Why is it important to know how much DNA half-siblings share (or how much any relationship should share)? Genetic genealogy is NOT an exact science. You might think it will give you a concrete answer but it doesn't.
DNA tests for genealogy (family history) are very different than a paternity test. That means you need to understand how much DNA to expect to be shared between relatives. Not understanding this can have serious consequences!
How much DNA do half-siblings share?
Half-siblings share approximately 25% of their DNA.
This is about the same as a grandparent and child share. Let's think about this.
You have four genetic grandparents. Two for your mother and two for your father. If you inherited an equal amount of DNA from each grandparent, that is 25% (1/4) of your DNA from each grandparent.
With a half sibling, you share one parent (two-grandparents). At most you could share 50% of your DNA with them. So why do you share the same as with a grandparent? This is where you MUST understand how DNA segments are inherited---which is how genetic genealogy works.
Genetic genealogy is not perfect portions of DNA but also, it isn't as simple as, "we share one parent so that's half my DNA."
You get half your DNA from each biological parent. That is the only "perfect" portion (FYI, there are some little quirks in DNA testing for genealogy where this can be a little off so sharing 49% with your biological mother is not significant. That has to do with technical issues for this type of DNA test. Just remember that genetic genealogy DNA tests will NEVER be perfect. As you're about to find out, it's all statistics.)
So how can two family members be estimated to share the same amount of DNA when they are such different relationships?
The Autosomal DNA Test and Inherited DNA
You are one whole person (you have 100% of your own DNA). Each parent can't pass the full 100% on to you. You don't have 200% DNA, that's just not possible!
A person has approximately 6800 cMs of DNA (cMs are centimorgans, which is a unit of measurement for DNA.) You can't inherit 6800cMs from each parent, then you'd have 13,600cMs (200%). You get your 6800cMs by inheriting 50% of your DNA from each parent. And, in theory, 25% from each grandparent.
You'd think that would mean you share 100% of your DNA with your full siblings, then. But you know you're not identical twins, you're two very different people.
This has to do with those other cMs you didn't get from each parent. This is where "in theory" and "statistically" become very important words. That's why an ancestry DNA test is approximate and more accurately, random.
In theory (or "statistically"), you and your full brother might inherit half of the same DNA from your two shared parents (this applies to full sisters, too, your biological sex doesn't affect the statistical average for this type of DNA, it's the same for any full biological siblings). The estimate isn't actually exactly 50% but we'll get to that in a moment, let's finish looking at the half-siblings.
Now let's apply this to your half brother or your half sister.
Statistically, full siblings should share DNA close to 50% so half-siblings should have less DNA, statistically half or 25%. It's almost a coincidence that's the same as a grandparent.
Speaking of grandparents, let's look at the random nature of DNA inheritance starting with grandparents because you do get 50% from each parent. But when we start looking at grandparents, great grandparents, a 1st cousin, etc., things change
Why Genetic Genealogy is "Random"
Statistically, we should inherit 25% of our DNA from one grandparent. That is what a mathematical model would show. In theory (as in, it is possible), you might inherit nearly all the DNA your mother got from her biological father and only a very small amount of DNA from her biological mother.. Then you'd match your maternal grandfather at say 49% and your maternal grandmother at 1%. That accounts for the full 50% you got from your mother but is hardly what is statistically predicted.
This extreme case hasn't happened but you see how wildly those percentages can swing due to the random nature of how we inherit autosomal DNA. When we're talking about half-siblings and even full siblings, it's even worse. Although it's unlikely you'd get all the DNA from one grandparent and none from the other, it's possible you and a sibling could inherit the opposite portions of DNA from each parent. (Once again, this is less likely to happen with full siblings, where they got exactly the opposite half from both parents but for a half-sibling, it becomes far more likely since it only occurs with the one shared parent).
If you were curious, I'm sure there is data out there on how likely this is but as much as I enjoy a nice romp with some statistics, I prefer doing genealogy so I can't tell you. Let's look at some better estimates.
Better Shared DNA Estimates
For explanatory purposes, I've gone with the traditional 50% and 25% estimates. But these aren't really the best estimates to use when talking about genetic testing for genealogy.
There are some better estimates for shared DNA so let's look at those for sibling DNA testing.
- On average it is estimated you should share 2613cMs with your full sibling (think of this as 33-40%.
Remember, there are various issues so we're not talking about a full 6800cMs of your genetic material being tested and remember random inheritance---that extreme example of sharing 49% with one grandparent.
