Did you see the new DNA percentages?

Question

Hi WikiTreers,

Today we changed how we list autosomal DNA test connections on profiles [example]. They now appear in descending order of biological connection. The family members who are expected to share the most DNA are listed first.

The estimated average percentage of shared DNA appears with each connection, and so does the expected X chromosome percentage, if any.

Most of you reading this will already know it, but I should emphasize that these numbers are only expected averages. Once you get beyond your parents, DNA mixing is random. I know that many people who see these percentages will assume that they are something more. If you have suggestions for the explanation on the help page please post here.

In addition to profiles, the percentages appear on DNA test detail pages and DNA Ancestor Confirmation Aids.

If you mouse-over a percentage number you can see the generational distance. For example: "Joe is 5 generational degrees from Jane; they are expected to share about 6.25% of their DNA." Since mouse-over isn't possible on mobile devices, the degree numbers also appear on test detail pages.

Showing these percentages will have some practical utility, but I think there is a less tangible appeal as well. I see these numbers as reminders that we have something more than family connections in common with our ancestors and cousins. We literally carry a part of our ancestors in each cell of our bodies.

If you like this new feature as much as I do, we have Ales Trtnik to thank. We have been wanting to add these percentages for years but I was a bit intimidated by the project. Ales was done within a few days of tackling it.

Do you see a problem with any of the numbers? They have not yet been reviewed by serious genetic genealogists. We need your help. If you have a good handle on how DNA inheritance works, can you review the percentages on your test connections and let us know if you spot problems?

During testing I thought I had spotted a problem, but it turned out that my family is just more inbred than I thought. :-) Ales's code accounts for multiple biological relationships between the profiled person and the test-taker. However, it is only examining eight generations from the test-taker. Deep endogamy won't move the needle. (Maybe down the road we could try to expand that a bit. That is, maybe we could still only show test connections that are within eight degrees of biological separation, but when calculating percentages perhaps a few more generations could be considered.)

Please post if you have questions, comments, or suggestions, and most importantly, if you spot errors. Thanks!

Onward and upward,

Chris

P.S. To answer one question that might come up about privacy: Test connections are made and percentages are calculated for all members who enter a DNA test for themselves. However, they are only displayed on a profile if the profiled person's family tree is public or if you're on the Trusted List.

Comments

I love the new DNA percentages, it really gives you an idea of how far back you can go before its lost so to speak. This leads to a question, if my son does a YDNA, besides his biological fathers side, will it also show his grandfather's (my dad's) side? I'm considering an mtDNA for me, but that is my mother's side only, it won't show the maternal line of my father's side. Before we order the yDNA for my son, we wanted to verify if it's both lines or just the one? Thank you so much.. Regina

---

Y-DNA is only transmitted from fathers to sons. You got no Y-chromosome from your father, and you have no Y-chromosome to give your son. So no, the Y-DNA test of your son will give you no information about your father’s line. If your father is passed, you could test a brother, if you have one, or a first cousin , son of a paternal uncle, or even an agnate second or third cousin if you can track one down.

---

Hi Barry,

 Thank you so much. That's what I thought but wanted to confirm. My father passed in 07, an only child. I do have 3 brothers. I guess I'll have to convince one of them to do the test. Again, I appreciate your help.. Regina

---

Re: my comment of November 13, 2020: is it still the case that DNA results are propagating off of private pages and onto public profiles connected through biological relationships? There was a post to the GGTT Facebook Group today (76,000+ members) where someone was claiming that their mother's DNA had propagated, shown as a half-relationship to her nephew, and that she was very upset that this was visible to the entire world. Comments were turned off, and the poster said they would contact WT, but I wanted to make sure WT is aware of this particular privacy concern.

---

I gather from Barry's comment that the person with privacy concerns did not give enough information to identify the profiles they are concerned about.

I wonder if they are seeing this information on a public profile because they are on the trusted list(s) for the private profile(s).

---

I saw that post, and it sounded like the problem was that the person included their aunt's information in a DNA confirmation statement on their profile.

---

That was part of the post, but it was definitely unclear from the post if that was the whole issue. That might have been, but the post was long and not particularly well-written. It was so long, and the confirmation statement a small part of it, that it seemed to me there were other issues.

