PLOS Is Changing What It Expects from Authors: Code Sharing, Data Deposits, and the Push Beyond APCs

PLOS (Public Library of Science) has been one of the most consequential organizations in academic publishing since it launched PLOS ONE in 2006 and helped normalize open access at scale. In 2026, the organization is making two moves that will affect how researchers prepare manuscripts for submission: a mandatory code-sharing policy now in force at PLOS Biology, and a broader strategic report published in May 2026 that proposes replacing the APC-based business model with something PLOS calls a “knowledge stack.” Both changes matter for medical authors, even those who primarily publish in clinical rather than biological science.

The code mandate is immediate and operational. If you submit to PLOS Biology from January 2026 onward, you are required to make all author-generated analysis code publicly available without access restriction at the time of publication. That is a harder line than most journals draw, and it will catch researchers who assumed that uploading data was enough. The broader strategic shift is longer-term, but it signals where PLOS expects the rest of the publishing ecosystem to follow.

What the January 2026 Code Mandate Actually Requires

PLOS Biology published an editorial in February 2026 titled “Formalizing our commitment to code sharing,” clarifying a policy that took effect for all submissions received from January 1, 2026. The requirement is not a gentle recommendation to share code if you happen to have written some. It is a mandatory condition of publication. All author-generated code required to replicate the study's findings must be made publicly available without access restrictions, either at the time of publication or, when journals allow it, during the editorial process.

What counts as “code” under this policy is broader than many researchers expect. It includes data analysis scripts, statistical code, custom algorithms, and any programming that shaped the results reported in the paper. It also explicitly includes custom code written inside commercial software packages: if you ran your analysis in Matlab, SPSS, or Excel with custom macros, that code falls within scope. The policy does not restrict what programming language you use, and it does not require you to apply any specific license to the deposited code, only that you indicate what license you have applied.

The mechanics of deposit matter. PLOS Biology requires code to be in a permanent, public repository that issues a citable digital object identifier or equivalent persistent identifier. GitHub and similar version control platforms are not considered permanent repositories on their own, because they can be deleted or taken private after publication. The accepted workflow is to archive a snapshot of the GitHub repository in a permanent repository such as Zenodo, CodeOcean, or the Software Heritage archive, and to cite that archived version in your Data Availability Statement. A link to a live GitHub repository without an archived snapshot does not satisfy the requirement.

There are narrow exceptions for code that cannot be publicly shared due to legal restrictions (contractual obligations with companies whose data you used), ethical restrictions (code that could be combined with restricted data to re-identify patients), or dual-use concerns. In those cases, PLOS Biology recommends using a controlled-access repository, such as a restricted Zenodo deposit, that allows interested researchers to request access. If even that is not possible, the Data Availability Statement must name a specific contact point, with sufficient information for a reader to submit a request. Stating that code is “available on request” without a named contact is not sufficient.

PLOS Computational Biology has operated under a similar mandatory code-sharing policy since 2022. The PLOS Biology policy extends that precedent to a broader life-science audience. What is not yet clear is the timeline for analogous requirements to reach PLOS Medicine and PLOS ONE, but the trajectory is visible. Authors who begin depositing code as part of their standard pre-submission workflow are unlikely to be caught off guard.

What PLOS Medicine and PLOS ONE Already Require for Data

Before turning to the strategic picture, it is worth being specific about what PLOS Medicine and PLOS ONE already require from medical authors, because compliance gaps here are common and catching papers at submission.

Both journals require a Data Availability Statement in every submission. The statement must specify where the data underlying the manuscript's figures, tables, and conclusions can be accessed. PLOS defines the minimum data set as the data required to replicate all study findings reported in the article, including related metadata and methods. Saying that data are “available from the corresponding author on reasonable request” is not acceptable at either journal. The data must be in a public repository or included as Supporting Information at the time of submission.

PLOS has integrated its submission systems with Dryad and FlowRepository to make the deposit step easier. Both journals also require that each submitting author have a verified ORCID iD, and that author contributions be described using the CRediT taxonomy of 14 roles. These are not new requirements, but they remain sources of last-minute friction for research teams that assume the corresponding author can handle them alone.

For clinical studies with protected health information, the data sharing requirements create real complexity. Patient-level data cannot normally be deposited in an open repository. PLOS Medicine handles this through a controlled-access deposit model, where the Data Availability Statement explains what data can be shared and under what access conditions, identifies the repository used, and provides a point of contact. Teams submitting clinical data studies should confirm the access structure with their IRB and legal office before they commit to any specific repository in the manuscript.

