The authors' assertion that registration requirements necessarily impede interoperability (pages 3 and 5) conflates governed access with an inability to perform automated analyses.

Registration and authentication are not inherently incompatible with computational interoperability, nor does scientific utility necessarily depend on anonymous, unrestricted redistribution of sequence data between databases. GISAID provides an established example of how governed access can coexist with integrated analytical capabilities, allowing users to analyze curated pathogen genomic data and associated metadata within a trusted environment. By bringing analytical tools to the data rather than requiring their repeated transfer across independent platforms, this approach facilitates scientific analysis while preserving data governance and contributor recognition. The article should therefore distinguish between data accessibility, data redistribution and analytical interoperability, rather than treating unrestricted data movement as the only viable model for scientific collaboration.

GISAID provides an important counterexample to the suggestion that registration requirements are inherently incompatible with timely sharing, genomic surveillance and computational analysis as shown by the long history of being the platform of choice at the forefront of outbreaks.

As a result of this oversight, the authors' suggestion that PABS registration requirements could lead to the progressive removal of pathogen sequences from the global scientific ecosystem (page 6) is speculative.

GISAID is dedicated to the curation and maintenance of a comprehensive global resource of pathogen genomic sequences and associated metadata, providing an established infrastructure for timely data sharing, genomic surveillance and outbreak response.

Its success was particularly evident during the COVID-19 pandemic, when, across 222 countries and territories, the average country shared approximately 90% of its SARS-CoV-2 sequences exclusively through GISAID. Far from contributing to the fragmentation or loss of pathogen data, GISAID has facilitated their global availability at an unprecedented scale. Today, GISAID accounts for the overwhelming majority of available sequences for several major viral pathogens, including influenza, SARS-CoV-2, RSV, dengue and mpox.

These observations challenge the assumption that governed access necessarily discourages data sharing or diminishes the global pathogen knowledge base. Rather, they demonstrate that a trusted data-sharing framework that recognizes contributors' rights, supports data stewardship and provides integrated analytical capabilities can successfully incentivize widespread and sustained sharing of pathogen genomic data. The more relevant concern for pandemic preparedness is therefore not registration itself, but whether the proposed PABS framework preserves the trust, incentives and infrastructure that have enabled timely and comprehensive pathogen data sharing

Figure 1 illustrates the movement of data between several biological databases but does not establish that this distributed architecture is necessary, optimal or efficient for pathogen genomic surveillance. The authors suggest that “all these data flows are necessary to answer diverse biological questions”, citing the metabolic potential of a novel virus as an example of information that cannot be obtained from nucleotide sequences alone (page 4). This argument is questionable on several grounds.

First, characterizing viral metabolic potential is hardly among the most immediate priorities during pandemic response, where rapid identification, genomic characterization, evolutionary surveillance and assessment of emerging variants are generally more relevant. 

Second, many of the annotations illustrated, including protein translation, functional prediction and structural modelling, can be derived computationally from nucleotide sequences using established bioinformatics tools, without requiring sequential data transfers between independently maintained databases. Indeed, the databases depicted largely represent specialized resources that apply different analytical methods to sequence data, with additional dependencies on database synchronization, annotation and update cycles that may introduce delays.

Third, Figure 1 oversimplifies the nature of these database interactions by presenting automated record exchange, cross-referencing, functional annotation and computational prediction as components of a broadly interconnected workflow, without adequately distinguishing their respective computational requirements, dependencies and update schedules. Importantly, the existence of automated connections between databases does not establish that all the illustrated operations occur automatically for every deposited sequence, nor that such an architecture represents the most efficient approach for time-sensitive pathogen surveillance.

A more efficient alternative for time-sensitive pathogen surveillance is to integrate these analytical capabilities directly within the primary data-sharing platform, allowing sequences to be processed, annotated and analyzed as they become available. This is the approach adopted by GISAID, where curated genomic data, associated metadata and integrated analytical tools are brought together within a trusted environment. By enabling analysis where the data resides, this architecture reduces unnecessary data movement, minimizes dependencies on external database updates and facilitates more timely access to analytical insights while maintaining established data-governance arrangements. The omission of this established approach from the discussion presents an incomplete and potentially misleading picture of the technical architectures available for pathogen genomic analysis and interoperability.

The authors should explain why their proposed architecture which requires data to be repeatedly transferred across independent platforms, is preferable to existing integrated environments.

As with Figure 1, the authors define interoperability in terms of databases automatically exchanging, linking and building upon one another's data Interoperability (page 3-4). For pathogen surveillance, the objective should not necessarily be to maximize the movement of data between databases.

Instead, it should be to maximize the ability of researchers to derive reliable and actionable insights from available data while respecting appropriate governance requirements.

This distinction is especially relevant to GISAID's approach, where integrated analytical services can reduce the need to move data out of its governed environment.

Although the authors acknowledge the importance of attribution and addressing inequities in scientific research, their proposed mechanisms for recognizing data generators focus predominantly on scientific publications and citation practices (page 8). Downstream monitoring of publications, patents and commercial products may complement benefit-sharing arrangements, but it should not be presented as an adequate substitute for governance mechanisms operating at the point of data access and use.

Retrospective monitoring may be incomplete, particularly where data contribute indirectly to research or commercial development. A more comprehensive framework should consider how attribution, accountability and traceability can be supported throughout the research lifecycle. This overlooks the broader rights and legitimate interests of those who generate, curate and share pathogen genomic data, including appropriate recognition beyond publications, accountability for downstream data use, and safeguards against unacknowledged redistribution or exploitation. Publication attribution alone does not adequately address these concerns, particularly when sequence data are increasingly used in automated analyses, computational models and commercial research that may not result in conventional scientific publications.

GISAID's data-governance framework recognizes that incentivizing timely data sharing requires more than retrospective citation; it also requires clear conditions governing data access and use, contributor recognition and accountability. A meaningful discussion of equitable pathogen data sharing should therefore consider how these interests are protected throughout the data-use lifecycle, rather than relying primarily on downstream publication monitoring.

The authors propose standardized PABS Terms of Use that could apply across databases without disrupting existing interoperability (page 8). The proposal to adopt standardized Terms of Use is reasonable but insufficiently developed. The authors should explain how compliance, accountability and enforcement would operate in the absence of mechanisms that identify users and establish applicable obligations. Standardized terms and registered access cannot be mutually exclusive; registration provides the means to audit data use and is required to support the practical implementation of governance arrangements intended to protect contributors and ensure responsible data use.