---

Thanks.  We are looking for my Greatgrand father on my mother side. Now I understand that it would be more possible to find it thru a male in our family.  My nephew took one so it would be more possible to be found.

---

I have I think a classic generation jump, my father's sister married my maternal 3rd cousin (and I never knew that when I was a page-boy at the wedding) so the children are my 1st cousins paternal, and 3C1R maternal. Unfortunately I'm the only one with a DNA test.

---

lol, Gary, one of my sisters-in-law is actually 3rd cousin to my father, so my brother's children are actually also my 4th cousins.  Pedigree collapse continues.  

Answer 1

Thank you Aleš and Chris,

This has pointed out possible X chromosome connections which were easy to overlook.

I believe I found an error.  WikiTree shows that on average I share 100% of my X chromosome with my maternal uncle (which is incorrect).  My uncle’s one X chromosome is (usually) a combination of his mother’s two X chromosomes.  My mother (his sister) has an exact copy of her father’s X chromosome and her other X chromosome is a combination of her mother’s two X chromosomes.  My one X chromosome is a combination of my mother’s two X chromosomes. 

Comments

Thanks, Peter! We will get this fixed.

---

A son receives 100% of his X chromosome from his mother (which is usually a mix of his mother’s two X chromosomes), and a daughter receives 100% of her father’s X (and her other X is usually a mix of her mother’s two), but siblings almost never share 100% of their X chromosome (WikiTree currently indicates they do).

---

I have located the problem. X-cromosome from your test is correctly set to 50% on your uncle's profile. the error is in the other direction (replication of your uncle's test that is displayed on your profile).

It is corrected. I think it will be correct now and I hope nothing else will break.

It will take a day or two to rebuild the data. I will let you know when it will be updated.

---

Percentages are now updated. You can continue the validation.

---

Thank you. I have been trying to figure the X chromosome value out for some time. Peter explained it very well and after reading a few times I think I am starting to understand.

---

Have waited a few days for new data propagation, but I believe we're still displaying percentages incorrectly for the X chromosome. If I'm right, at the root of the problem might be that the Coefficient of Relationship applies only to the 22 autosomes; it isn't applicable to the X because chromosomal crossover occurs on the mother's gamete at approximately the same rate as an equivalent-size autosome, but doesn't occur at all on the father's X.

What still seems to be happening--because brothers are being displayed as 100% X-sharing...but oddly, not all of the time (on Peter's profile, note that Peter and Roberts-7273, Roberts-7275, and Roberts-11791 are all shown as full siblings, but Roberts-11791 is the only one displaying 50% X-sharing)--is an assumption that male children inherit the same X chromosome, intact, from the mother. But they don't.

A modeling done by Blaine Bettinger indicated an average 1.71 crossovers per gametogenesis for the X chromosome in females (and keep in mind that one crossover equates to two inherited segments; two crossovers to three segments); with 14% of the time no crossovers occuring; 1 crossover 30.7% of the time; 2 crossovers 32.7%; 3 crossovers 17.3%; 4 crossovers 4%; and 5 crossovers 1.3%. Given those figures, male children of the same mother will see either two or three X chromosomal segments inherited over 63% of the time, with four segments an additional 17% of the time.

If we're simply modifying the CoR to indicate no recombination for males of the same mother, that won't work. With a very broad brush from Bettinger's work--because I know, offhand, of no peer-reviewed studies that have been done specifically for this--we can expect crossover about 85% of the time and, with a fairly long chromosome like the X (about 156 million bp under GRCh38, and representing about 196cM), there seems no good reason to expect any two sons of the same mother would carry identical X chromosomes more than roughly 15% of the time.

I honestly don't know if a valid X chromosome algorithm can be built, at least one similar in simplicity to the CoR. What absolutely cannot be relied upon is anything like the X chromosome inheritance chart Roberta Estes shows about a quarter of the way down this page on a 2012 blog post. It's true that a male receives 100% of his X chromosome from his mother, but around 85% of the time that particular recombined mix of the X will be unique to that one child. I don't know if there's any reasonable way to extrapolate sharing across the same or previous genetic generations. Which is one reason none of the DNA testing companies use the X chromosome, by itself, for any matching purposes.

