Much of the 'junk' DNA did serve a purpose at one point; deactivated genes for instance. Much of it still serves a purpose now, such as coding elements and transcription factors (see the work of Sean Carroll for more info on this point). Some of it is there for epigenetic and structural modifications such as the methylation of cysteine residues, (and similarly the acetylation of histones) which actually changes the shape of the DNA helix itself (and this affects transcription). And some of it is there simply to take up space. Intron splicing, for example, requires a minimum distance between the exons to function properly; longer is okay, but too short and you'll start skipping out on pieces of genes that *should* be there. And, following one of the older theories about the purpose of the 'junk' DNA, it acts as a buffer space to limit the damage caused by mutations that *will* happen.
So yes, the "junk DNA" isn't necessarily useless; but in many cases its sequence isn't necessarily meaningful either.
To use a car analogy: Sometimes it's like analyzing the composition of your engine block, where changes in the trace elements can have an affect of the performance of the vehicle as a whole. And sometimes it's analyzing samples of the air residing in your door panel (between the exterior sheet metal and plasticky interior) It's there to take up space and its composition really doesn't matter overall.