Short abstract:

Would you trust a Photoshopped gel electrophoresis image? What are the dynamics of trust building in a world of recursive fraud? Two different faces of reproducibility are highlighted in a historical analysis of the digitization of gel electrophoresis images.

Long abstract:

The Voinnet affair (around 2015) recounts the story of a rising biologist forced to retract a hefty number of his publications in retrospect, upon allegations of scientific misconduct concerning gel electrophoresis images. Actually, for the last ten years, within molecular life sciences, the reproducibility crisis discourse reflects a crisis of trust in scientific digital images.

Beyond the contentious perception of "questionable research practices" associated with a digital turn in the production of images, we highlight transformations of gel electrophoresis as a family of experimental techniques from the 1980s to the 2000s. Our aim is to analyze the evolving epistemic status of electrophoresis images.

To this end, we discuss the advent of two tiers of gel electrophoresis, each with different standardization procedures, different epistemic statuses of generated images and different ways of generating (dis)trust in images. The first tier is characterized by specialized devices processing images as quantitative data. The second tier is described as a routine technique making use of image as qualitative "virtual witnessing".

The difference between these two tiers is particularly apparent in the ways images are processed, even though both tiers involve image digitization. Our account thus highlights different views on reproducibility within the two tiers. Comparability of images is insisted upon in the first tier while traceability is expected in the second tier. In the second tier, Photoshop entails distrust, whereas a form of collective disciplining in the first tier implies a general sentiment of trust.