PDA

View Full Version : Big difference in siblings



Mikegp
06-27-2017, 11:05 PM
AncestryDNA found me to be 54% Great Britain
43% Euro West
2% Scandinavia
1% Ireland

My brother's results were 66% Euro West
30% Ireland
4% Trace regions

I realize siblings DNA often differs but this seems excessively so. Any thoughts?

geebee
06-28-2017, 06:09 PM
If you're full siblings, I think that is a bit excessive. I might be able to believe the amount of variation in Euro West, but 1% versus 30% for Ireland? 54% Great Britain for you, and I presume none for your brother?

There can be significant variation, since you probably only share about 50% of your DNA. But, what are the odds that your brother's 30% Ireland or your 54% Great Britain would all be limited to regions you don't share?

EDIT: This is one of the ways in which 23andMe's analysis is superior to either Ancestry's or FTDNAs. It isn't just a matter of being more accurate -- though I believe that in most respects it is -- but it's also because Ancestry Composition actually shows the segments believed to belong to given ancestries. This makes comparisons among family members much more useful.

For example, I can see where my father's various ancestries are supposed to be located. So I know what actually makes sense in terms of what my siblings and I inherited, or that I passed on to my daughter from my father -- and what doesn't make so much sense.

To give one example, my siblings and I have all have tiny amount of Native American ancestry. We have documentation of such ancestry, and it also shows up in our DNA. It's found in the form of 6-8 segments for each of us, totaling around 2.0%. The segments aren't exactly the same for all of us, but with six of us they do over overlap.

Since my father and daughter have also tested, it's clear that my siblings and I inherited all of these segments from our mother. In addition, it's possible to tell from which of her parents -- in most cases -- the segments came. (Our mother has a small amount of Native American ancestry on both sides.)

What's cool is that I can also make comparisons to a few cousins who inherited segments in some of the same locations, to see whether their segments are also identified as Native American. (The answer, generally, is yes.)

Mikegp
06-30-2017, 03:04 PM
Thanks geebee, I think I may have to do 23 and me for more info.

kujira692
06-30-2017, 03:53 PM
I think it seems within the realm of normal, it's just you've each inherited a different 50% from each of your parents. I've tested both my parents, my sister, and myself and it seems like she inherited more of the British/Irish elements, while I received more of the German/French ones:

Dad
17306

Mom
17307

Sister
17308

Myself
17309

Mikegp
06-30-2017, 04:33 PM
Thanks jujira, I see the differences but my 30 to 1 still seems excessively different.

Mikegp
06-30-2017, 04:50 PM
One more detail. The only known ancestor from Ireland is my GG Grandfather who immigrated to the US in 1849. He was from County Tyrone in Northern Ireland and was Protestant which means he most likely was descendant of the colonizers from England, Wales or Scotland in the 16th and 17th centuries.
Of course there is the possibility of intermarriage between Native Irish and transplanted British.

Dewsloth
06-30-2017, 07:17 PM
Thanks jujira, I see the differences but my 30 to 1 still seems excessively different.

How about 47% to ZERO :) http://www.anthrogenica.com/showthread.php?11123-Geno-2-0-Results&p=253181&viewfull=1#post253181

Also check out how different my Evil Twin is (Gedmatch looks at your parents and extrapolates someone who got the other 50% of each of your parents' DNA):

http://www.anthrogenica.com/showthread.php?9594-Gedmatch-Evil-Twin-Phasing-Admixture&p=246726&viewfull=1#post246726

Me:
# Population Percent
1 East_Med 27.76
2 North_Atlantic 26.53
3 West_Med 14.38
4 West_Asian 14.16
5 Baltic 7.97
6 Red_Sea 6.24
7 South_Asian 1.02
8 East_Asian 0.9
9 Northeast_African 0.85
10 Oceanian 0.2

Single Population Sharing:

# Population (source) Distance
1 Italian_Abruzzo 6.87
2 West_Sicilian 9.49
3 Tuscan 10.47
4 East_Sicilian 10.98
5 Central_Greek 11.04
6 Greek_Thessaly 11.95
7 South_Italian 12.19
8 Ashkenazi 12.6
9 Sephardic_Jewish 14.6
10 North_Italian 15.41

