Support Research Integrity (2025) | Honesty includes honesty in referencing and in acknowledging others' work; organisations follow this edition from April 2026 | Cite what the group actually used |
| European Code of Conduct for Research Integrity (2023) | Citing "selectively or inaccurately" is an unacceptable practice; data, code and software are citable products of research | No courtesy citations; datasets get entries |
| NIH intramural guidelines (9th edition, amended 2026) | Recommends a welcome letter or lab compact for new trainees; its section on AI says every claim and citation in the published text should be verified | A reading list beside the welcome letter; a check before submission |
| University of Glasgow code of good practice (version 3.8) | The PI or supervisor carries overall accountability for good practice and makes sure the group knows the university's and funders' policies | A named keeper |
| Imperial College London code for supervisors (2016) | Supervisors may suggest background reading before a student starts, and guide students to read original literature critically | Where a core reading list comes from |
UKRI's policy treats inadequate record-keeping as poor practice, normally put right with training rather than discipline (UK Research and Innovation, 2026b). The European code, from ALLEA (2023), is sharper about citing: its unacceptable practices include expanding a bibliography "to please editors, reviewers, or colleagues", a line worth reading to any group whose shared list has turned into a list of each other's papers.
What none of them says
None of these documents requires a shared reading list or reference list, or names a tool or style for one; the nearest is Imperial's note that a supervisor may suggest background reading before a student enrols. The advice below on keepers, tags and journal clubs is ours, built on those principles.
Data belong in the list too
UKRI's common principles on research data say published results should always say how to reach the supporting data, and that anyone using research data should acknowledge its source (UK Research and Innovation, 2026a). NIH's intramural research office summarises the NIH data management and sharing policy, in force since 25 January 2023: data behind a publication are shared by the time it appears (NIH Office of Intramural Research, 2023). The University of Glasgow (2026) wants data of long-term value in a repository, citable by a persistent identifier such as a DOI, for at least ten years.
So a group's list holds its datasets and software too. Find sources has Dataset among its kinds of source, and Cite a source builds a dataset's entry from its DOI: a 2026 Zenodo deposit came back in APA 7 labelled "[Data set]". The search listing spelled one author with an accent the registry record lacks, so check names on the repository's page.
How do you build a core reading list for new members?
Keep it short, ordered and explained: the papers the group's work rests on, its methods, the codes it works under, and its own recent papers, each with a line on why it is there. NIH's guidelines recommend a "Welcome Letter" or "Lab Compact" setting out what trainees and supervisors can expect of each other (National Institutes of Health, 2025). They never mention a reading list, but one sits naturally in that letter.
- List what a newcomer needs first: foundations, methods, the policies above and the group's own papers.
- Search each topic in Find sources, which covers some 300 million published works, by topic or claim, narrowed by kind of source, years, open access or times cited.
- Bookmark candidates to the reading list on the keeper's free EdCitation account, and read them before choosing.
- Make each entry from its DOI or ISBN in Cite a source, tagged "core" and with its topic.
- Put the one-line reasons and the reading order in the welcome letter.
- Share it from the Team page, which invites each newcomer by email to see the list; they create a free account to open it.
- Once a year, paste it into Verify references, so a retraction is caught before a newcomer builds on it.
A worked search for the data-practice reading
On 24 September 2026, research data management FAIR principles returned 704,135 works, led by Wilkinson and colleagues' 2016 Scientific Data paper setting out the FAIR Guiding Principles, open access and cited 18,902 times. Set to open access only, cited 100 times or more and most cited first, the search returned 12,876 works, still led by that paper, then a 2022 paper on FAIR principles for research software (505 citations). Four lessons came off the same page:
- An addendum is a record of its own. A 2019 addendum to the Wilkinson paper came ninth. Read it with the paper, and cite the paper.
- Filters narrow; reading decides. The twelfth result, a review of fairness in AI data management, was not about FAIR data.
- Records disagree. Built from their DOIs, the paper and the addendum give the third author as "Aalbersberg, Ij. J." and "Jan Aalbersberg, I.": one person, filed two ways.
- Most cited is not most relevant. The looser words reproducibility research practices, most cited first, opened with a 2000 guide to structural equation modelling. Sort by citations once the words and filters are tight.
How should a group organise sources across projects and papers?
In one list, with a tag for each project, grant and paper, so a source used by three projects is entered once and corrected once. Organising references with tags sets out a general scheme; a lab adds its own: core for newcomers, one per grant using the funder's reference, one per paper in preparation added when a citation enters the draft, journal club with the date, our papers so everyone cites the group's work alike, and data for datasets and software.