Answer 2

Totally, totally cool   This is so great and thanks to the team.

Answer 3

I'm excited to play with this more and see the value it will add.

My only picky thing is I would prefer full siblings have the ~ in front of the 50%. The range for full siblings is about 38-60%. I share only 38% with my full sibling. Siblings get different parts of the inherited DNA from their parents. The only time I say a relationship is definite based on cM without further exploration is parent/child.

Comments

And by the way, I think it's cool that it's adding percentages based on relationship in the tree. I see my half siblings are showing as ~25%

---

Ah, of course. Those tildes were a last-minute addition, and they're done in a rather unsophisticated way. If the percentage is less than 50% it gets a tilde. We could add another conditional that says if it's 50% but the test-taker is a sibling it doesn't get a tilde. I hesitate, though, because it will require extra processing. We might be better off removing all the tildes and expressing the uncertainty surrounding these percentages in other ways.

---

I expect many people will not understand the meaning of the tilde. Frankly, when i first saw it, the tilde was so small on my display I thought it was a minus sign. Say what? 

---

Several others, as well as myself, thought it was a minus sign too, Bennet. I was wondering how someone could have a negative amount of DNA 

---

I also thought it was a minus sign and found that confusing!

---

" We could add another conditional that says if it's 50% but the test-taker is a sibling it doesn't get a tilde."

Even with fullsiblings "50%" is only an average.  It could be anywhere from 0% to 100% (e.g., identical twins).

Answer 4

Aleš is a genius.  Thank you so much!!  I spot checked the percentages, comparing my own research with the % numbers for shared autosomal DNA on ancestor profiles.  I don't need to know how you did it, Aleš, but the % numbers match the amount of expected, average, shared cMs.  That is so amazing to me!

Just to let you know: There is no endogamy in those first 5 generations (starting with me and going backwards in time) in our twig.  Notice I did not say 'no known endogamy.'  That is the purpose of researching every twig and every descendant 'other' ancestors. (And, no, I will not be putting them all on WikiTree until I am cloned.) The 'possible' endogamy (shared ancestors in biological parents) begins with generation 9 (I am generation 3), where all descendant ancestors have not been identified and researched (by me).

Again, I am just simply amazed!  Thank you, Aleš and CEO Chris!!

Comments

Thanks, B! I'm so glad to hear that the percentages are holding up under your scrutiny. We know there are improvements to be made, but it's reassuring that we're on the right track. And yes, Ales is a genius.

Answer 5

Just wanted to extend another Thank You  to all who have contributed to this improvement. For people doing DNA-based research, this should be a real help. 

Comments

Thank you, Bill! That's much appreciated.

Answer 6

Thanks for this useful tool! I see that others have commented on X values, that was something that caught my eye this morning. Mine seem a bit off where they do play in...

Answer 7

When I look at my 2nd Great Grandmother 

Theresia is the second great grandmother of Laura (Confident)

1. Laura is the daughter of Shirley Ann (Hempen) Pennie (1928-2015) 
2. Shirley is the daughter of Viola D (Woerner) Hempen (1899-1985) 
3. Viola is the daughter of Martin Woerner (1851-1912) 
4. Martin is the son of Theresia Koenig (1814-bef.1914) [confident]
This makes Theresia the second great grandmother of Laura.

She does not have the DNA confirmed sign because I have to add the DNA confirmation write up.  I will work on that shortly.  .  Yet the person I match who is showing on her profile is not showing on my profile

see profiles  Koenig-679 and mine Pennie-22

Comments

Hi Laura,

It appears that you and Kathleen (Fields) Renziehausen are 4th cousins. DNA test info only propagates 8 steps from the test-taker (I don't believe that any of the test propagation functionality has changed recently). Since you are 10 steps away from Kathleen, her DNA test info will not propagate to your profile (and vice versa - yours won't propagate to her profile either).

Hope that makes sense.

Answer 8

This is AWESOME! Thank you!!!

I do have a question (or two). I do not know if this is really stupid or not. If so, I apologize in advance. 

My paternal great grandfather, Alenza Berryann, was a twin to his brother, Alonzo Berryann. Is there a way to know the approximate percentage of dna I share with my great uncle Alonzo? (I have not found any of his descendants yet.) 