Evil Twin:
1 North_Atlantic 48.36
2 Baltic 27.41
3 West_Med 17
4 West_Asian 2.43
5 East_Med 1.99
6 Amerindian 0.77
7 East_Asian 0.64
8 Northeast_African 0.62
9 Sub-Saharan 0.62
10 Siberian 0.17

Single Population Sharing:

# Population (source) Distance
1 Orcadian 6.04
2 Southeast_English 6.1
3 Southwest_English 6.56
4 North_Dutch 6.72
5 Danish 7.17
6 Irish 7.19
7 Norwegian 7.55
8 West_Scottish 7.66
9 North_German 7.8
10 Swedish 8.24
:behindsofa:

Mikegp
06-30-2017, 08:39 PM
So which one do you believe is accurate? If any.

geebee
06-30-2017, 10:56 PM
One problem with Evil Twin phasing is that it really doesn't completely work unless both parents have been tested.

You can use it with just one parent and child, but there's no way to actually know what the unherited SNPs are of the untested parent.

Plus, you can't always be sure which SNP you inherited from which parent at any given location, if the SNPs are not homozygous for the parent or the child at that location. So if I have AC and my father has AC, either my A or my C could have come from my father. But if he has AA or I have AA, then I know I inherited an A from him.

This also means that if both of the parents plus the child are heterozygous, then you still can't know from which parent the child inherited which SNP.

Bottom line is, if you only have one tested parent, the Evil Twin tool at GEDmatch won't work as well as if both are tested. And if it's just you and your brother, it won't work at all.

There's another problem with the Evil Twin idea. That is, what are the actual odds of having a sibling who has inherited exactly the opposite DNA from a full sibling? For two siblings who inherit 100% of the same DNA, that's easy: they're identical twins. It happens often enough not to question it.

But with the Evil Twin idea, you're basically talking about "anti-identical" twins. We know the mechanism behind identical twins; it isn't random chance. What's the mechanism behind "anti-identical" twins?

This is basically like talking about two series of coin tosses with exactly opposite results. You'd have to have either: every chromosome inherited from each parent has the same crossover points, BUT each child has the segments reversed (in terms of which came from which grandparent); or, you'd have to inherit 23 unrecombined -- and opposite -- chromosomes from each parent. Odds on that? Just a guess, but I'd say "astronomical" would probably be an appropriate term.

Unless, that is, someone can find a mechanism that is non-random, as with the formation of identical twins, who exist because a single fertilized egg gives rise two completely separate embryos. And remember, the proposed mechanism for anti-identical twins would have to involve both sides.

I suppose it might be possible to have a Merely Naughty Twin, who was only opposite on one side?

EDIT: To be fair, for one sibling to inherit 30% of an ancestry while the other inherits only 2% doesn't require exactly opposite inheritance. But, the Evil Twin phasing does. So it's probably not a good example of how different inheritance can be for two siblings in real life.

What is possible and what is probable, that's the question. It's possible that a full sibling could look like a half sibling in a DNA test. That is, by pure chance they might share DNA from just one parent.

Or, it's possible for them to share DNA from neither parent -- opposite inheritance. Has it ever happened? Maybe, but if someone claims to have seen it I'm still going to say, interrogate the milk man.

Dewsloth
07-01-2017, 06:11 AM
One problem with Evil Twin phasing is that it really doesn't completely work unless both parents have been tested.

You can use it with just one parent and child, but there's no way to actually know what the unherited SNPs are of the untested parent.

[snip]

Or, it's possible for them to share DNA from neither parent -- opposite inheritance. Has it ever happened? Maybe, but if someone claims to have seen it I'm still going to say, interrogate the milk man.

Speaking for myself, both parents are tested, and not only are they matches for Y and mtDNA, respectively, FTDNA's FF is confident they're my parents :lol:

geebee
07-01-2017, 08:39 AM
Speaking for myself, both parents are tested, and not only are they matches for Y and mtDNA, respectively, FTDNA's FF is confident they're my parents :lol:

Glad to hear it, though I didn't have your paternity in mind. :biggrin1: My only point was that if you did find two supposed siblings with 0% in common, it would be reasonable to assume they weren't actually siblings at all. And if they had only half identical matching, they would likely be only half siblings -- not full siblings who happened to inherit opposite DNA from one parent.

