On generating PPI on STRING database for my handful of proteins i) with a confidence cut off score of 0.7 ii) without any additional interactors, I am hardly getting a decent network. There seems to be hardly any interaction :( However on reducing the score or adding additional interactors , I am able to observe some interaction which can be explained.
In such a case, would you recommend setting my cutoff to a lower value of 0.6 or 0.5 ? Or would you recommend adding additional interactors? Which of these would be scientifically more reliable?