Also, my maternal grandmother, Lois (Johnson) Whittingham, was a twin. Her twin, Lorraine Johnson, did not survive a day, but it would be cool to know approximately how much dna I could have shared with her.

Thank you for your help, in advance! 

Missy 

Comments

I decided not to do anything about the twins. For now we have only a sticker for multiple births, and that doesn't define if they are actual twins with same dna or just siblings. So there is no point in considering twins any differently than siblings.

If we will actually have DNA twins at some point, I could adjust the algorithms to compensate for that. But it will only double the percentages for some relatives.

---

Missy, the percent that you might have shared with your grandparents' siblings isn't known, although a range can be estimated.  As an example, I have two nephews & one niece that have tested, as have my parents (their grandparents). Nephew A shares 18.08% of his DNA w/his grandmother; Nephew B shares 21.46% w/his grandmother; Niece A shares 23.32 w/grandmother.  With their grandfather, Nephew A shares 29.26%; Nephew B 25.27%; Niece A 26.54%.  With me, Nephew A shares 24.39%; Nephew B 22.81%; Niece A 28.46%.

Looking at the shared cM project tool will give you ideas of the ranges.  If you want to discuss in more detail, you should start a new G2G thread.

---

Thank you, Aleš and Darlene.

I will check out the link. I just have not seen on g2g (yet) how twins are reflected in DNA and how it is inherited. I would LOVE to find descendants of my great uncle. The most dna I share with a cousin is 573 cMs with a 1C1R (she is here on WT) and her dna is on my profile and vice versa. That is the most dna I share with any cousin and we are not descendants of twins. I am interested in knowing if I could share even more dna with a twin’s descendants than without a non twin’s descendants. Maybe I will post a new thread.

Thank you!

---

Missy, we did have a brief discussion back just at the onset of the coronapocalypse  in this G2G thread. The bottom line is that unless twins are monozygotic--identical; forming from a single zygote, or egg--there's no genetic distinction among twins and full siblings.

In the case of monozygotic twins and our current microarray tests, they effectively will be indistinguishable. In other words, if your great-grandfather, Alenza, was an identical twin to your great-grand-uncle Alonzo, any of Alonzo's descendants would represent expected sharing amounts with you just as if Alenza had been their ancestor. Alonzo's children wouldn't look like 1C2R to you, but like great uncles/aunts. And their children wouldn't look like 2C1R but 1C1R to you; and their children in turn wouldn't look like 3rd cousins to you, but like 2nd cousins.

It gets more interesting if one set of identical twins marries another set of identical twins. But I have a feeling I've only confused things, so I'll go away now...

---

Edison,

You are awesome! That is EXACTLY what I wanted to know! Thank you so much for explaining it to me in terms I can understand.  I think I get it now. We do not know if they were identical or not, because uncle Alonzo died before we (my aunts, uncles, cousins and I) were born. So if I can find descendants of Alonzo, the DNA %s COULD prove whether my great grandfather and uncle were identical. Correct? That would be awesome!

I am still new to DNA. I can triangulate (thanks to MyHeritage and GEDMatch) and write DNA confirmations, but that is my extent of DNA knowledge. Thank you again! 

---

You're very welcome, Missy. Awesomeness is only a figment of my own imagination--terms like "curmudgeon," "contrarian," and things I can't write on G2G are far, far more common--but the thought is much appreciated!

And, yep, the atDNA sharing amounts found in the descendants of Alenza and Alonzo may give you evidence regarding the nature of their twinship. I say "may" because the strength of the evidence diminishes quickly with each generation. In general, assuming there's no pedigree collapse, the greater the generational distance the greater the possible variance, as stated as a probability, of DNA sharing amounts.

Put another way, the theoretical sharing between you and aunts/uncles is 25%; between 1st cousins, 12.5%; between 2nd cousins, 3.125%, and so on. So you want to catch 'em as early as possible!

In other words, it's going to be more likely that a 4th cousin (at around 0.195%, or 13cM) could look like a 3rd cousin (0.781%, or 53cM), than it is a 3rd cousin would look like a 2nd (3.125%, or 212cM). The closer to the source the better.

