I'm using the RRHO package to compare ranked gene sets. The associated publication mentions that "the highest intensity point on the [RRHO] map can be used to extract the most statistically significant overlapping gene set", but there is no information on how this is accomplished. How exactly can this be accomplished?