QUESTION:
A reader contacted me recently to ask: Is it true that half-relationship DNA “fades” faster? And if that’s the case, why? Do you have a blog post about it somewhere?
The reader was referring to this tweet:
FACT CHECK:
To be honest, the precise meaning of the tweet is not clear, but as it is public and is confusing people, I’ll try to clarify.
Interpretation 1: If the tweet means to say that half relationships share less than full relationships, of course that’s true! We all intuitively understand that closer relationships share more DNA than distant ones and that full relationships share more DNA than half ones, on average.
Perhaps a better way to convey this concept is with the Shared cM Project Tool at DNA Painter. It presents summary data for dozens of relationships based on more than 60,000 data points collected by Blaine Bettinger.
As you can see in the snippet above, half relationships (to the left in the graphic) share roughly half the DNA as the equivalent full relationships (on the right). For example, a full first cousin (1C) averages 866 cM while a half first cousin (h1C) averages 449 cM. A full niece/nephew averages 1740 cM while a half niece/nephew averages 871 cM. This is exactly what we expect!
But averages are only part of the story. We also have to consider the range of variation, because DNA inheritance has a certain amount of randomness to it, thanks to recombination. Some 1Cs are reported to share as little as 396 cM while some h1Cs can share as much as 979 cM. Histograms of simulated data show the same thing: the averages for these two relationships are distinct, but there’s a lot of overlap.
Similarly, the amount of shared DNA drops by half, on average, with each generation. For example, first cousins once removed (1C1Rs) average 433 cM in the Shared cM Project compared with 866 cM for 1Cs. Again, both relationships have large ranges and quite a bit of overlap.
In summary, we expect half relationships to share, on average, half the amount of DNA of full relationships and also the average amount of shared DNA to halve from one generation to the next. We need to consider the ranges as well as the average, though.
Interpretation 2: Another possible interpretation of the tweet, suggested by the explicit mother–daughter example and the “fade faster” wording, is that we should expect to inherit less than half of our parent’s DNA when the relationship is half rather than full. That is false.
In the tweet’s example, the daughter shares less than half what her mother does (55 cM vs 360 cM), but that’s just normal variation. It’s not an indication of a large-scale pattern of inheritance that’s somehow special to half relationships. If we had hundreds of similar parent–child matches to compare, we’d see that sometimes the child shares more than half of the parent’s amount and sometimes less.
Again, we can refer to the Shared cM Project data to prove this to ourselves. Half 1Cs average 449 cM while their children (h1C1Rs) average 224 cM. If half relationships “faded” faster than full relationships from one generation to the next, the average for h1C1R would be less than 224 cM.
Another way to think about it is this: the mother’s body has no way of knowing that she only shared one grandparent with her cousin. Thus, there is no biological mechanism for it to have treated those segments of DNA any differently than her other DNA when she was making her daughter’s egg. Every segment had a 50-50 chance of being passed on, regardless of how the mother came to inherit it.
In conclusion, yes, half relationships share, on average, half the amount of DNA of full relationships, but half-cousin DNA does not “fade faster” from one generation to the next. These are important concepts to grasp when using DNA to solve genealogical mysteries.
~~~~~~~~~~
Do you have a suggestion for a DNA fact-check? Message me!
My great granddaughter and I share a match with a fellow, each of us with almost the same cMs.
Current social forces leave us wary of statements beyond our own immediate experience.
So we tend to judge the world by what we see. Contrariwise, we can only understand a more general experience of, say match cM, if we share them, for example in the Shared cM project. We can often only understand the full scope of human family activities if we join a Family History Society or an online community.
This is where the work you have been and are doing is so valuable. For example, contributing our individual experiences of prediction tools to a shared understanding of their performance.
I suspect they’re failing to realise that the amount of DNA you share with any ancestor drops rapidly with each generation. With half-relationships you start with less DNA as it is so it’s not disappearing any faster