The rise of open data practices among bioscientists at the University of Edinburgh

TL;DR

This study shows that bioscientists at the University of Edinburgh have increasingly adopted open data practices, with more research data being shared over time.

Contribution

The study provides a longitudinal analysis of open data practices in biosciences and evaluates the performance of the ODDPub tool for detecting data-sharing in publications.

Findings

The percentage of publications sharing all relevant data rose from 7% in 2014 to 45% in 2023.

Genomic data was shared more frequently than image data or data on human subjects.

Publications with data availability statements or preprints showed better data-sharing completeness.

Abstract

Open science promotes the accessibility of scientific research and data, emphasising transparency, reproducibility, and collaboration. This study assesses the Openness and FAIR (Findable, Accessible, Interoperable, and Reusable) aspects of data-sharing practices within the biosciences at the University of Edinburgh from 2014 to 2023. We analysed 555 research papers across biotechnology, regenerative medicine, infectious diseases, and non-communicable diseases. Our scoring system evaluated data completeness, reusability, accessibility, and licensing, finding a progressive shift towards better data-sharing practices. The fraction of publications that share all relevant data increased significantly, from 7% in 2014 to 45% in 2023. Data involving genomic sequences were shared more frequently than image data or data on human subjects or samples. The presence of data availability statement (DAS) or preprint sharing correlated with more and better data sharing, particularly in terms of completeness. We discuss local and systemic factors underlying the current and future Open data sharing. Evaluating the automated ODDPub (Open Data Detection in Publications) tool on this manually-scored dataset demonstrated high specificity in identifying cases where no data was shared. ODDPub sensitivity improved with better documentation in the DAS. This positive trend highlights improvements in data-sharing, advocating for continued advances and addressing challenges with data types and documentation.

Figures