If, for example, you can locate a descendant of Alonzo who solidly shows on the paper trail as being your 3rd cousin, but the sharing amount is around 200cM or higher, then there's only a tiny chance that could genetically be a 3rd cousin. That would provide strong evidence--as long as the paper trail shows no additional paths for DNA inheritance matching yours except the parents of Alenza and Alonzo--that your great-grandparent and great-grand-uncle were identical twins.

If there's a generation closer to Alonzo, what on paper would be your 2C1R, the evidence could be stronger still. Because the approximate sharing at that level would be around 106cM. The expected amount for 1C1R--which is what you'd be genetically if identical twinship is in play--is around 425cM. You can see the gap is even larger. At 400cM and above, the odds are pretty near zero that, genetically speaking, the match is a 2C1R.

Hope you find the right cousin to test!

---

Thank you, Edison! I am on the hunt now!  I was lucky to find a much older cousin who was my 2C1R and the cousin had her dna tested! (She just passed away this year at 104!) So those older cousins are out there! 

Answer 9

Question from an average WikiTree user/contributor.

How to interpret the percentage against GEDmatch.

Looking at Thornton-425, I share 0.78% of DNA as does my (known) 2^nd cousin. But when I use the GEDmatch Autosomal One-to-one comparison with the 1.56% (tester/person), “No shared DNA segments found.”

Comments

Heya! I'm not sure I fully understand...but this is the kinda stuff I should be able help with a bit. Thornton-425 shows to be your 4g-uncle. At that relationship level, yep, you should share somewhere around 0.78%...which equates to about 53cM the way the testing/reporting companies show it. That's the same amount of sharing (again, approximately) that you'd expect with a 3rd cousin.

But since that uncle is seven generations removed from you, he's also probably six or seven generations removed from another test-taker who's contemporary to you and a direct descendant of your most recent common ancestor, that uncle's parent, your 5g-grandparent. If we say seven generations each, then that's 14 instances of meiosis (chromosomal recombination), which happened at each birth event.

That contemporary relative would be your 6th cousin...which is exactly what that person showing 1.56% sharing with Thornton-425 seems to be. She's Thornton-425's 4g-granddaughter.

Sixth cousins have only about a 4% chance of sharing any meaningful DNA at all, and if they do the theoretical average amount would be less than 1cM, so the companies are unlikely to report it. Is this making any sense?

The estimated sharing shown on ancestors' profiles are what you should have in common with the ancestor. In this instance the common ancestor looks to be Abraham's father, Thomas, so the shared DNA would be further reduced with each birth, each generation, back down the tree all the way from your 5g-grandfather. WikiTree has always put that "It may be possible" link on profiles up to eight genetic steps from you, which is equivalent at its farthest to your sharing DNA with a 6g-grandparent...or a 3rd cousin. Same eight birth events. Which is why that 6th cousin who appears on Abraham Thornton's profile doesn't also appear on yours. As far as DNA is concerned, you're too far removed for it likely to be detectable.

Edited: pesky typos...

---

And maybe I should add that there's an easy way to convert the displayed percentages to approximate centiMorgans for all but full sibling relationships (they're computed a little differently).

On all the reporting companies except 23andMe, multiply the percentage by 68. I mentioned the 0.78% sharing above; multiplied by 68 you get, voila!, 53cM. For 23andMe, you need to multiply by 68.81...they count stuff just slightly differently, plus they include the amount of X chromosome sharing, which no other company does.

That same formula will work for full siblings only if you substitute the ~50% sharing with ~37.5%. Long story.

---

I give you a gold star for the explanation! Understood and appreciate the time you gave towards it.

FYI, on FTDNA it says "those who share the same direct maternal ancestor within the last fifteen generations should have mt DNA results that match exactly."

Though this was not part of my question, I thought I would just through it out there because of being "too far removed."

thank you, jennifer

---

Yep. There are two kinds of what is called "uniparental" DNA because, well, we get it only from one parent. The Y chromosome is passed down from father to son to son, and mtDNA--mitochondrial DNA, which isn't part of our nuclear DNA at all but comes from the vitally important, tiny organelles that exist by the hundreds in all our cells except red blood cells--is passed down only from the mother. Neither of these go through recombination when the gamete is formed, as do our 22 autosomal chromosomes and the X chromosome. As a result, yDNA and mtDNA can be unchanged for many generations, and in fact mtDNA, partially because is so tiny and has so little material that can change, or mutate, can go unaltered for many more than 15 generations. A November 2018 research paper (Andersen MM, Balding DJ (2018) "How many individuals share a mitochondrial genome?" PLoS Genet 14(11): e1007774; https://doi.org/10.1371/journal.pgen.1007774) gives us an estimate of about one mutation every 70 generations.

