# Is a half-sibling the same as a first cousin, from a genetic standpoint?

Mathematically it makes some sense, since in both cases they share one set of grandparents. But perhaps additionally having the same parent is stronger than merely having two of the same grandparents.

Half siblings have [about] 25% DNA in common, cousins share [about] 12.5%.
I tried to work this out with Punnett squares and alleles, but it’s just easier to link to a page that describes it far better than I can.
https://customercare.23andme.com/hc/en-us/articles/212170668-Average-Percent-DNA-Shared-Between-Relatives

What about first cousins with one identical twin parent? Twins of each other of course. So for example, two first cousins who have fathers that are identical twins.

Since genetically speaking, they have the same father and different mothers, I would imagine it would be the same as half siblings.

ETA yup
" In fact, this is why at the DNA level they are really more like half siblings they share 25% of their DNA"

Makes perfect sense. Thanks.

So identical twin fathers with identical twin mothers (it could happen) - cousins from each family share the same amount of DNA as siblings.

Yep, I’m in that sort of arrangement with my cousin. We were born less than three months apart and our fathers are twins. It was only about a decade ago that I realized that, genetically, that made us half-brothers.

Cousins and half-sibs each share one set of grandparents, but for cousins they arise from different offspring of those grandparents, while half-sibs arise from the same offspring of those grandparents. This is an extra reduction by half of relatedness, so, as stated, half-sibs on average share a quarter of their genes, while cousins on average share an eighth of their genes. (This of course only refers to the genes that vary between people, if you’re comparing the two humans to an oak tree, the two humans share essentially 100% of their genes.)

``````      Granddad-Grandmom
/\
/  \
X-F1a  Y-F1b-Z
|      |   |
F2a    F2b F2c
``````

F1a and F1b are siblings. X and F1a have a kid, F2a. F1b has two kids, F2b and F2c, with two different partners, Y and Z. Assume X, Y, and Z are unrelated. It’s easy to see that F2b and F2c are closer to each other than they are to F2a on the pedigree.

The amount of shared genes are on average, and will be a distribution. It is very possible to get cousins, who just by chance, are more related than average, say 0.18 instead of 0.125, and half-sibs who are a related less, so 0.18 instead of 0.25. This isn’t even taking into account in-breeding. If X and F1a happen to be slightly related, that will increase how closely F2a is related to the rest of the F family, because F2a will be getting some F genes from both parents’ sides.

I don’t understand. My first cousin and I share one common set of grandparents. Why do we not therefore (twins and other hypos aside) share approximately 25% of our DNA?

First cousins have two shared ancestors: one Grandmother and one Grandfather. Let’s first figure out how much DNA from Grandma the cousins share.

We already know that a child share 50% of his DNA with his dad. And since Dad got 50% of his DNA from Grandma, we can figure out that Grandma and Grandchild will share 25% of their DNA.

(½)(½) = ¼ (25%)*

How much DNA would Grandchild share with his Aunt? The Aunt inherits half of the Grandma’s DNA. So Grandchild and Aunt share 12.5% of DNA from Grandma.

(½)(½)(½) = ⅛ (12.5%)*

And if you go down one more generation, Cousin shares 50% of his DNA with Aunt, so first cousins (Grandchild and Cousin) share 6.25% DNA from Grandma.

(½)(½)(½)(½) = 1/16 (6.25%)*

But Grandchild and Cousin have two ancestors in common, Grandma and Grandpa. So you can add up the shared DNA that they inherited from each shared ancestor:

1/16 (6.25%) DNA shared from Grandma
+
1/16 (6.25%) DNA shared from Grandpa
= ⅛ (12.5%) shared DNA total*

I also come from a family with complicated genetics. My grandfather and his two brothers married my grandmother and her two sisters (separately, it wasn’t a group marriage). This means I share a larger than usual amount of genes with all of my paternal second cousins because we all have four great grandparents in common rather than the usual two.

Also worth considering are siblings who marry siblings. Those cousins share both sets of grandparents. I’m not sure how you work out how closely related they are to each other in a way that takes that into account though.

(slightly ninja’d)

Not sure how it works out with 3 siblings having kids with 3 siblings, but when it’s 2 and 2, the cousins are apparently called ‘double cousins’ and share 25% of their DNA.

As the Big Lebowski said, this is fucking with old duder’s head. I share 25% of my DNA with my grandmother. Got it.

My first cousin shares 25% of his DNA with the same grandmother. Gotcha.

But there is also the grandfather. So even when you take the other side into account, that would offset to 12.5% which should then be added back together.

I don’t know why you go into the aunt part of the equation. It seems like you are double discounting.

IOW: why is the DNA between my father and his sister HALVED to 12.5% in your analysis? Should be the (roughly) same 25%, no? His sister (my aunt) are not another generation. They had the same parents.

I mean, I somewhat get where you are coming from. Take my sister and I. We have the same parents but do not have 100% the same DNA (again no twins). There will be some variance, but I don’t know how, for purposes of analysis of our children’s DNA (the first cousins) that it (my and my sister’s common DNA) drops to 50%. That is the most it could possibly drop, AFAIK, and under the most extreme and almost impossible chance.

ETA: And as I think it must be less than 25% for first cousins, but it seems that 12.5% would be an absolute floor. But let me think more and I might respond later.

Not my analysis. I linked to where I got it. I’m not claiming to understand it, just posting what I found.

