PDA

View Full Version : 23&me - "relatives"



Dante
04-13-2016, 10:40 AM
Please move if the wrong spot but I am curious about the "relatives" of 23&me.

I understand some of it.

Example large segments [e.g. 1.2%, 7 segments] shared should indicate closer relation - given as largest segments typically are marked 2nd cousins.

However, I've read and it makes sense given recombination, that smaller segments don't necessitate that someone is 6th cousin - they could actually be 3rd cousin.


What I don't get is the haplogroups on 23&me. Haplogroups are supposed to be passed down father to son, mother to daughter / son correct? So how can a "2nd cousin" on 23&me then have haplogroups that are not even related to what I, my sibling and by association parents possess? I am not talking one haplogroup been off but both maternal & paternal having no relation.

We're R1b* / U5 - one "2nd cousin" is J1b / E1b*. It is sort of curious because unless I missed someone the only potential Jewish in the family tree are from the 1700s which I don't think should put us as 2nd cousins with someone. That the guy is private via 23&me of course is a hassle to figure out exact relation.


My point is whom on 23&me to take as actual relatives? And whom might just be "noise" in a way.

People that share same haplogroups?

People with large segments?


I'm asking cause trying to figure something out relative wise & most of the relatives that aren't private or the public ones that have familial info [there's public with just haplogroups] are pretty obvious to how they're related.

A Norfolk L-M20
04-13-2016, 01:18 PM
I can answer the cousin and haplogroup question.

yDNA passes down the strictly paternal line - father to son. mtDNA passes down the maternal line - mother to daughter. Most of your cousins will not be on the same trajectory. For example, you have eight great grandparents. You'll share only two of them with your second cousins, but none of these shared great grandparents, may pass directly through a strict paternal or maternal line to either party. You do not for example, inherit a haplogroup from your mother's father, or from a father's mother - the line is broken. Go back a few more generations, and you may have a hundreds of third or fourth cousins that do not share either haplogroup with you, or even if they do so, not by any recent common ancestry.

Of my twenty first cousins, I only share yDNA with one of them - we've both passed down from sons of a shared paternal side grandfather. The rest passed down via sisters of my father, or on my mother's side. The Y chromosome does not pass through women. Equally with mtDNA on the opposite trajectory. My two male cousins on my mother's side do not share yDNA with me - because they passed through son of maternal grandparents. They instead get my maternal grandfather's yDNA. Neither will they share mtDNA, as they would inherit that from their mother's to whom I am not related by blood. You see, I have full first cousins that will not share either haplogroup with me.

On the 23andMe relatives question, I'm a beginner myself, but yes, I understand that 23andMe's relationship interpretation of lengths and numbers of common segments is indeed a bit far fetched. I have so many 23andMe 3rd Cousins that I can find no common relationship with, even though we should share a pair of G.G grandparents.

AJL
04-13-2016, 02:47 PM
I can answer the cousin and haplogroup question.

yDNA passes down the strictly paternal line - father to son. mtDNA passes down the maternal line - mother to daughter. Most of your cousins will not be on the same trajectory. For example, you have eight great grandparents. You'll share only two of them with your second cousins, but none of these shared great grandparents, may pass directly through a strict paternal or maternal line to either party. You do not for example, inherit a haplogroup from your mother's father, or from a father's mother - the line is broken. Go back a few more generations, and you may have a hundreds of third or fourth cousins that do not share either haplogroup with you, or even if they do so, not by any recent common ancestry.

Of my twenty first cousins, I only share yDNA with one of them - we've both passed down from sons of a shared paternal side grandfather. The rest passed down via sisters of my father, or on my mother's side. The Y chromosome does not pass through women. Equally with mtDNA on the opposite trajectory. My two male cousins on my mother's side do not share yDNA with me - because they passed through son of maternal grandparents. They instead get my maternal grandfather's yDNA. Neither will they share mtDNA, as they would inherit that from their mother's to whom I am not related by blood. You see, I have full first cousins that will not share either haplogroup with me.

On the 23andMe relatives question, I'm a beginner myself, but yes, I understand that 23andMe's relationship interpretation of lengths and numbers of common segments is indeed a bit far fetched. I have so many 23andMe 3rd Cousins that I can find no common relationship with, even though we should share a pair of G.G grandparents.

Well said.
The chances of you matching a haplogroup with an autosomal cousin because you share a recent autosomal ancestor are almost nil.

A second cousin could be related through your father or mother, and then through their father's or mother's parents, and then through their son or daughter, and then through their son or daughter. That already gives you eight possible pathways up, and then there are eight possible pathways down. If you are male you have one mtDNA line and one yDNA, but half the database only has mtDNA. You can quickly see how unlikely it is that you'll have a same-haplogroup match that's relevant to your common ancestor.


We're R1b* / U5 - one "2nd cousin" is J1b / E1b*. It is sort of curious because unless I missed someone the only potential Jewish in the family tree are from the 1700s

How do you know this cousin is Jewish? If this is old J1b (i.e. J1-M365.1), it's found scattered lightly around Europe, while mtDNA E1b comes form the Pacific rim of Asia. This is precisely why you should not look too closely at the haplogroups of your cousins to try to figure out whether they match. Use the autosomes to figure out if they match, then worry about what the haplogroups mean later. If you have enough DNA relatives you will inevitably have cousins with African, Asian, European, and Amerindian haplogroups. It doesn't mean they'renot related, nor that you share these ancestral haplogroups with them.

ffoucart
04-13-2016, 02:57 PM
Y and MT haplogroups are markers of a regional origin, not of your relation to somebody else.
A simple example: you don't inherite the haplogroups of your maternal grandfather, but you are still closely related by your autosomal DNA.

As it was said above.

Little bit
04-13-2016, 03:40 PM
Here are a couple of nice charts which show how haplogroups get passed down and how easy it is to carry a different haplogroup than even close relatives:

8824
https://dna-explained.com/2013/01/24/what-is-a-haplogroup/

8825
https://blog.23andme.com/23andme-and-you/genetics-101/ancestry-at-23andme-what-can-you-learn/

First cousins would share grandparents with you. I share my mtdna haplogroup with my maternal first cousins from my aunts but I would not share a mtdna with paternal first cousins or a maternal first cousin descended from an uncle. As a matter of fact, I don't share a mtdna with my dad, with whom I share 50% of my DNA and I know from testing that my maternal grandfather is K1a41d, while I am J1c3i. And yet, we share 25% of our DNA.

anglesqueville
04-16-2016, 07:51 AM
By the way, one more frenchy on our forum. Who tells that french are not interested in genetics? Wellcome Dante!