Safeguarding Scientific Accuracy in a Rapidly Changing World

Recently, I have become more aware of the fact that scientific mores are changing when it comes to publishing results or theories. After a period of increased rigor (no more “unpublished results” in papers, please), the rules seem to be relaxing somewhat. One example: referring to papers that have not undergone peer review (yet). Preprint repositories, such as bioRxiv, were around from the nineties but they really multiplied and got intensely used from about 2010 onward. To be honest, I have one manuscript, reflecting unfinished business, in such a repository myself. It has been there since 2013 and was, to my surprise, quoted on several occasions. I would still hesitate to do so myself and I will explain why.

Current Biology is an excellent journal with, for example, high-quality reviews that are a joy to read. Recently, I encountered a great paper about the evolutionary arms race with defense systems and anti-defense systems constantly evolving in prokaryotes on the one hand and phages and plasmids on the other [1]. While perusing it, I stumbled upon a surprising fact: more than 5% (6 out of 103) of the references were to non-peer reviewed manuscripts on bioRxiv (deposited in 2023 and 2024). This example illustrates a pronounced development over the last decade: the use of preprints in such manner is now widely accepted [2]. The positive aspects seem clear. Researchers can get rapid feedback and community review, as well as hasten dissemination, and all this in open access form. In mathematics, meeting abstracts and preprints are even standard and encouraged by major journals (the actual publication process is almost an afterthought). On the minus side stands the possible loss of one of the safeguards of scientific quality: peer review. However, on balance, the fact that the preprint is publicly available and is clearly identified as still having to undergo peer scrutiny, does help in neutralizing the potential harm of sub-standard research.

Indeed, the quality of the process of peer review itself could be under pressure. At the risk of anecdotal misrepresentation, I have the strong impression of seeing more mistakes in published articles nowadays as compared to a few decades before. This could just reflect a greater “publish or perish” pressure, as well as ever more complicated techniques and algorithms associated with massively increased datasets. However, rapidly evolving new areas of expertise, time pressures, and the “courtesy” basis of relatively unrewarded reviewing seem to be taking their toll. I will give just one example of an error that should have been spotted before publication, having to do with one of my own hobbyhorses, the mitochondrial electron transport chain (ETC). A highly interesting PNAS perspective looks at how modern-day ETCs could tell us something about how the earliest stages of evolution might have harnessed the potential available in electrochemical gradients [3]. Alas, in their figure 2 the authors make an unfortunate mistake, by having the reduced carrier Q (carrying two electrons and two protons) giving off two protons to the intermembrane space (IMS) before contacting complex III which is then shown as “pumping” another four protons into the IMS (the same kind of mistake is depicted in the chloroplast ETC). But those two protons coming from Q are part of the calculation for complex III! Actually, complex III does not transport protons the way complexes I and IV do. It oxidizes two QH₂ at the IMS side, releasing four protons, and reduces one Q at the matrix side taking up two protons (the complicated two-step “Q-cycle”). The net effect is the transport of four protons for the oxidation of one QH₂, but without a proton channel inside the complex. This might be the oldest way of coupling a gradient to electron transport, as the authors also deduce and correctly depict in their next figure. However, the previous depiction obscured the fact that this old mechanism is still in operation [3]. For further details, see [4].

In an era where many of the most powerful people have rather strained relations with the truth and scientific facts, accurate fact checking and peer control in science is more important than ever, so let's institute mechanisms to further strengthen it.