In EdCitation a reference carries as many tags as it needs, and filtering by one shows that project or paper alone. The list is saved to the account rather than to one computer, and goes out copied with its italics or as Word, BibTeX or RIS.
Who keeps the list
One named member, often a postdoc or lab manager answering to the PI, since Glasgow's code puts overall accountability for good practice with the PI or supervisor. Everyone else sends the keeper the DOI or ISBN of what they have read; the keeper adds, tags and corrects, and the Team page lets the group see the result. When a student leaves, the list stays with the group, and the keeper can download it as RIS or BibTeX for the leaver's own reference manager (moving references between tools).
How do you run a journal club that leaves a record?
Meet on a fixed schedule, circulate the paper beforehand, share the choosing, and record every paper discussed. Deenadayalan et al. (2008) reviewed health journal clubs: of 101 articles, 21 formed the evidence and 12 described effectiveness, at moderate quality, and over 80% of the papers reported gains in knowledge and critical appraisal. Clubs that worked met regularly, had a clear purpose and a trained leader who chose the papers, circulated them in advance, used an established appraisal method and summarised what was found. No paper followed the evidence from a club into clinical practice, so this is good sense with support, not proof.
Lonsdale et al. (2016), writing about a bioinformatics club they ran, record the papers discussed and who led, partly so newcomers can see what has been covered, pair a specialist paper with a review, and treat a flawed paper as good material. They admit they never found the best way to get the papers to everyone.
A shared list answers that. The chosen paper goes on the reading list, Share beside it sends that one source to a colleague, and the Team page lets the whole group see the reading list. Afterwards its entry joins the reference list tagged "journal club" with the date, and that tag's filter is the club's history. Cite a source screens each paper for retraction, which matters for a paper chosen for its flaws; retracted papers in your reference list explains the next steps.
How do co-authors keep citations consistent across papers?
Agree three things before drafting, and check the merged manuscript before submission: one entry per work, built from its identifier; which version of a work to cite; and the target journal's style. Every author answers for the list. The ICMJE criteria (International Committee of Medical Journal Editors, n.d.), which the University of Glasgow (2026) adopts, ask each author to know which co-author did which part and to have confidence in the integrity of the others' contributions.
The preprint or the published version
APA's guidance is to cite the final published version where you can, and the preprint when that is the version you used (American Psychological Association, 2020). The trap is that both stay public, as NIH's guidelines note: one co-author cites the preprint from a lab meeting two years ago, another the journal article, and the manuscript credits one study twice.
Bar and colleagues described antibody-guided biotinylation in a 2016 preprint and then in Nature Methods. From its DOI, Cite a source gave the preprint in APA 7 as:
Bar, D. Z., Atkatsh, K., Tavarez, U., Erdos, M. R., Gruenbaum, Y., & Collins, F. S. (2016). Biotinylation by antibody recognition - A novel method for proximity labeling. bioRxiv. https://doi.org/10.1101/069187
The journal article came back with its revised title, volume 15, issue 2, pages 127–133. Three details matter to the keeper. The article's Crossref record lists the preprint's DOI as its preprint, so the pair is on the record. The preprint entry names bioRxiv, the server readers know, and in Vancouver it is marked "[Preprint]". And the article is dated 2017 in Harvard, when it appeared online, but 2018, its print issue, in APA 7, Chicago, IEEE and Vancouver. Build every entry of one paper in one style from the record, and the years agree.
The check before submission
We put the pair into a short test manuscript of our own, drafted as three co-authors might, with the FAIR paper and the Lonsdale paper beside them. Uploaded as a Word file to Verify references, it came back:
4 in-text citations · 4 entries · 0 orphan citations · 0 uncited entries.
All four entries were verified. The report did not say that two describe one study, because both are real and both are cited; that choice is the authors', and APA's answer is the article. For a manuscript checked at every revision, References from a file keeps it in My papers so each run shows what changed. A free account allows three of those checks a month and Pro has no limit; the paper itself stays in the browser where it was checked. The full routine is in checking references before journal submission.
When a paper moves to a journal in another of the six styles, set the list to that style and copy it again. Co-authors in LaTeX take the BibTeX download (BibTeX for LaTeX and Overleaf), and colleagues in Zotero or EndNote take RIS.
Why keep a research group's shared sources in EdCitation?
EdCitation is the strongest choice for a group list that must be exact, because each entry is found in the publisher's record, never written: two members who add one paper get one identical entry, and a correction is made once, where everyone sees it. In the runs above, a preprint, its journal version and a dataset came back from their DOIs ready to use, and four citations met four verified entries; an author filed two ways, an accent and the choice between versions were left to people.