Autosomal DNA, by contrast, recombines on the general order of 45 times with the formation of each egg, and 26 times with the formation of each spermatozoa. That means we get (not counting the X or Y chromosomes or mtDNA) half our DNA from our mother and half from our father...but we'll get about 146 chunks of autosomal DNA in total from our four grandparents--73 from each pair of them--because of the recombination that occurs during meiosis. This is what assures enough genetic diversity for the survival of the species and why no two siblings (unless they're identical twins) have the same genome.

As a result, at each step of full cousinship, the amount of shared DNA will be less by a factor of four. For example, assuming no pedigree collapse, your 2nd cousins will share only about 25% as much DNA with you as do your 1st cousins (about 3.13% compared to 12.5%), and your 3rd cousins only about 0.78% compared to 3.13%.

That's how the sharing diminishes so quickly: each step of full cousinship means we're introducing two additional birth events, two additional times the parents' genomes have recombined. With 1st cousins, there are four birth events: your parents, grandparents, aunt/uncle, and then your 1st cousin. When we get to 3rd cousins, we already have eight birth events in the mix; 5th cousins, 12 birth events, and so on. And when you think of it in terms of (very) approximately 73 different chunks, or segments, of DNA coming from each set of parents, it's easy to see how we all end up so similar, yet so different. And what allows Aleš to write the nifty algorithm that takes that genetic diversity into account. 

Edited to add: And thanks for the gold star for effort! Some might tell you to not encourage the strange-looking zoo animal because, whenever he has a free moment, he'll tend to write way too many words. But we in the zoo thank you for taking the time to read it!

---

I respectfully disagree with you. Your explanations were the PERFECT amount of words because you greatly enlightened me and I'm sure other readers will gain much knowledge not only from you but from the input of all the others above.

Answer 10

Recommended X chromosome reading:

http://smithplanet.com/stuff/x-chromosome.htm

Comments

Thanks for that!  Good read.

---

Hi Peter.

I added a link to this article from the help page. https://www.wikitree.com/wiki/Help:DNA_Test_Connections#X-chromosome_percentages

Do you think this addresses Edison's concerns or should we add more explanation?

Thanks!

Answer 11

Really cool! This correctly handles my double cousin Orra Marsh. 8 degrees away, but still 1.5% average DNA shared.

Answer 12

This is pretty awesome stuff !  Aleš and Chris and team you are amazing.  The numbers seem correct on the several profiles I have looked at for myself and my Uncle's DNA.

Answer 13

Thanks much, Chris and Ales...

As you said, Chris, 'Onward and Upward.' In this case though, it's descending order.

Answer 14

Thank you for this so very very much, I've been putting off working with DNA, and that has helped push me to make the leap into understanding and being able to work with it, I just did my first one for a person and their 3rd cousin for one of my profiles and "get it" so now can go add in the corresponding "XDNA statement" to go with the autosmonal one I did, and then go further with it on people beyond 3rd cousins... As this will allow so much great proving of lines further back..

Comments

Thanks, Kristina! That's nice to hear.

Answer 15

First off, I apologize for my complete lack of knowledge of all things DNA and hope this isn't a silly question.....

I am connected through 2 separate routes to my ancestors Giuseppantonio  Venditti and Arcangela  Criasia . They both show that I should have ~4.69%. Does that seem correct?

You can see that I show up twice on Giuseppantonio  & Arcangela's descendant lists but here is a shortened version of the chains:

Chain # 1: Giuseppantonio  & Arcangela's son, Agostino

-->his daughter Maria Michela Venditti

-->her daughter Maria Filomena Simone

-->her son Michele Andreola 

-->His son Nick Andreola Sr

-->Me

Chain #2: Giuseppantonio  & Arcangela's daughter, Antonia

-->her daughter Fioralba Maria Rossacci

-->her daughter Louisa Fino

-->Her son Nick Andreola Sr

-->Me

Comments

Hiya, Nick! There are never any silly DNA questions, only silly answers. Wait. That didn't come out quite right...

