The current landscape of author guidelines in chemistry through the lens of research data sharing

Introduction

When asked whether they have read an article in the last year, e.g., in a graduate course on scientific writing, every researcher will raise their hand. When inquiring who then was satisfied with the completeness of information provided, some hands are lowered. Only few remain raised when continuing to ask who was able to find and access the actual underlying research data so that they can reproduce the work or build upon it.

Publishing the underlying research data alongside journal articles is widely considered good scientific practice [1, 2]. In simplest terms, providing these data ensures transparency and is crucial for others to be able to build upon the published results. However, sharing data is not yet as common as sharing results, and more sharing requires a cultural shift [3]. As the primary method of communicating research results, journals garner a great amount of influence on community behavior. This includes how researchers handle the data on which their published results are based. For example, data availability statements included in the back matter of a manuscript not only provide information on how a reader may find and access the data—requiring such statements also encourages the author to think about whether and how they will make their research data available to others in the first place.

Journals are able to guide researchers in making their data discoverable by recommending that the data underlying the published research results should be deposited in an appropriate recommended repository. Ideally, journals and their publishers place these recommendations in their author guidelines and data policies.

For several decades, publishers and journals have been providing instructions to scientists in the form of author guidelines, with the aim that articles published in a journal have a consistent style, length, or, e.g., a maximum number of figures. Predominantly chemistry journals have also included specific recommendations and directives concerning chemical information. As of 1953, The Analyst provided instructions [4] on abbreviations and noted that units and formulae may only be used in addition to the names of elements and compounds. Guidelines from the 1970s in Chemical Communications [5] required authors to follow symbols or nomenclature as established by, e.g., the International Union of Pure and Applied Chemistry (IUPAC).

Early examples of guidelines specifically addressing data, such as deposition into the Cambridge Structural Database (CSD) with the Cambridge Crystallographic Data Centre (CCDC) for small structures as required by Angewandte Chemie in 1992, were less common [6]. More detailed instructions also began to appear, including how to list physical measurement data such as boiling and melting point as well as spectroscopic data within the text [7].

Data repositories have also provided guidance for research data. For example, the Protein Data Bank (PDB) for protein structures provides instructions for journals [8] on how deposition to PDB should be requested in their respective author guidelines and the CCDC has worked with many publishers to improve the guidance for crystal structure deposition [9]. Therefore, sharing data along with research results has started to go hand in hand.

While titled “author” guidelines, these are important for different stakeholders, e.g., for editorial staff and reviewers, who check and can request adherence to the guidelines, including the scientific and quality aspects of the research data. Fig. 1 illustrates the relevant stakeholders with respect to author guidelines, as well as aspects covered in author guidelines and their subsequent impact on the publication of research outputs.

Fig. 1:

Author guidelines are relevant for different stakeholders who can obtain information on topics and scope of a journal to present, review or edit research outputs compliant with a journal’s guideline.

An important aspect in the publishing of today and the future is an increased level of machine-readability to support finding, navigating, and analyzing the publication space, a concept in line with the FAIR Data Principles [10]. This requires unique and persistent identifiers (PIDs) to exactly pinpoint digital items in that (vast) space. The most prominent example is the Digital Object Identifier (DOI) [11], the use of which began at the end of the last century and is now commonplace in academic publishing. The ORCID iD [12] identifies researchers (i.e., authors) and is on its way to reach a similar uptake as the DOI in just 10 years. Similarly, the Research Organization Registry (ROR) may be used as an identifier for institutions [13].

Several efforts are underway to improve the infrastructure, processes, and uptake of research data sharing in chemistry. To set the stage, a workshop on “FAIR Publishing Guidelines for Spectral Data and Chemical Structures” was organized by some of the authors (LRM and VFS) and held on March 29–30, 2019 in Orlando, FL (USA). Among other topics [14, 15], a survey and checklist on Chemistry Journal Data Submission and Sharing Policies was presented [16]. This checklist was created for a Research Data Alliance (RDA) Publisher Forum presentation [17], a broad community venue for discussion on RDM efforts across domains and publishing, such as the respective Chemistry Research Data Interest Group [18] and the Data Policy Standardisation and Implementation Interest Group [19].

The NFDI4Chem [20] consortium aims to develop infrastructure to support chemists in making their research data findable, available, interoperable, and re-usable, or FAIR (see FAIR Data Principles [10]). The adoption of the FAIR principles requires a shift throughout the community, where research data is readily shared on established and appropriate repositories and standards on content, metadata, and data formats are adhered to. Open research practices, such as the use of research registries and preprint servers, are encouraged to increase transparency throughout the community.

IUPAC is leading a project on FAIR chemical data standards under the global WorldFAIR Initiative led by CODATA and the Research Data Alliance to advance FAIR data sharing. The goal of WorldFAIR Chemistry is to review gaps in FAIR implementation of chemical data and enable the development and adoption of chemical data standards in research workflows to enable downstream data reuse [21, 22].

In this article, we are building on the initial chemistry data checklist work [16], and present the results of a large-scale analysis of author guidelines from several publishers and journals active in chemistry research. The main aim is to provide an overview of how well the publishing landscape currently supports criteria needed to enable FAIR data sharing and how it supports an interoperable data ecosystem and considers FAIR digital research data as first-class research output in publications.

Methodology

To assess the current journal requirements for research data in chemistry, the author guidelines and policies from 42 journals, released by 13 publishers, were reviewed. The selected journals were sourced based on where the various chemists in the NFDI4Chem consortium commonly publish or source information and are considered to be common and established and therefore representative of the prevailing publishing landscape. The full list of journals are available as table JournalGuidelinesDataCollection in the supplementary data published on RADAR4Chem [23].

