Reverse Translation: Using Computational Modeling to Enhance Translational Research.

Abstract

Animal models provide powerful tools for examining human disease; however, translation of findings from these models to human patients is often challenging. To this end, we discuss modern tools to support the process of selecting and validating animal models with relevance to humans. We draw from data mining and computational modeling approaches to examine how large data sets may be leveraged to identify suitable models with the greatest translational potential. Rodent models for human disease are indispensable tools for modern biomedical research, particularly because of the high degree of control they afford in examination of complex pathophysiological phenomena. There is no doubt within the scientific community that animal models are both useful and necessary; however, the value of any animal model depends on its ability to replicate the human condition. Curing disease in a mouse is fascinating but short on merit without application to the understanding of human disease. Thus, there is great need for a robust process to translate a human disease into an appropriate rodent model, termed reverse translation, to distinguish it from the more familiar act of moving findings from animal models into human patients (translation). Studies on the quality of rodent models have reported disparate results depending on the specific physiological system of study. For example, studies on the acute inflammatory response to blunt trauma, burn, and endotoxin have found a poor correlation between the response in mouse and that in human, leading to the conclusion that murine models have limited translational potential.1 This weak association may be due, in part, to a highly variable response in mice that is not observed in humans, underscoring the low success rate in translating mouse findings on acute inflammatory response to human. However, rodents fare better as a model in chronic inflammation, where there is a much better …