Abstract
In genome-wide association studies, statistical procedures such as logistic regression or the Cochran-Armitage test for trend can be used for the analysis of genetic variants while taking the underlying genetic model into consideration. Should a researcher was to investigate the role played by the interaction between two or more of these genetic variants (epistasis), it would suffice to add an interaction term, which corresponds to their product, to a logistic regression model. However, we question whether such a model is able to capture the intricacy of the genetic architecture in complex traits. Therefore, this paper proposes a network-based model that allows graph theory to characterize both genetic variants and their interactions in a genome-wide context, including a simulation study that proves its applicability.