The paper A comprehensive evaluation of module detection methods for gene expression data says that:

"We first want to provide an overview of the characteristics and performance of current module detection methods to guide the biologist in their choice"

.

My problem is that I have no gold standards for my gene dataset and I would like to know if I can use the paper results to choose a module detection method without the need of external validation measures and how can I justificate this generalization? I am not finding the justificative about how can I use this guiding.