Figure 2
From: Extracting expression modules from perturbational gene expression compendia
Performance on artificial data. Performance of ENIGMA versus other coexpression measures and biclustering methods on (A) modular and (B) non-modular toy datasets. The ENIGMA-T curve shows the performance for the ENIGMA coexpression network at several non-corrected p-value thresholds, ENIGMA-N stands for the ENIGMA coexpression network at FDR = 0.05, and ENIGMA-M for the final clustering result.