Wednesday, August 29, 2018

Moralistic Fallacy on Ancestry Testing

Responding to an unmitigated idiot.

This may be the most scientifically illiterate piece on genetics I have ever read.
What worries me most, however, is that companies offering personal genetic testing customarily seem to report back to those sending along a sample of their spit that they are a mix of different “ethnicities.” This is more than simply statistical nonsense. I fear doing this can also be dangerous. Claiming that it is possible to map ancestry in this fashion may be giving discredited old ideas about ethnicity and race new visibility.
This is the moralistic fallacy: It would be a bad thing, very worrisome, if genetic tests can detect ethnic and racial differences; therefore, they cannot detect them.  This guy is either retarded or mendacious.
{snip} Despite the wars, border tensions, and other types of violence that stem from perceptions of human difference, we are approximately 99.9 percent identical to every other human on Earth in terms of our genetics. 
How much do we share with chimps (albeit some more than others, eh)?
In short, we are all cousins. 
Lie.
Even when you drill into the specifics of the remaining 0.1 percent to learn something more detailed about your biological ties, don’t get your hopes up that you can identify real ancestors very far back in time.
That’s an absolute lie.  Even with the flaws of these tests, they CAN and DO tell you what your majority ancestry is.  How come an Irishman will test out as Irish and not Pakistani?
Here’s the basic math. We inherit roughly half our genes from our mothers and half from our fathers. If one or both of them should be unknown to you, it is a safe bet gene profiling may help you track them down. But how far back across the generations can you go and have similarly assured success? Go back, say, five more generations to your great, great, great, great grandparents. Assuming there hasn’t been a lot of inbreeding in your ancestry (the further you go back in time, by the way, the more likely it occurred), you should have 64 of them. Only about 1.56 percent of your genes may come down to you from any one of these 64 ancestors. Good luck should you go looking for them many generations back—or their living descendants.
See below.  This “argument” is completely irrelevant in evaluating ancestral sources that have made a significant contribution to your genome.  Yes, possibly, that single “Indian princess” from centuries ago may not show up, she will be lost in the meiotic shuffle.  But no one is going to mistake a Dane for a Japanese, and that’s what this idiot is really worried about.
Now go even further back in time to the 17th or 18th century. The number of folks on average living then who could have contributed to your genetic endowment is so large (more than 1,000), and their possible genetic contribution so small (about 0.098 percent for 10 generations back), it would be smoke and mirrors to assert claims about who they were in person. In fact, most of these people left no trace of themselves in your genome.
So what?  The tests are not looking for particular ancestors.  They are evaluating a person’s genome and ascertaining what the ethnic mix represented in that genome (as it is) is.  Trace ancestry may be lost, but the major components will be present; obviously, some ancestors will be – must be - represented in your genome (or else you wouldn’t exist), and it is proportionately likely that what is represented will derive from those ancestors that make up the bulk of your genealogy.  If 50% of your ancestors were French and 2% were Chinese (and 48% something else), the Chinese ancestry may or may not be detected – the luck of the meiotic draw – but you can rest assured that a substantial amount detected (but obviously not precisely 50% of course) will be French (or whatever level of specificity a given test reads French as).
In short, while it can be hard to get your head around the statistics involved, go back more than a few thousand years and you are genealogically related to almost everyone on Earth. 
That’s true and why Der Movement’s ranting about “purity” is nonsense.  But – BUT! – you are going to be much more related, and more recently related – more recent common ancestors – with some groups than others.  Germans are going to have most, or all, of their most recent common ancestors with other Europeans, particularly Northcentral Europeans, and Nigerians will in turn share such ancestors with sub-Saharan Africans, mostly West Africans.  Sure, if you go back far enough, Germans and Nigerians will share ancestors.  If you go back even farther, both groups will share ancestors with mice and groundhogs.  People take ancestry tests to ascertain their ancestry in the sense that they differ from others; they want to know what makes them unique.  And that’s the level the tests, correctly, evaluate.  They are looking for your ethnic and racial family, and they WILL identify the major components of that.
Genetically speaking, however, very few of these very distant ancestors contributed something of themselves biologically to your genome.
Yes, but you are getting a representative sample, you mendacious dumbfuck.  When an American polling company performs a political survey, they do not poll every single American citizen.  They take a representative sample.  Think of a genetic test as taking a representative sample of your ancestral background.  There will be some statistical error, some noise, but if performed correctly, the “poll” of your genome will correctly ascertain - within reason - the likelihood of your ancestral mix.

As an example of the power of this representative genetic survey, it is possible to detect Neanderthal gene variants in modern Eurasians.  So, yes, if you are of European or Asian descent, that small and prehistoric fraction of your genome derived from Neanderthals can be identified.  And yet this low-life shitheel pretends we can’t distinguish a Swede from an Angolan - that it’s “nonsense” to estimate ancestral proportions derived from extant ethnic groups.  Really now, isn’t it time to call out bullshit when you read or hear it?

Of course, the reason why this turd is so hysterical is that the tests actually WORK, in the broad fundamentals.  Yes, I am critical of these tests and how they are over-interpreted, and how they ignore kinship metrics.  But those are details – important details, but still details.  While I would be suspicious about ancestral proportions less than 10%, and certainly less than 5% (and less than 1% is a bad joke), when it comes to majority ancestry, the tests are on target.  Don’t believe that?  Put it to the test (no pun intended).  If you know your own background, get tested.  If your friends know their background, get them tested as well.  Then observe how well the tests detect the bulk of your ancestry, and know that the deniers are lying to you.

By the way, all of this moron’s arguments would apply to the field of population genetics, you know, all those peer-reviewed studies that can distinguish race with close to 100% accuracy, distinguish German-, French-, and Italian-speaking Swiss, and identify people who are even just ¼ Jewish.  

What a despicable piece of lying filth this scumbag is.