Quote:
Originally Posted by Eugene80
I was thinking that 1% is 1/100, which is a lot less than 1/16 [4] or 1/32 [5] or even 1/64 [6] generations ago. When approximately could it have occurred?
Also, is 1% pretty much noise, or does it have tangible meaning?
|
There's a lot of qualifications to this. If you are talking about DNA ethnicity then it also depends on what. Major world DNA regions such as European, Sub-Saharan African, and East Asian/Native American being the most separated are the most likely to identify separately.
In this case 1% is very likely not noise. Even half a percent or less isn't necessarily noise. At smaller percentages like this what is more important is not the total percent but the cM size of the segments. For example of the 1% comes from a slew of 3-5 cM segments, those all "could" be noise (though might not be, just can't really tell easily). If the 1% comes from segments much larger, say 20+ cM then it really is far more likely not to be noise (everything is degrees of likelihood, not certainties). For context people consider 10+ cM the typically threshold for a valid amount of shared DNA with someone who is likely a cousin within a genealogically significant time period. So personally I find that to be a relevant threshold for ethnicity segments. At 1% I'd guess you have some larger segments, though that'd be the way to be a bit more confident.
Back to what this means for regions. If you are say of full European descent, so the vast majority of your DNA comes up as that and you get a Sub-Saharan African or East Asian/Native American 1% then it's probably not noise. The DNA profile is different enough that it's probably something along those lines.
Now if you are say of European descent and you are wondering if the 1% is say North African, West Asian, or say a specific European nationality, lets say Italian. Then don't put much stock in that. It's [b]really[/] hard to distinguish DNA from surrounding regions. Partially because it's hard to say what region a DNA really comes from! German and French is often merged for example because there has been so much back and forth and the modern national lines are just arbitrary for the most part. So whose to say what DNA is German say 500 years ago, 300 years ago, or say 50 years ago, compared to say now. So don't put much stock in the breakdowns within the broad regions since it's just a guess and those regions for the most part don't have distinct DNA profiles that are historic to just that place since they intermixed all around them. The furthest removed places though have developed DNA distinct enough to be different and identifiable to a reasonable degree of accuracy however.