It's tangentially related, but people don't realise just how accurate and consistent DNA replication is. I don't have the error rate for humans, but E.coli is 1 error every ~1,000,000,000,000 replications (give or take a 0. I also asume DNA error correction is taken into account here) For context, humans have ~3,200,000,000 nucleotides.

This incredibly low error rate means that organisms that are related to each, even if a common ancestor was a few dozen generations ago, will have very similar DNA. As a result, the broad "99.9%" statistic would likely be accurate for both coding and non-coding DNA.