Figure 7
From: R-Gada: a fast and flexible pipeline for copy number analysis in association studies
Variable ranking for MCDA. CNV blocks (recurrent CNVs) are ranked according to equation 4, that is their maximum correlation across class axes within the principal component subspace. Variables that are more relevant to the discrimination have higher correlations. The figure after the dot refers to relevant level (-1: loss, 0: no gain, 1:gain) of the CNV to discriminate one of the populations.