Ensuring Reproducibility and Transparency in Experimental Methods and Data Analysis: A Guideline for Authors Submitting to Horticulturae

Simple Summary

This editorial outlines explicit recommendations for authors submitting to Horticulturae to ensure repeatability/reproducibility and transparency in their experimental methods and data analysis. It emphasizes the importance of detailed reporting, including software versions, accessible data sources, and public repository submissions, to encourage scientific rigor and facilitate peer review.

As members of the Editorial Board and Academic Editors of Horticulturae, we are committed to maintaining the highest standards of scientific rigor and integrity in the articles published in our journal. One critical aspect of this commitment lies in ensuring that the research we disseminate is reproducible and transparent. Repeatability or reproducibility—the ability for independent researchers to replicate a study using the descriptions provided—is fundamental to the advancement of science, enabling validation, fostering trust, and supporting further discovery.

Over recent years, we have observed recurring challenges regarding the description of the materials and methods and data analysis protocols. Often, manuscripts lack the necessary level of detail and clarity required to fully replicate experiments or computational analyses. This issue jeopardizes the reliability and impact of the reported findings and complicates the peer review process.

To assist our global community of authors and to uphold the quality of publications, we provide the following guidance aimed at enhancing reproducibility and transparency.

1. Detailed Reporting of Experimental Procedures

It is imperative that the Materials and Methods section include comprehensive descriptions of every step in the experimental workflow. This encompasses sample origin, conservation, preparation, growth conditions, treatments applied, instrumentation parameters, data collection, and analytical methodologies.

Biological materials (isolates, strains, cell lines, insect specimens, and other voucher material) must be deposited in recognized biological resource centers or culture collections, with appropriate accession numbers provided.

Authors should ensure that the description is sufficient for an independent researcher to replicate the study without requiring additional information. To this end, authors must indicate the actual manufacturers of all materials used, not local suppliers.

2. Specification of Software and Tools

When data analysis involves software tools or computational pipelines, authors must specify the software name, exact version used, and any relevant settings or parameters.

We strongly encourage the use of analysis platforms that support the development and execution of scripts, as scripted workflows inherently promote reproducibility when applied to the same datasets. Examples of widely adopted scripting-based tools include R [1] and Python [2], which enable precise control over data processing and statistical analyses.

It is important to acknowledge the significant limitations often encountered when relying on spreadsheet software, such as Microsoft Excel, for complex scientific data analysis [3,4]. Spreadsheets often struggle with scalability and performance, becoming slow or unresponsive with large datasets and sometimes crashing. Manual data entry and manipulation increase the risk of human errors, data corruption, and inconsistencies, while version management and change tracking features are limited, making it challenging to ensure data accuracy.

Additionally, spreadsheets are ill-suited for handling unstructured or semi-structured data and face difficulties in analyzing complex relationships within datasets. Their limited automation capabilities mean many tasks are repetitive and prone to manual errors, which reduces reproducibility. Collaboration on spreadsheets is also challenging, as typically only one user can work on a file at a time, and real-time updates are rarely supported natively. Furthermore, the proliferation of multiple spreadsheet files creates significant security risks and complicates data access management.

In contrast, scripting-based analysis offers numerous advantages that address these limitations. Scripting languages can efficiently handle massive datasets without performance degradation, making them ideal for large-scale data analyses. They provide superior flexibility for complex data manipulations, transformations, and analyses that are unattainable in spreadsheets.

Scripts offer precise control over data manipulation and analysis, supporting consistent and reproducible results, with the added benefit of version control and change tracking via code repositories. Automation is enhanced, enabling end-to-end workflows from data import and cleaning to advanced statistical modeling and visualization, which reduces manual effort and risk of errors.

Moreover, scripting facilitates integration with various data sources and supports collaboration through shared, version-controlled repositories like GitHub [5] or GitLab [6]. Advanced functionalities, such as custom functions, statistical models, and interactive dashboards, can be developed to further extend analysis capabilities.

The

Table 1. Comparison between spreadsheets and modern data tools.

