Its been a few years since I left the genetics world behind, and I was just a sysadmin, but even then I'm pretty sure they had determined non coding regions to play a vital role in protein formation/folding.
Non-coding = "not translated into proteins", not "isn't used". Some non-coding DNA produces non-coding RNA, including transfer RNAs (https://en.wikipedia.org/wiki/Non-coding_RNA), which help in protein construction.
This makes sense given we know that some nontranslated RNA has biological activity , like tRNA and rRNA. I was under the impression that "junk" DNA referred to DNA that was not transcribed.
nearly all DNA is transcribed at some low rate (this has been experimentally determined) but it seems probable that most long stretches of non-coding DNA contain little to no functional elements and could be replaced or removed with no observable functional effect.
It's possible that upstream UTRs impact RNA transcription rates, but I'd be surprised if UTRs mediate protein folding.