By the way, this estimated number of cMs is based partly on pure science, partly on statistical modeling, and partly on real-life data.
If you only share 35ish% of your DNA with a full sibling, someone where you share both biological parents, you are going to share less with a half-sibling. At best, you will inherit exactly the same DNA from your shared two grandparents as your half sibling does. But this is statistically unlikely!
- The estimated shared cMs for a half sibling is 1759cMs.
BUT there is a range, determined mainly from that real-life data. Half siblings share between 1160cMs to 2436cMs. You'll notice that upper number is nearly as many cMs as you'd expect full siblings to share. That would be the case where you and your half sibling inherited all the same DNA from your shared parent (all the same DNA from the shared two grandparents).
It doesn't really help you use your DNA test results when I just give you these basic numbers, even with the explanation of why.
But there's a really easy tool for you to get this information for yourself.
How to Find GOOD Estimates for Amounts of Shared DNA
You can understand you should statistically share 50% of your DNA with a full sibling and therefore half of that with a half sibling. But when your sibling DNA tests vary from these amounts, you might want to panic.
Don't!
Instead, check out the free "Shared cM Project" tool at DNA Painter.
This online tool combines the theoretical with real-life data to give better estimates and most importantly, ranges, to help you determine if an odd amount of DNA is just a matter of random inheritance or something more.
It can take a while to really understand how to use the information but it should at least help you understand that not sharing exactly 50% of your DNA with an immediate family member is no cause for alarm.
Why is this important for genealogists?
If you're ONLY concerned with an unexpected amount of shared DNA indicating your close relatives aren't who you think (or you think you might have close family you didn't know of), the "why" isn't as important.
But if you want to do genetic genealogy, the random nature of autosomal DNA is really important.
It's also important to understand that if you're looking at parentage, siblings, or grandparents, it's very different than researching great-grandparents, 2nd cousins, or distant cousins. It is also a big deal when you have half relationships but not necessarily a big deal when you're dealing with double cousins.
Taking a look at the Shared cM Tool can actually help you start to understand this (well, the part that doesn't involve double cousins, at least).
There are a lot of ways to start to understand how genetic genealogy works. In part this has to do with how your brain works. For me, I notice how there's a cluster of close relatives, including sibling relationships, that could be confused with each other, but not with cousin relationships.
As you start dealing with cousin relationships, things become much more vague. Once you're getting to fourth cousin (or closer cousins removed several times) you might not share any DNA with them. For any type of cousin, there's a lot of chance for confusion. I can tell you this from experience, not just by looking at the table! I find using this table, instead of just the estimates on the testing site results, helps me understand why using genetic genealogy isn't a quick road to an exact relationship.
What to keep in mind when using DNA for genealogy
- Genetic genealogy is basically an exercise in statistics. It is not designed to provide exact relationships like a paternity test does.
- Autosomal DNA (and other genealogical DNA tests such as Y-DNA, where the Y chromosome is tested) involve a lot of RANDOM inheritance. That's how humans work.
- A DNA testing company provides you an estimated relationship to a match. A computer estimated this based on statistics. Some companies have additional tools that compare family trees---a computer did this as well---the results depend on if the trees are correct so those are "estimates," too, just not statistical estimates.
- It's a good idea to use traditional genealogy, not just genetic information. After all, an amount of cMs is meaningless without knowing about the people it is shared between.
- Use your human brain for all genealogy, genetic or otherwise. Relationships have to be possible. Remember many estimates are provided by a computer so they might not be possible or even reasonable, except statistically.
- Use tools beyond the testing company, like the Shared cM Tool at DNA Painter to help you work with your DNA results.
Finally...
Keep in mind that most DNA tools aren't designed to work with situations like double first cousins. There isn't good data for situations of "endogamy" or "pedigree collapse" because there is a big difference between having one random set of double first cousins versus having every second cousin being some type of cousin to every person in your tree.
You need to do some traditional research into your own tree (if possible) and those of your matches to help determine if you are dealing with situations like this and how extreme they are. Being related to your matches in more than one way will affect genetic genealogy, even if it doesn't affect the estimated relationship based on the number of centimorgans.
In conclusion, it is important to know how much DNA half-siblings share because it can help you accurately place them on your family tree. Keep in mind that DNA tests for genealogy are not exact, so don't expect concrete answers. Do your research and use your best judgement to avoid any misunderstandings.
Looking for more basic DNA info? Check-out our sister blog Genealogy Questions (try this post about "what is a 3rd cousin?").
