[[On the other hand, if non-coding has important functions then their differences are significant enough to consider in percent similar calculations.]]
That’s the key point of hte discussion- the link i gave cites several articles, done by ‘secularists’ btw, throughout history, that indicate we should be counting it all, or at least much of what had been deceitfully thrown out when doing comparisons
But there's no "deceit" if we understand that over 90% of DNA is non-coding and mutations there have no visible effects on survival & reproduction.
But coding regions are quite different -- there any mutations can be fatal resulting in Natural Selection weeding those mutations out of our gene pools.
Yes, I "get" that some people now suspect some non-coding DNA does have some functions.
However, it's not yet been shown that DNA mutations there effect survivability, and therefore in terms of coding percent similarities, such mutations are irrelevant.
Bottom line: if you wish to include all non-coding DNA in your comparison of chimps & humans such that it shows only, what is it, 85% similarities, that's fine.
So long as you properly identify how you arrived at your number, it's legit.
But in terms of DNA that truly matters -- protein coding DNA -- the figure of 98% (plus or minus) similarity between chimps & humans is still valid.
The reason is Natural Selection working on Descent with Modifications, just as Darwin predicted.