FACT and FAIR with Big Data allows objectivity in science: The view of crystallography

Abstract

A publication is an important narrative of the work done and interpretations made by researchers securing a scientific discovery. As The Royal Society neatly states though, "Nullius in verba" ("Take nobody's word for it"), whereby the role of the underpinning data is paramount. Therefore, the objectivity that preserving that data within the article provides is due to readers being able to check the calculation decisions of the authors. But how to achieve full data archiving? This is the raw data archiving challenge, in size and need for correct metadata. Processed diffraction data and final derived molecular coordinates archiving in crystallography have achieved an exemplary state of the art relative to most fields. One can credit IUCr with developing exemplary peer review procedures, of narrative, underpinning structure factors and coordinate data and validation report, through its checkcif development and submission system introduced for Acta Cryst. C and subsequently developed for its other chemistry journals. The crystallographic databases likewise have achieved amazing success and sustainability these last 50 years or so. The wider science data scene is celebrating the FAIR data accord, namely, that data be Findable, Accessible, Interoperable, and Reusable [Wilkinson et al., "Comment: The FAIR guiding principles for scientific data management and stewardship," Sci. Data 3, 160018 (2016)]. Some social scientists also emphasize more than FAIR being needed, the data should be "FACT," which is an acronym meaning Fair, Accurate, Confidential, and Transparent [van der Aalst et al., "Responsible data science," Bus Inf. Syst. Eng. 59(5), 311-313 (2017)], this being the issue of ensuring reproducibility not just reusability. (Confidentiality of data not likely being relevant to our data obviously.) Acta Cryst. B, C, E, and IUCrData are the closest I know to being both FACT and FAIR where I repeat for due emphasis: the narrative, the automatic "general" validation checks, and the underpinning data are checked thoroughly by subject specialists (i.e., the specialist referees). IUCr Journals are also the best that I know of for encouraging and then expediting the citation of the DOI for a raw diffraction dataset in a publication; examples can be found in IUCrJ, Acta Cryst D, and Acta Cryst F. The wish for a checkcif for raw diffraction data has been championed by the IUCr Diffraction Data Deposition Working Group and its successor, the IUCr Committee on Data.

FIG. 1.

The Big Data pyramid labeled for typical crystallography dataset file sizes for raw diffraction images at the base, the processed structure factors in the middle, and the derived atomic coordinates and their respective atomic displacement parameters in the top of the pyramid.