Reduce the Number of Animals per Experiment
The Problem with Small Experiments
Manual or purely random distribution of animals into treatment groups often proves suboptimal, especially for small-scale animal studies (N<12 per group) [Bertsimas et al. 2015]. The common workaround—increasing the number of animals per group (N>12)—comes at a higher cost, raises significant ethical concerns, and is often not feasible.
Optimized Balancing: Same Quality, Fewer Animals
Randmice's advanced algorithm identifies the optimal group assignment by iteratively testing combinations and minimizing inter-group heterogeneity. By precisely controlling the distribution, our method allows you to achieve the same level of homogeneity between groups with fewer animals than traditional random assignment. This directly translates to more efficient and ethical preclinical research.
The figure below vividly illustrates this principle: transitioning from 8 mice per group with random assignment to just 6 mice per group with Randmice's optimized balancing maintains equivalent inter-group heterogeneity—effectively saving 2 animals per group without compromising statistical validity.
8 mice/group (random assignment) to 6 mice/group (optimized balancing) yields the
same inter-group heterogeneity, resulting in 2 mice saved per group.
Comply with the European 3Rs Ethical Framework
The European 3Rs ethical framework—Replacement, Reduction, and Refinement—is paramount for humane animal research. This framework is embedded in national and international legislation concerning the use of animals in scientific practice. Key European directives include:
- Directive 2010/63/EU
- Regulation 2019/1010
By significantly reducing the number of animals needed per experiment without compromising statistical quality, Randmice directly supports the Reduction principle of the 3Rs. Implementing optimized balancing in your experimental protocols helps justify 3Rs compliance to ethical committees and regulatory bodies, promoting both scientific rigor and animal welfare.