- Free without an account: Find sources, Cite a source (APA 7, MLA 9, Chicago 18 author-date, Harvard, IEEE, Vancouver), Check your paper and Verify references.
- Free with an account: the reference and reading lists, saved to the account; tags, filtering by tag, and a style per reference; copies with italics and Word, BibTeX or RIS downloads; the Team page, which sends email invitations to see either list, each person creating a free account; and Share for one source.
- Paid: Pro, at $8 a month, removes the monthly limit on References from a file and brings Mechanics QA, Check AI and Check plagiarism; Max, at $24 a month, brings Theoretics QA and the Library. A department can take an Institution licence; see Pricing.
Invited members see the list rather than edit it, so the keeper's account is where it changes. EdCitation does not store the group's PDFs or cite inside Word as you type, as Zotero and Mendeley do, and it writes no part of anyone's paper. Lecturers building a module's list for students should read our guide to class and cohort reading lists instead.
Quick questions
Should a research group keep one reference list or one per project?
One, with a tag for each project, grant and paper, so a source is entered once and a correction reaches every project.
Can every lab member edit the group's list in EdCitation?
No. The Team page invites members by email to see the reading list or the reference list, each creating a free account, and changes are made from the keeper's account. Members send the keeper DOIs, or build entries with Cite a source.
Do funders require a lab to keep a shared reading list?
Not in any document we read. UKRI, the Concordat, the European code and NIH's guidelines ask for honest citing, good records and a group leader who sets that culture; a shared list is one practical way to meet them.
Should we cite the preprint or the published version?
APA's guidance is the final published version where you can, and the preprint only if that is what you used. In one paper, cite one version of a study, with the same year in the entry and every citation.
What makes a lab journal club work?
The review by Deenadayalan and colleagues found that successful clubs met regularly, had a clear purpose and a leader who chose the papers, circulated them in advance and summarised what was found.
References
- ALLEA. (2023). The European Code of Conduct for Research Integrity (Rev. ed.). https://doi.org/10.26356/ECOC
- American Psychological Association. (2020). Preprint article references. APA Style. https://apastyle.apa.org/style-grammar-guidelines/references/examples/preprint-article-references
- Deenadayalan, Y., Grimmer-Somers, K., Prior, M., & Kumar, S. (2008). How to run an effective journal club: A systematic review. Journal of Evaluation in Clinical Practice, 14(5), 898–911. https://doi.org/10.1111/j.1365-2753.2008.01050.x
- Imperial College London. (2016). Code of practice for research supervisors [PDF]. https://www.imperial.ac.uk/media/imperial-college/administration-and-support-services/registry/academic-governance/public/academic-policy/codes-of-practice-for-research-degrees/Code-of-Practice-for-Research-Supervisors.pdf
- International Committee of Medical Journal Editors. (n.d.). Defining the role of authors and contributors. ICMJE Recommendations. Retrieved 24 September 2026, from https://www.icmje.org/recommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-and-contributors.html
- Lonsdale, A., Sietsma Penington, J., Rice, T., Walker, M., & Dashnow, H. (2016). Ten simple rules for a bioinformatics journal club. PLOS Computational Biology, 12(1), Article e1004526. https://doi.org/10.1371/journal.pcbi.1004526
- National Institutes of Health. (2025). Guidelines and policies for the conduct of research in the Intramural Research Program at NIH (9th ed., amended September 2026). Office of Intramural Research. https://oir.nih.gov/system/files/media/file/2026-06/guidelines-conduct_research.pdf
- NIH Office of Intramural Research. (2023, February 17). 2023 NIH data management and sharing policy. NIH Sourcebook. https://oir.nih.gov/sourcebook/intramural-program-oversight/intramural-data-sharing/2023-nih-data-management-sharing-policy
- Research Integrity Concordat Signatories Group. (2025). The Concordat to Support Research Integrity 2025. UK Committee on Research Integrity. https://ukcori.org/wp-content/uploads/2025/12/The-Concordat-to-Support-Research-Integrity-2025.pdf
- UK Research and Innovation. (2026a, September 15). Making your research data open. https://www.ukri.org/manage-your-award/publishing-your-research-findings/making-your-research-data-open/
- UK Research and Innovation. (2026b, April 1). Policy on the governance of good research practice. https://www.ukri.org/publications/ukri-policy-on-the-governance-of-good-research-practice/policy-on-the-governance-of-good-research-practice/
- University of Glasgow. (2026). Code of good practice in research (Version 3.8). https://www.gla.ac.uk/media/Media_490311_smxx.pdf