NES directional contraints

[ivokwee]

I wonder what you think of adding directional (sign) contraints. In the example data of the vignette, the top geneset 'Metabolic pathway' is for example negative in transcriptome, positive NES in proteome, and negative in metabolome. Should a true enrichmed pathway not be all of the same sign positive or negative? The combined p-value does not reflect that.