The “Redefining Publishing” Report: What PLOS Said in May 2026

In late May 2026, PLOS published the results of an 18-month research and design project, funded by the Gordon and Betty Moore Foundation and the Robert Wood Johnson Foundation, under the name “Redefining Publishing: Practical pathways to open science.” The report is a strategic statement about where PLOS believes scholarly publishing needs to go, and it contains two arguments that medical authors should understand.

The first argument is about article processing charges. PLOS helped popularize the APC model as a mechanism for funding open access: authors or their institutions pay a fee so that the published article is free to read. The May 2026 report acknowledges that this model has become a significant barrier. Stakeholders consulted during the 18 months said they want models that move beyond per-publication charges, that are transparent and predictable, that reflect regional economic disparities, and that align publishing support with a wider range of research outputs. The report does not announce the end of APCs at PLOS journals, but it signals clearly that PLOS is conducting a comprehensive economic analysis to explore a non-APC business model. That is a substantial commitment from an organization that currently charges $2,477 per article at PLOS ONE and higher at some of its specialty journals.

The second argument is about what counts as a research output. Current publishing treats the journal article as the dominant unit of scientific work, the thing that gets cited, indexed, and credited in promotion and funding decisions. Data, code, protocols, and other outputs are supplementary, attached to articles or deposited elsewhere and mostly invisible in citation and recognition systems. PLOS describes this as a problem, and the report proposes a concept called the “knowledge stack” as an alternative.

The Knowledge Stack: What It Means in Practice

The knowledge stack, as PLOS describes it, is a structured, open, machine-readable record that connects a published article or preprint with its associated research outputs: data, code, methods, materials, and protocols. The goal is not simply to make those outputs available (which data availability policies already attempt) but to make them visible as a coherent whole, with proper attribution and credit for everyone who contributed to each component.

In the current system, the researcher who spent a year building the statistical pipeline for a clinical trial gets a line in the CRediT taxonomy, possibly a middle-author position, and nothing in the citation record. Their code, deposited in Zenodo, sits there largely unseen unless another researcher happens to search for it. The knowledge stack concept would make that code a citable, discoverable first-class object within the publication record, linked to the paper and credited to its author.

PLOS has identified data and code as the starting points for practical implementation, because these are the most mature in terms of existing infrastructure: repositories exist, DOI systems work, and funders are already mandating them. Protocols, methods, and research plans are the next layer. PLOS's stated next steps involve developing publishing capability for data and code first, rather than attempting to transform the entire system at once.

For medical authors, this has immediate implications even before any formal knowledge stack infrastructure exists. Teams that deposit well-documented code and data with proper DOIs are already in a better position than those who attach a supplementary file. If PLOS moves forward with making those deposits citable and linked within its platforms, researchers who got into the habit of good deposits early will benefit from the discoverability gain. Researchers who skipped the step will need to backfill it.

Why This Matters Beyond PLOS Submissions

PLOS's influence on publishing norms extends well beyond its own journals. PLOS ONE introduced article-level metrics and peer review that evaluated scientific soundness rather than perceived significance, and that approach spread across the industry. PLOS's data availability policies were among the first to move from encouragement to requirement, and other publishers followed. When PLOS publishes a strategic report about changing the fundamental model of scholarly publishing, it is reasonable to expect other organizations to respond.

Several converging pressures make the knowledge stack argument more credible in 2026 than it would have been even three years ago. Funder mandates around data and code sharing have proliferated: NIH eliminated its embargo on public access in July 2025, and funders like HHMI and the Gates Foundation have their own requirements. The FAIR data principles (Findable, Accessible, Interoperable, Reusable) are increasingly cited in grant applications and journal policies alike. And the CRediT taxonomy, now mandatory at most major publishers, has created infrastructure for distinguishing contributions more precisely than the author list alone. The knowledge stack concept builds on all of that existing scaffolding.

It is also worth noting what the PLOS report says about APCs in the context of equitable publishing. A consistent criticism of the APC model is that it shifts the paywall from readers to authors, leaving researchers without institutional support unable to publish in open access venues. This is an acute problem in medical research conducted in low- and middle-income countries, where the research burden of disease is high but institutional APC subsidies are rare. The stakeholders PLOS consulted during its 18-month study specifically asked for models that reflect regional economic realities. Whether a non-APC model can be made financially sustainable at the scale PLOS operates is genuinely uncertain, but the organization is committing resources to exploring it.

