From crisis to opportunity: Linking research on animal personality and reproducibility

生命科学正面临着可重复性危机。最初，这场危机是由于在不同时间和地点重复调查结果的单一令人震惊的失败而产生的。如今，系统研究表明，不可复制研究的普遍性实际上确实超过了50%。从一个相当愤世嫉俗的角度来看，费特的实验室定律“永远不要复制一个成功的实验”因此具有了全新的含义。在这方面，动物研究受到了特别的审查，因为在研究伦理和社会影响方面的利害关系都很高。为了解决这一问题，必须确定重现性差的来源，并消除这些故障。从跨学科的角度来看，我们在这里开发了一种全新的方法来解决重现性差的问题，并根据生物变异增加实验结果的力量。更具体地说，通过行为表型研究的例子，我们建议整合进化生物学的知识，并把稳定的行为个体间的差异，所谓的“动物个性”，作为一个迄今为止完全被忽视的动物实验中的变异来源。我们的项目计划包括四个工作包（WP），其中WP I-III为WP IV中再现性的原理验证研究建立了必要的实验基础。简而言之，WP I旨在澄清如何可靠地评估实验室小鼠行为中的暂时性差异，而WP II则研究人格差异是否可以解释动物实验中广泛宣称的“残余变异”的相当大比例。它还打算验证小鼠个性的“预测标记”，这些标记可用于以最小的努力识别不同的个性类型。WP III研究了动物的个性和示范性治疗之间的潜在相互作用，从而阐明个性是否是一个因素，可以严重影响动物研究的结果。最后，WP IV旨在将动物个性的概念与再现性研究联系起来，并提出了一种在实验设计中考虑个性差异的新策略，以提高统计功效，从而提高再现性。研究结果有望对动物个性的研究以及基础和临床前动物研究中动物实验的规划和实施产生影响。

Life sciences are facing a reproducibility crisis. Originally, the crisis was born out of single alarming failures to reproduce findings at different times and locations. Nowadays, systematic studies indicate that the prevalence of irreproducible research does in fact exceed 50%. Viewed from a rather cynical perspective, Fett’s law of the lab “Never replicate a successful experiment” has thus taken on a completely new meaning. In this respect, animal research has come under particular scrutiny, as the stakes are high in terms of both research ethics and societal impact. To counteract this, it is essential to identify sources of poor reproducibility as well as to iron out these failures. Taking an interdisciplinary perspective, we here develop an entirely new approach to tackle poor reproducibility and increase the power of experimental findings in light of biological variation. More specifically, by the example of behavioural phenotyping studies, we propose to integrate knowledge from evolutionary biology and put stable behavioural inter-individual differences, so called “animal personalities”, into focus as a so far completely neglected source of variation in animal experiments. Our project plan comprises four work packages (WPs), from which WPs I-III build the essential experimental basis for a proof-of-principle study on reproducibility in WP IV. Briefly, WP I intends to clarify how temporal consistencies in the behaviour of laboratory mice can be assessed reliably, while WP II investigates whether personality differences explain a considerable proportion of the widely declared “residual variation” in animal experiments. It furthermore intends to validate “predictive markers” of mouse personality that can be used to identify different personality types with minimal efforts. WP III examines potential interactions between the animals’ personality and exemplary treatments, thereby elucidating whether personality is a factor that can critically affect the outcome of an animal study. Lastly, WP IV aims at linking the concept of animal personality to reproducibility research and propose a novel strategy for considering personality differences in the experimental design to improve the statistical power, and hence the reproducibility. Findings are expected to have implications for both the study of animal personality and the planning and conduct of animal experiments in basic and preclinical animal research.