Here's an interesting thing about the Evil Twin, by the way. If I had been my Evil Twin, I'd have 0 matching with the self I actually am; but, my Evil Twin would not have 0 matching with any of my siblings. Instead, he'd match each sibling everywhere I don't, and nowhere that I do. Since I match four out of five siblings at over 50%, he'd match them at a bit under 50%. And since I match my fifth sibling at a hair under 50%, my Evil Twin would match her at a hair over. But, even my Evil Twin would still have very similar matching to his siblings in the real world.

The only thing I'm skeptical of is that you would ever see a pair of siblings who happened to inherit fully opposite DNA from even one parent, let alone from two.

However, it's clear that if you were your Evil Twin instead of yourself, you'd have very different scores in whatever GEDmatch ancestry calculator you used. But frankly, I don't fully trust most of those calculators. In fact, it's difficult to trust any calculator that doesn't show you where on each chromosome it's seeing your various ancestry segments.

geebee
07-01-2017, 08:52 AM
Speaking for myself, both parents are tested, and not only are they matches for Y and mtDNA, respectively, FTDNA's FF is confident they're my parents :lol:

Glad to hear it, though I didn't have your paternity in mind. :biggrin1: My only point was that if you did find two supposed siblings with 0% in common, it would be reasonable to assume they weren't actually siblings at all. And if they had only half identical matching, they would likely be only half siblings -- not full siblings who happened to inherit opposite DNA from one parent.

Here's an interesting thing about the Evil Twin, by the way. If I had been my Evil Twin, I'd have 0 matching with the self I actually am; but, my Evil Twin would not have 0 matching with any of my siblings. Instead, he'd match each sibling everywhere I don't, and nowhere that I do. Since I match four out of five siblings at over 50%, he'd match them at a bit under 50%. And since I match my fifth sibling at a hair under 50%, my Evil Twin would match her at a hair over. But, even my Evil Twin would still have very similar matching to his siblings in the real world.

The only thing I'm skeptical of is that you would ever see a pair of siblings who happened to inherit fully opposite DNA from even one parent, let alone from two.

However, it's clear that if you were your Evil Twin instead of yourself, you'd have very different scores in whatever GEDmatch ancestry calculator you used. But frankly, I don't entirely trust all of the GEDmatch calculators. In fact, it's difficult to trust any calculator that doesn't (1) use phased data; and (2) show you where on each chromosome it's seeing your various ancestry segments.

Volat
07-01-2017, 09:33 AM
Identical (monozygotic from the same egg and sperm) twins are genetically the same. It’s not possible to determine who is twin father of a child doing DNA testing if the father is an identical twin.

Non-identical twins (from two separate eggs and sperms) are just as regular siblings who happened to be conceived and born during the same pregnancy. In fact, non-identical twins can be conceived from different fathers if the mother had sexual intercourses with two mens within a couple of days, when one egg is fertilised by a sperm of one man, and the other egg is fertilised by the sperm of another man. Such cases are known.

geebee
07-01-2017, 10:36 AM
Identical (monozygotic from the same egg and sperm) twins are genetically the same. It’s not possible to determine who is twin father of a child doing DNA testing if the father is an identical twin.

Non-identical twins (from two separate eggs and sperms) are just as regular siblings who happened to be conceived and born during the same pregnancy. In fact, non-identical twins can be conceived from different fathers if the mother had sexual intercourses with two mens within a couple of days, when one egg is fertilised by a sperm of one man, and the other egg is fertilised by the sperm of another man. Such cases are known.

That isn't what's being discussed here. The "Evil Twin" refers to a utility available at GEDmatch. Using the DNA files of a child and at least one parent (preferably both), the utility calculates the DNA the child did not receive. The child himself/herself is the "Good Twin"; the hypothetical opposite is the "Evil Twin".

That's why I also referred to this "Evil Twin" as an anti-identical twin. Whereas identical twins have essentially the same DNA, an anti-identical twin shares no DNA even with a full sibling -- he or she inherits only the DNA that the parents did not pass on the other child.

Unlike identical twins or fraternal twins, I don't believe that these anti-identical twins are likely to actually exist.

EDIT.

This statement is actually no longer true:


It’s not possible to determine who is twin father of a child doing DNA testing if the father is an identical twin.

