Skip to main content
Fig. 6 | BMC Bioinformatics

Fig. 6

From: Bioconductor’s EnrichmentBrowser: seamless navigation through combined results of set- & network-based enrichment analysis

Fig. 6

Combination of methods improves individual rankings on the GEO2KEGG benchmark set. a Combined ranks by rank sum (solid lines) were computed for the SBEA-combination of the 4 set-based methods (blue) and the NBEA-combination of the 4 network-based methods (green). Depicted is the number of GEO datasets (y-axis) for which the combination yielded a ranking of the corresponding target pathway, which was at least as good as obtained from x of the individual methods. As an example, the green point at x=3 and y=15 indicates that the NBEA-combination returned for 15 of the 27 datasets (55.6 %) a ranking of the target as good or better as obtained for 3 of the 4 individual methods (i.e. only one method yielded a ranking better than the combination). For comparison, the dashed lines depict corresponding results when, instead of re-ranking by rank sum (solid lines), ranks are averaged across methods. b shows the total number of unspecific pathways ranked at least as good as the target pathway (y-axis) for the 4 network-based methods and each pairwise combination. Unspecific pathways are defined in the main text and listed in Additional file 1: Table S1. Corresponding values for the NBEA-combination of the 4 network-based methods and the ALL-combination of all 8 methods (4 set- and 4 network-based) are indicated with the green and the brown dotted line, respectively

Back to article page