Anyway, if I have the inheritance chains correct, through your paternal grandfather, Michele, Giuseppantonio and Arcangela are your 4g-grandparents. And through your paternal grandmother, Louisa, Giuseppantonio and Arcangela are a generation closer: your 3g-grandparents.

To Antonia Venditti, your 2g-grandmother, her profile shows "~6.25%" sharing with you, which is correct...works out to the same as would be a 2nd cousin or great-grand aunt/uncle; ~425cM in centiMorgans.

Agostino Venditti, Antonia's brother, is also your 3g-grandfather. His profile is showing "~3.12" with you (or ~212.5cM). Which is also correct...though that calculation isn't taking into account that Agostino is both a 3g-grandparent and a 2g-grand uncle. This may be an example of an issue with the algorithm, I'm not sure (you should also show ~6.25% with Agostino, not ~3.12%), but we aren't concerned with that right now; it doesn't impact the sharing estimates with Giuseppantonio and Arcangela.

So... Giuseppantonio and Arcangela are both your 3g-grandparents and your 4g-grandparents, genetically speaking. That means you'd be expected to share (theoretically/hypothetically/approximately) about 3.125% down one inheritance path, and 1.5625% down the other.

Et viola! Give or take 4.6875% combined. So, yep, the ~4.69% is correct!

As a practical matter, this is also a good example of how the genetic-sharing impact of pedigree collapse lessens quickly over successive generations. I haven't looked further at your part of the tree, but let's say that a third child of Giuseppantonio and Arcangela--let's say Maria Rosa Venditti--has a descendant of a contemporary generation to yours who takes one of our popular microarray DNA tests. Let's also assume that her line has no pedigree collapse between your mutual 4g-grandparents, Giuseppantonio and Arcangela.

On paper, you would be 5th cousins, sharing a pair of 4g-grandparents. At that relationship level, you would only be expected to share about 0.0488% of your DNA, call it about 3.3cM. But you also have a path to Giuseppantonio and Arcangela as your 3g-grandparents. If we consider that path instead when comparing you to your new-found cousin, the two of you would be 4th cousins 1x removed, and would be expected to share about 0.0977% of your DNA, or about 6.6cM.

At this stage it's simple addition. Your new cousin is both 5C and 4C1R to you. So your expected sharing would be 0.1465%, which would work out to 9.962cM. As you can see, we're talking very small differences over the course of those six generations thanks to the dances of meiosis and independent assortment going on in the chromosomes at each birth event. The difference is so small, in fact, that the extrapolated calculations from our microarray tests can't even really measure it with great accuracy. The numbers would be more interesting between you and, say, a descendant of Maria Michela Venditti, but it is an example of how single instances of pedigree collapse don't affect current-generation sharing as much as some might expect. Which is a good thing. Otherwise, in regions where populations are small and isolated, there wouldn't be sufficient genetic diversity to support the species.

See? Going from your simple question about 4.69% DNA sharing to a statement about genetic survival of the human species... I told you DNA answers could get silly.

Edited: Had to take a second look at Venditti-14. I think--but don't know--that this, and Venditti-13, may be examples of the algorithm not taking into account all inheritance chains for the percentage calculations.

---

Grazie Mille Edison! 

I value your insight and GREATLY appreciate your ability to explain in such a way as to help me achieve a modicum of clarity 

As my paternal line has a number of these multi-connections, I've printed out your response & will refer to it when I re-examine the relationships noted with cousins I've connected with on the 2 DNA websites I use.

Otherwise, in regions where populations are small and isolated, there wouldn't be sufficient genetic diversity to support the species.

I guess that must be considered the 'Saving Grace' for this group!

---

Sei il benvenuto!

DNA trivia of the day: Trying to determine the underlying reason behind why meiosis does the chromosomal shuffle that it does (which metaphysics strikes me as kinda pointless, but that's just me) there are basically two camps. 1) The purpose is to have an integrated, self-contained method of DNA repair; i.e., if a coding gene is messed up on one of the two chromosomes, the gene from the other chromosome is swapped in; 2) The purpose is to assure adequate genetic diversity down to the smallest possible population numbers.