You can't do it with standard DNA testing, but you can if you look at a sufficiently large number of markers. Even identical twins develop a tiny amount of divergence in their DNA, and by examining that divergence (which is passed on to the child), it's possible to tell which twin is the father.

http://genetics.thetech.org/ask-a-geneticist/identical-twin-paternity-test

ArmandoR1b
07-01-2017, 01:09 PM
I think it seems within the realm of normal, it's just you've each inherited a different 50% from each of your parents. I've tested both my parents, my sister, and myself and it seems like she inherited more of the British/Irish elements, while I received more of the German/French ones:

Dad
17306

Mom
17307

Sister
17308

Myself
17309

The larger amounts of ethnic DNA that we inherit from a parent isn't spread out in the parents in a way that it can actually be missed to the degree that Mikegp mentioned. For the same reason, you are missing way too much Great Britain from your mother for that to be accurate. It's not like the DNA recombination just happens to miss so many regions of your parents combined 58% Great Britain that you only get 8% when you should get between 25-30%. It's not statistically possible. Your parents combined British/Irish is 102% so you should get between 40-60% but you only get 25%. Your sister is closer to what the results should be. You should only have a difference of about 10% from your sister.

Family groups of parents and children and the testing of identical siblings with multiple companies to show how much of a difference in the results there can be and how little sense the results make from some of the companies. Comparing the results of a lot of people with well documented genealogies also shows some companies have a worst track record than others. AncestryDNA isn't the worst but it isn't the best either.

Ron from PA
07-01-2017, 02:30 PM
In my opinion, I don't think they can tell Europe West/Britain apart with any confidence.

Volat
07-01-2017, 02:35 PM
This statement is actually no longer true:


It’s not possible to determine who is twin father of a child doing DNA testing if the father is an identical twin.

You can't do it with standard DNA testing, but you can if you look at a sufficiently large number of markers. Even identical twins develop a tiny amount of divergence in their DNA, and by examining that divergence (which is passed on to the child), it's possible to tell which twin is the father.

http://genetics.thetech.org/ask-a-geneticist/identical-twin-paternity-test

Okay! It's now has become possible by analysing 6 billion marks in DNA of each identical twin, the mother and the child. How many labs will offer such service? The price is still high for an average household in the developed world let alone the rest of the world.


This is why you need a more comprehensive test. Instead of the usual 15 or so markers that a standard test looks at, you need a test that looks at six billion or so markers. Only then do you have any chance of telling which identical twin is the real dad.

You can only get this much information from sequencing the entire set of DNA (or the genome) of each twin, the mother, and the child. Until a few years ago this would have been way too expensive. But nowadays, with the cost coming down, you could probably do something like this for tens of thousands of dollars.

As you can tell, this won’t be available to everyone any time soon. But if you really need to know and you have lots of money, you have at least a chance of finding out.

MitchellSince1893
07-01-2017, 05:01 PM
This may have already been discussed, but
http://www.anthrogenica.com/showthread.php?11073-How-Reliable-Are-Home-DNA-Ancestry-Tests-Investigation-Uses-Triplets-to-Find-Out

Mestace
07-01-2017, 05:21 PM
Ancestry compositions are choosing arbitrarly a best fit that can look different when it isn't really, both of your compo likely end up being roughly the same genetic point. Better compare between admix in calculators, i have some differences with my sisters in say DNAland, however in calculators it remains in the 1%.

geebee
07-02-2017, 01:29 AM
In my opinion, I don't think they can tell Europe West/Britain apart with any confidence.

I'd agree with that assessment. If you click on "See Details" for Europe West, this is what you find:


Primarily located in: Belgium, France, Germany, Netherlands, Switzerland, Luxembourg, Liechtenstein

Also found in: England, Denmark, Italy, Slovenia, Czech Republic

And if you click on "See Details" for Great Britain, you get:


Primarily located in: England, Scotland, Wales

Also found in: Ireland, France, Germany, Denmark, Belgium, Netherlands, Switzerland, Austria, Italy

So even Ancestry seems to acknowledge that the regions overlap.

In fact, if you look at the concentric rings on the map on this page, the outermost one extends down into southern France and northern Italy. Does that mean some of the people who have lived in this regions for many generations are really British, or just that the component Ancestry labels as "Great Britain" isn't actually exclusively British -- just as the component called "Europe West" isn't exclusive to the Europe West region.

selectivememri
07-20-2017, 01:39 PM
when one gets a family member (parents for example) tested with ancestry, does the site do anything to change my results (either more refined percentages, phasing, etc?) or allow me to compare between the two?