When performing pathway analysis, should you only include proteins that were significantly changed (after performing t-tests, ANOVA etc), or should you include all your proteins from the dataset to get a complete picture?
Apart from the problem that it is not well defined what is meant with "significant", you will likey get a very biased picture if you work with some selection (selected by siginficance or whatever else). So you better should use all the data.