What PLOS Computational Biology's Code Mandate Shows About What Comes Next

PLOS Computational Biology implemented mandatory code sharing before PLOS Biology did, and the experience there is instructive. After the policy took effect, submissions initially had higher revision rates as authors adjusted to depositing code and writing adequate Data Availability Statements. Over time, the average quality of deposited code improved, and peer reviewers began using the deposited code more actively in their assessments. The policy also revealed something journals had not previously been able to see: how often published results depended on custom code that had never been described in the methods.

The PLOS Biology policy follows a similar trajectory. Authors who have not deposited code before will need to build that step into their workflow. For medical researchers, this is most relevant in studies with substantial computational components: clinical prediction models, genomic analyses, image analysis pipelines, pharmacokinetic modeling, and any study where a custom statistical analysis was written rather than run through a standard software interface. Many clinical studies that involve only descriptive statistics in Stata or SPSS will find the policy easier to satisfy, because standard software operations do not generate novel code in the same sense.

The peer review implications are also worth thinking through. When code is available during review, a methodologically careful reviewer can test whether the code produces the reported outputs. That is a stronger check than what peer review has historically been able to perform. It raises the standard for methods reporting and, potentially, for what reviewers are expected to do. Some researchers find this uncomfortable. Others see it as the review process finally having enough information to catch the kinds of errors that currently pass through undetected.

Practical Steps for Medical Authors Before Submitting to Any PLOS Journal in 2026

If you are preparing a submission to any PLOS journal this year, the following steps will prevent the most common delays. The requirements differ by journal, so check the specific submission guide for your target title before submitting.

Start with your data and code inventory before you begin formatting. For every figure, table, and conclusion in the paper, identify the underlying dataset and any code used to produce or analyze it. Write this inventory down. It will become the foundation of your Data Availability Statement and, for PLOS Biology, your code deposit.

For data, deposit to a public repository with a DOI before you submit. Dryad and Zenodo are well-integrated with PLOS workflows. If your data include protected health information, work with your IRB and legal counsel to determine the appropriate controlled-access structure, and name the access point explicitly in the manuscript. Do not wait until after acceptance to sort this out.

For code, if you are submitting to PLOS Biology or PLOS Computational Biology, deposit a snapshot of your code archive to Zenodo, CodeOcean, or Software Heritage. Get the DOI. Verify that the code is accessible without login. Document what the code does at the level a reader would need to run it. Then include the repository link and identifier in the Data Availability Statement.

What the Timeline Looks Like from Here

PLOS has said its next phase of the Redefining Publishing initiative will focus on developing practical publishing infrastructure for data and code. That means building the systems that would make deposited data and code genuinely citable and discoverable within PLOS platforms, not just technically available in a linked repository. ASAPbio is hosting a community call on July 21, 2026 with PLOS Chief Scientific Officer Veronique Kiermer to discuss the initiative publicly, which suggests the organization is actively engaging stakeholders rather than treating the report as a closed planning document.

The non-APC business model is a longer horizon question. PLOS has committed to a comprehensive economic analysis, but developing a sustainable publishing model that does not depend on per-article charges would require coordination with funders, libraries, and institutions at a scale that takes years. The Wellcome Trust and UKRI have already begun withdrawing support for hybrid APC models. HHMI has restricted APC payments to fully open access journals since January 2026. Those funder decisions do not directly push PLOS toward a non-APC model, but they create a financial environment where a well-designed alternative could find institutional buyers.

For now, authors submitting to PLOS journals are still paying APCs, still receiving peer-reviewed articles as their primary output, and still navigating the same submission systems. The code mandate is the concrete, immediate change. The knowledge stack and the post-APC ambition are in the planning stage. But following what PLOS does in the next 18 months will tell researchers something meaningful about where the broader open-access and research-data landscape is heading.

The practical implication for researchers is simple: treat code and data deposits as primary submission materials, not afterthoughts. Whether or not PLOS journals you target have a formal code mandate today, the direction of travel is clear. Teams that build this into their workflow now will not need to reconstruct or recreate code months after analysis is complete, which is both the most common failure mode and the most avoidable one.

Further Reading