If you’re specifically interested in using DNA to determine an evolutionary timeline, nuclear DNA mutates at a slower rate than mitochondrial DNA. I think that’s why the latter is used for this purpose rather than the former.
I always thought that the molecular clock was based on “neutral” mutations. E.g., if AAA and AAT code for the same amino acid (I don’t know that they do; I do know that there is a fair bit of redundancy in the code) then the mutation of one to the other would have little or no somatic effect and not be selected against (or for). Recently, I have read that even that may not always be so simple as the transcription efficiency of AAA might not be identical to that of AAT.
Incidentally, one of the early clues that “junk” DNA might not be junk is that at least some of it is highly conserved over time, suggesting it is important.