Through the lens of research data, our review criteria intend to survey how journals support authors in adhering to the FAIR Data principles and Open Science [24] practices during the publication process. Therefore, the review incorporated categories which focus on the data underlying the article, such as data publication through deposition into repositories. It also included aspects of the article (prior publication on preprint servers), its metadata (use of ORCID iDs), and the backmatter (implementation of data availability statements). Many of these criteria, e.g., minimum information standards and publishing the underlying data to journal articles in subject-specific repositories, reflect efforts being put forth by NFDI4Chem [20]. While many also reflect aspects of the FAIR Data principles, the publication process includes aspects that go beyond the scope of FAIR. Table 1 summarizes these criteria.

A team consisting of three scientists with expertise in chemistry and data management divided the journals up amongst themselves, grouped by the associated publisher. Each team member reviewed their groups of journals individually, then gathered feedback in group discussions. Furthermore, the reviews were cross-checked by the team members to ensure consistent rating and increase objectivity. The first round of reviews were carried out in September 2021, the results of which were presented at the Editors4Chem Workshop organized jointly by NFDI4Chem and IUPAC in November 2021. Following this workshop, a second review took place in January of 2022 to correct possible inconsistencies noted by the workshop participants and to incorporate changes to guidelines that occurred after the workshop. The results reported here and depicted in Fig. 2 pertain to the second and final review.

Fig. 2:

Overview of requirements and recommendations for FAIR and open data sharing in author guidelines. This circular plot shows the selected categories (outer legend) and criteria (inner legend) of the guidelines. Segments were classified into four categories: from “required” (dark green), “recommended” (green), “accepted” (light green) to “not mentioned” (red) pertaining to the article (upper right) and digital research data. E.g., the ORCID iD for the submitting author was required by 63 % of journals, recommended by another 10 %, and not mentioned by the remaining 27 %.

The selected author guidelines were reviewed for each category and assessed as to their level of requirement as shown in Fig. 2. The color scale indicates which criteria were determined to align with FAIR Data and Open Science practices from most (dark green) to least (red). Specifically, dark green signifies the portion of the journals that required a certain practice or stated that a practice is mandatory, light green signifies the portion that recommended a practice, also described as a practice authors are requested to follow or which is supported by the journal’s guideline. Yellow-green indicates the portion which stated that they accepted a practice, and red the portion that did not mention a practice at all. The absence of categories such as “is not recommended” or “is not accepted” is intentional, as no criteria fell into these categories.

No further input was collected from the journals aside from their online, publicly available author guidelines. Publisher guidelines were included in this evaluation only when referenced specifically within a journal’s guideline. The term “author guidelines” is interpreted to be inclusive, as linked or accompanying “data policies” or “publishing policies” were also considered.

Since author guidelines are continuously evolving, we collected snapshots of the web pages to document the state at the beginning of 2022 and provide these as supplementary information guideline_screenshots published on RADAR4Chem [23]. Perceptible changes of author guidelines towards FAIR research data, compared to before the Editors4Chem Workshop in November 2021 was held, were also noted and documented.

The review results were collected in the spreadsheet [23] JournalGuidelinesDataCollection, summarized and visualized via the accompanying Jupyter notebook using Python and the statistics language R.

As mentioned, this review aims to gain an understanding of the current state regarding research data in academic publishing in chemistry. Therefore, the data were evaluated to give percentages of journals that fell into the respective criteria for each category. This review does not intend to single out or rate individual journals or publishers according to the criteria.

Results and discussion

In the following we summarize and discuss the results of the survey evaluation data. Fig. 2 visualizes the results. The code to reproduce the figure is available as a Jupyter notebook online and also as part of the supplementary data, see the README in the supplementary data [23].

Outlook

This survey of the author guideline landscape through a chemistry lens has shown that publishers and journals are starting to include further aspects of research data sharing in their guidelines. The example of (missing) requirements on machine-readable chemical structures, minimum information standards, and open analytical data format standards shows that there is plenty of room for improvement, not only by publishers but also by infrastructure and resource providers. Promising developments were observed in terms of data findability by journals and publishers calling for the submission of data availability statements. Similarly, recommendations on research repositories increase data accessibility.

Many of these improvements will be supported by efforts such as the NFDI4Chem [3, 20] and IUPAC [49] establishing further repositories for data produced in the different sub-disciplines of chemistry, standardizing metadata and data formats for a majority of the analytical data coming out of chemical research, and spreading awareness through information and training sessions. Further guidance for implementing chemical data guidelines from policy to practice from IUPAC is under development in the WorldFAIR Chemistry project [21, 22], drawing from community practices in chemistry and more broadly [49]. Activities include FAIR analysis methods, recipes for incorporating chemistry standards, and resources for working with FAIR-enabled chemistry data [21].

As a continuum, FAIR has the unique ability to draw the research community and all of its varied stakeholders to its principles, with gradual incorporation of its recommendations in a stepwise manner that makes implementation easy. All it needs to succeed is a community that desires to be able to understand and re-use the results and knowledge of chemical research. Let us work towards a future where the participants’ hands in the room remain raised when asked “who read an article, went to the corresponding data, and was able to base new science upon it?”

Supplementary information

The supplementary data package [23] on RADAR4Chem contains a README with further instructions and the Jupyter notebook to summarize the guideline survey to reproduce Fig. 2. The dataset also includes screenshots of the author guidelines. Finally, the notebook rendered to HTML and PDF is included.