Improving animal-to-human translation using literature

Mouse

Animal experiments in Europe are subject to strict legal regulations that require an ethical assessment, including a harm-benefit analysis. This means that medical research involving the use of animals can only be ethically defended if the results of these tests contribute to the well-being of humans or animals. For human pharmaceutical development, we are thus interested in the correspondence between results from animal and human studies.

Our preceding scoping review has shown that this correspondence varies from 0 per cent to 100 per cent, but we have not yet been able to reliably determine which factors predict successful translation. Thus, we continue this work, with multiple projects, in the AI-aided Knowledge Discovery Lab. There, we compare animal-to-human correspondence for specific (groups of) treatments and specific tests, and further analyse factors which may contribute to successful translation, all using data available from literature.

Disease-preventing vaccines
One of our first studies using ASreview will assess the animal-to-human correspondence for disease-preventing vaccines, which has not yet been analysed (the preceding scoping review focussed on pharmacological treatments). This project consists of an umbrella scoping review, funded by ZonMWs more knowledge with fewer animals scheme. In it, we will summarise the evidence of studies analysing this correspondence at the meta-meta level. Because of the large body of available literature, screening for this review will be automated. Part of the literature will be screened manually, to ensure accuracy and as the basis for simulation studies to optimise efficiency. The full protocol for this systematic review will soon be available on the International prospective register of systematic reviews.

The type of data we collect in these projects can not only be used to improve animal-to-human translation, but also as a basis for the development of alternatives to animal testing; alternative (in vitro or in silico) tests should be at least as predictive as the currently used animal models.

People involved