Obviously a 0.5% difference over 3 billion base pairs is 15 million, not 3 million.
So how did that same source get both figures? They are obviously confused, or leaving out a critical bit of info.
Do you have a source for human variability similarly being 99.5%? The figures both you and I have cited have given it as 99.9%.
Either way, the data I have cited has given the human similarity as 99.9% and the neanderthal-human similarity as 99.5%. This is exactly what you would expect if the most distantly related human populations shared a common ancestor within the last 100,000 years, but neanderthals branched off some 500,000 years ago. Five times as much genetic difference, almost all of it in ‘neutral’ DNA regions of the genome.
And that is the point I've tried to make in all these exchanges.
.5% equates to 15 million DNA differences, not 3 million, and yet the only solid number I've seen regarding Neanderthals is 3 million base pair differences.
In everything else we agree -- 15 million for Neanderthals is the number we might expect. It's just that I haven't seen it reported yet. When I do, then will concede the point.
And your larger point is also correct -- that these numbers can be more complicated than it might at first seem.
An example of that is the question of "copy number" which turns out to increase intra-human genetic diversity from .1% to .5%.
So it is important that we know which numbers we are talking about before we start throwing them around.
But finally, let me suggest to you a possible reason why the reports are telling us that Neanderthal DNA is only 3 million base pairs different from human DNA.
This could be the case, couldn't it, even if our common ancestors lived 500,000 years ago -- if there was significant interbreeding between those last living Neanderthals and some of our earlier human ancestors.
Wouldn't a bit of trogloditic hanky-panky tend to reset the genetic clocks on both sides? ;-)