Dimension of Comparison	Spreadsheets	Modern Tools
Scalability	Limited: Capped at 1,048,576 rows. Performance degrades with file size.	Highly Scalable: Built to handle massive datasets (petabytes). Performance is consistent.
Reproducibility	Manual: Reliance on point-and-click actions. No auditable log of steps.	Automated: Code-based workflows are repeatable, version-controlled, and auditable.
Data Integrity	Poor: Lack of structured data types, relational integrity, and transactional support.	Robust: Databases enforce data types and relationships. Code provides a clear validation and cleaning step.
Collaboration	Fragile: File-based model leads to version confusion and conflicting changes.	Robust: Centralized databases and version control systems (e.g., Git) are built for simultaneous, auditable teamwork.
Automation	Manual: Repetitive tasks require manual effort or complex, hard-to-debug macros (VBA).	Automated: Scripts and data pipelines can automate entire workflows from data ingestion to final reporting.
Advanced Analytics	Limited: The Analysis ToolPak is outdated and rudimentary. Not built for machine learning.	Powerful: Supports machine learning, statistical modeling, and advanced data visualization.

provides a comprehensive comparison of spreadsheets versus modern data analysis workflows.

3. Transparency in Reference Databases and Datasets

Authors should clearly cite the reference databases and datasets utilized in their analyses, including version numbers and access dates when applicable.

For microbiome analyses, it is essential to report which taxonomic reference databases were used for taxonomy assignment, for example SILVA, commonly used for bacterial taxa [7], or UNITE for fungal taxa [8]. In other research fields, additional reference systems, such as BOLD (Barcode of Life Data System) for animal DNA barcoding [9], or GenBank, maintained by the National Center for Biotechnology Information (NCBI), for general sequence repositories [10], should also be properly cited.

It is equally important to specify whether results or identification have been curated or reviewed by experts in the field, as reference databases contain errors and taxonomies could be outdated. This practice helps readers and reviewers critically interpret the results and assess their reliability.

Where possible, links to publicly accessible repositories or databases should be included. This transparency supports the verification and reuse of data in future studies.

4. Avoidance of Inaccessible or Language-Restricted Resources

Manuscripts sometimes reference online platforms, portals, or databases that are restricted by language or geographic access, which presents a barrier to verification by the international scientific community. Authors should refrain from citing such resources without providing detailed methodologies or alternative access options available to all readers and reviewers.

It has been frequently observed that authors employ certain analytical platforms that, while described in English-language publications, are accessible primarily through interfaces in languages not widely understood by the global community. An illustrative example is the Majorbio Cloud Platform [11], which is referenced in many manuscripts. Although a detailed publication describing the platform is available in English [12], the platform itself operates mainly in Chinese, posing significant challenges for reviewers and researchers outside this linguistic community to verify or reproduce analyses.

To maintain transparency and reproducibility, authors who use such platforms must provide extensive methodological details alongside their manuscripts. This includes steps of data processing, algorithms applied, parameter settings, and any intermediate results critical to understanding and replicating the analyses.

It is essential for authors handling sequencing data to deposit raw data in publicly accessible repositories, such as the NCBI Sequence Read Archive (SRA) [13], and to reference associated BioProject accessions clearly. In addition, all newly generated sequences must be deposited in publicly accessible repositories, such as GenBank, the European Molecular Biology Laboratory–European Nucleotide Archive (EMBL-ENA) [14], or the DNA Data Bank of Japan (DDBJ) [15], with accession numbers clearly reported in the manuscript. This practice is vital for open science, enabling the validation and reuse of datasets by other researchers globally. Failure to do so not only limits the reproducibility of the study but also diminishes its contribution to the broader scientific community.

5. Provision of Sufficient Methodological Detail for Statistical Analyses

Statistical analyses and data-processing methods must be articulated with clarity, including the tests applied, assumptions considered, corrections for multiple comparisons, and criteria for significance. This level of detail is crucial to evaluate the robustness of results.

6. Supplementary Materials and Data Sharing

To further enhance transparency, authors are encouraged to include supplementary files containing raw data, detailed protocols, and code repositories where appropriate. Open sharing of such materials aligns with modern scientific best practices and facilitates the reproducibility of published research.

7. The Role of Authors, Reviewers, and Editors

We recognize that reproducibility is a shared responsibility. Authors must ensure completeness and clarity in their reporting; reviewers play a critical role in evaluating the adequacy of methodological descriptions; and editors help enforce standards and provide guidance. Importantly, all interactions between authors and referees during the peer-review process should be conducted in English. This practice ensures that editors can accurately interpret the review discussions and assess the work performed during the revision process, preserving the integrity and clarity of communication necessary for efficient editorial decisions.

Together, these efforts contribute to the robustness of the scientific record.

In conclusion, we urge all authors submitting to Horticulturae to carefully consider the reproducibility and transparency of their work. Detailed and accessible reporting not only strengthens individual manuscripts but enriches the collective knowledge base, advancing horticultural science globally.

We appreciate the efforts of our authors in adhering to these guidelines and thank the community for their continued commitment to high-quality research.