FWIW, the article is written by someone at Stanford in the genetics department and the site (The Tech Initiative) appears to be affiliated with Stanford Medicine. So my money would be on it being correct. I’m guessing it’s more that they didn’t go deep enough into detail for how the math works out over it being wrong.

ETA, and every time I look at the math it reminds me too much of stats class to want to put myself through that again.
Stats was one of the classes I hated the most in college. I’d sooner do Diff Eq again than stats (and I’d actually really enjoy Multivariable Calc (Calc III in my school), that was, in all honesty, one of my favorite classes).

I don’t doubt that it IS correct. It is like most of these things on a board which fights ignorance for others to come along and help understand. Some of us need to see it intuitively. I appreciate your input even though you will keep me up for a few more hours ciphering.

I will give this a shot. You share 50% DNA with your mom. Your sister does, too. But the 50% each of you shares is not the same 50%. This is due to the recombination that happens during meiosis.

As you look at all the corresponding pieces of genetic material you and your sister have, each of you has a 50% chance of matching your mom on each piece. 0.5*0.5=0.25. So on average, you and your sister will have 25% of your genetic material matching from your mom. Same goes for your dad. Leading to the 50% average figure for siblings: 0.25+0.25=0.5.

Did that help? Now first cousins. The recombination during meiosis randomizes which 50% of the genetic material goes from parent to child, so you can’t just say that because you and your cousin each share 25% from grandma it’s the same 25%. Rather, the shared fraction between you and your cousin (if we repeat the piece-by-piece comparison we did for your sister) will be, on average, 0.25*0.25=0.0625. Double that to account for grandpa and you get 12.5%.

Let me know what I messed up, dopers

Snoe above has it best.

You get half your DNA from one parent, half from the other.

So your dad for example has 46 chromosomes - 23 pairs (Hence, “23 And Me, Inc.”). You get one chromosome from each of those pairs. So does your sibling. BUT! If we consider each pair as labelled 1a-1b, 2a-2b, 3a-3B, etc. you may get 1a and 2b and 3b and 4a and so on. Your sibling may get exactly the same or exactly the opposite… but odds are you’ll share about half. (Well, 11 or 12…)

(There’s also the possibility that genes during “copy” process may crosslink, and so you get part of say, 2a and part of 2b to make a whole 2, but that is not that common…Usually you inherit chromosomes intact)

It’s like flipping a coin - heads or tails? If you do it 23 times in a row and mark down - 1H, 2H, 3T, 4H…etc., do it twice, the odds are that about half the flips in the first set will correspond with the second set, and half will be the opposite.

So you get from your dad about 50% the same chromosome set as your sibling -statistically. Same applies with mom, so full siblings share 50% total. (With half-siblings, other parent(s) not related, 50% of one of the pairs are from the same parent, the other half of each pair of the chromosomes for each sibling is from someone with no match. So half-siblings are 25% genetics in common.

If one of those full siblings (lets say “Ann” and “Billy” are the siblings) has a child (with a partner unrelated to the rest of the family) then that child will have half their chromosomes in common with the parent. So Ann’s child (“Chloe”) will have 50% of the same chromosomes Ann has. Statistically, though, these are selected from both the ones Ann and Billy have in common, and those Ann has that Billy does not have. Again, statistically, Chloe could inherit exactly all the same chromosomes Ann and Bill have in common, or none of them - but odds are pretty strong it will be about half-and-half again, so Chloe will match Uncle Billy 25%.

Similarly, Billy’s child “Dan” will have 25% roughly chromosomes that match what Aunt Ann has.

So if we compare Chloe and Dan - Each has (again, statistically odds are ) half their chromosomes from from their parent who is a sibling, so half from either grandmother or grandfather - one quarter from each. But the odds they both ended up with exactly the same is vanishingly small. the average is in the middle. Each inherited 25%, but from different sets. so the average is 12.5%, the middle value.

Basically, consider, say, chromosome pair 1 -
Dad has 1a-1b; let’s say Mom has 1c-1d
Child gets one at random from each parent The possible pairs for a child are:
a-c
a-d
b-c
b-d
The child gets one of these pairs. If the first child gets one pair - let’s say, “1a-1c” the second child could get any one of these 4 pairs -
1a-1c(same as sibling)
1a-1d, 1b-1c - (one n common with sibling)
1b-1d (nothing in common with sibling)

Repeat this 23 times, and statistically the number of chromosomes in common will be very close to half. (Flip a coin fairly, see how likely you end up with significantly less than 50-50 heads and tails.)

To be pedantic, crossovers are a necessary part of human (mammalian?) meiosis, with (quick search) about 55 crosses on average in sperm production and 75 in egg production. There are still coin flips as you describe, but multiple coin flips per chromosome.

Two unrelated parents produce a child related exactly 50% to each of them. On average, the child will be related 25% to each grandparent (assuming the grandparents are unrelated to each other), but any given child will deviate from that 25%. Just by chance, due to crossover events in the parent, a child may be related a bit more to grandma than to grandpa.

In real genetics, this is a distribution. For example, with a dataset of full siblings I see that the range is from 0.35 to 0.60. A histogram:

``````Relatedness N
0.35-0.39   2
0.40-0.44  30
0.45-0.49 188
0.50-0.54 284
0.55-0.59 109
0.60-0.64   6
``````