Our paper on a two-sample test for the detection of a systematic difference between Gaussian graphical models appeared online in the Journal of the Royal Statistical Society - Series C.
In the paper the two-sample problem is addressed from the perspective of Gaussian graphical models (GGMs). It concentrates on the particular situation in which partial correlations (i.e. edge strength measures between node pairs in a GGM) are systematically smaller/larger (in an absolute sense) in one of the groups. Biologically, systematically weaker/stronger partial correlations represent an inactivate/active pathway. Read more.