Me, I think the reason is an "all of the above" answer. And it's a pretty nifty mechanism. Without it, severe population bottlenecks could mean the extinction of any animal species, and we humans have been there more than once. Even without natural disasters like the Toba Event, studies (like Amos and Hoffman, 2009, Proceedings of the Royal Society B) have shown something we can intuit: humans deciding to migrate in small numbers and across large distances (think first leaving Africa and movement across the Bering Strait into North America) resulted in what probably were sequences of genetic bottlenecks. Without meiosis and independent assortment, we wouldn't be here today.

Easter Island is happy about it, too. Until the early-1960s when airline flights first started coming to Easter Island, the indigenous Rapanui population was the most endogamous in the world. Prior to that, only about 3% of all births on the island were from (mostly) exogamous unions.

What's also fascinating is that in bottleneck situations rare alleles are lost faster in the population than is heterozygosity, as if the DNA is favoring increased stability over variability. By removing the genes that are rarer in a given population, more room is made for that diploid repair function where a damaged bit of a chromosome can be repaired by the analogous bit on its sister chromatid before the ovum or sperm is created. In some measure, that's why today we identify certain characteristics with certain populations, like straight black hair in Eastern Asians or pale skin in Scandinavians.

In highly endogamous populations--like the Easter Islanders pre-1960ish--effective DNA sharing percentages can't really be estimated accurately; the pedigree collapse is almost total within just a few generations. On the other hand, mostly exogamous populations with inbreeding here and there will see the results of that pedigree collapse--results as in percentage sharings--diminish very quickly generation over generation. So we can calculate the estimated effect of, say, cousins marrying once or twice in a tree within the past several generations.

See how I managed to bring it back around to DNA percentages? <cough> And if you'll excuse me, I have to get to my weekly meeting of Oversharers Anonymous...

Answer 16

Thanks for addressing double cousins and other interesting relationships in the results!

Answer 17

This has been a great addition.

However, the addition of percentages underscores the need to find a way to extend the DNA info to cover more of the living cousins who have a very decent chance of having DNA matches.

In my family tree, I see DNA connections and now percentages on ancestors like Magdalena Brodhead (born 1704), who ostensibly would share 0.39% of auDNA with me if she could get a test. However, the profile of R Clark, a 3rd cousin 1x removed with whom I have a good auDNA match, shows no indication that I might possibly have a match with him, although our theoretical match is close to that same 0.39% shown for my long-dead 6G grandmother (and our actual match is larger than 0.39%). All the worse, R. Clark and I both have a high likelihood of carrying significant X chromosome material from our female MRCA Maria Marsh (as can be seen on her profile), but there is no hint of that on his profile or mine.

Now that it is possible to display DNA percentages, would it be possible to display DNA matches on the profiles of relations like 3C1Rs, 4th cousins, and any cousin who could be a significant X-chromosome match? These are relatives that few of us are likely to be personally acquainted with, yet they are the people with whom it might be most interesting to compare data. To be most useful, WikiTree DNA tools should help us identify them.

Answer 18

This is great for identifying potential relatives that I may have a DNA match with.

Has something gone a bit wrong in recent days?

William Wilkerson has two 100% matches:

https://www.wikitree.com/wiki/Wilkerson-2255

He is also a 100% match with Lillian Jensen (but Lillian is not shown on his page):

https://www.wikitree.com/wiki/Jensen-10055

Lillian is showing as a 1.56% match with an ancestor of mine:

https://www.wikitree.com/wiki/Lowson-18

As far as I can tell she only has her mother and father in her tree (but father is unlisted).

Thanks,

James

Comments

It is a glitch, They are working on it: https://www.wikitree.com/g2g/1188545/100-dna-connection-but-no-relationships

---

This problem should now be fixed. Thank you for reporting and helping, James and Lincoln.

---

Thank you for fixing and letting us know.

Answer 19

I found this question today and looking at my profile, I have three 100% X matches that I have no idea of who they are 

I have no relationship with them in the relationship finder either.

Something wrong?

Comments

Thank you for reporting this, Maria. Autosomal connections have been temporarily removed from profiles while we investigate. We'll post updates on this other G2G post.