Skip to main content
. 2017 Jan 24;3:2. doi: 10.1038/s41540-017-0003-6

Fig. 3.

Fig. 3

The potential role and number of influencer proteins. a The model of action of an influencer protein, which directly affects a first neighbour of a differentially expressed cancer-related protein. Network centrality differences are indicated by the size of the circles. Influencer proteins having lower network centrality parameters could be better drug targets than those first neighbours that are too central and multi-functional. b The number of influencer proteins of colon cancer in the overlap of the five network resources used in the current study. The number of proteins that are already drug targets are shown in parenthesis. c–e The effect of three influencer proteins, IFN-γR, FRAT1 and DUSP3 in the JAK/STAT, WNT and MAPK signalling, respectively, in normal, cancer and treated cases. At all three examples influencer proteins could be better targets than the first neighbours of differentially expressed cancer-related proteins, which are too central proteins making them difficult to pharmacologically target. Note the colour codes for the proteins: Green node: cancer-related, expressed protein; green empty node: Cancer-related, not expressed protein; orange node: first neighbour protein; grey node: unaffected protein, influencer; arrow: stimulation; blunted arrow: inhibition. For the sake of clarity we present the heterodimer of IFN-γR1 and IFN-γR2 as “IFN-γR”