Escaping from the 1600s: Advancing FAIR scientific knowledge with reborn articles

Text-based literature remains the primary expression of scientific knowledge. Since the first scientific article published in 1665, we have managed the switch from physically printed articles to PDFs, but nothing more. While PDF publications can be easily shared electronically, they remain unstructured text-based documents that machines cannot easily interpret (i.e., they are not machine-reusable). This limits our ability to use digital support tools to efficiently extract and organize knowledge from scientific articles. Rather, to reuse most scientific results (e.g., for synthesis research), we must first extract them from articles and organize them into databases, which is time consuming and prone to error. 

Here, we present reborn articles, which offer a novel approach to producing scientific knowledge. By integrating with programming languages commonly used for data analysis, like R and Python, reborn articles allow researchers to produce scientific results in a machine-reusable format from the outset. This means subsequent data users can download the results of a reborn article as a CSV file with just a click of a button and bypass post-publication data extraction. To support the production, publication, and reuse of reborn article data, we developed ORKG reborn, a FAIR knowledge online infrastructure. 

Using an ecological dataset, we showcase the production of a reborn article, and its impact on knowledge integration and synthesis. Building on the author’s original data analyses conducted in R, we developed an accompanying R script to produce machine-reusable descriptions of the original statistical models that were automatically harvested by ORKG reborn, eliminating manual data entry. We envision that the use of programming languages, like R, to facilitate the production of machine-reusable scientific knowledge could feasibly be streamlined into existing FAIR data management requirements that are already implemented by many academic publishers. Broad adoption of the approach across research communities could transform the way we share and synthesize scientific knowledge.