Figure 3.

Statistical comparison of GSECA results with conventional strategy. A. Gene clustering and enrichment analysis was performed for 7 functional gene sets corresponding to the functional cluster 2 of myogenesis dataset. K-means and SOM clustering was performed with 16 different settings for the gene numbers to be clustered (5 – 50%) and cluster numbers (5 – 100). The significance levels (Y-axis) are illustrated with the color lines corresponding to 16 settings (shown in the bow below). For comparison, the unadjusted significance level or normal P value of GSECA algorithm are demonstrated as asterisk. B. The significance level for 2 functional gene sets of erythropoiesis are similarly calculated and compared with those of GSECA results. C and D. The comparison results of 6 and 3 regulatory motif gene sets with significance enrichment in the functional cluster 2 of myogenesis (C) and cluster 1 of erythropoiesis (D) are similarly demonstrated.