 I've already talked a lot about enrichment analysis, but I haven't said anything about interaction enrichment. I recently showed you how in the String database you can not just pull an interaction network, but also perform gene-set enrichment analysis and visualize the results on the network. However, String also does PPI enrichment analysis. At the top of the analysis tab, you will see numbers like these, the number of nodes, the number of edgils, the number of expected edgils, and typically a very significant P-value. When you see that, it tells you that you have more interactions among these proteins in the network than you would expect if they had been random proteins. In other words, the input list that you provided does not look to be random, which is reassuring if it came from your ownomic study. One could argue the alternative would be really bad. Try out the enrichment analysis at StringDB.org.