Question

can my Autosmal DNA be matched against the Stewart family to see if iam related by the Fitzalan family

Answer 1

Hi, Bill. If the connection between the two lines is where I suspect--the FitzAlans who were High Stewards of Scotland circa the 12th century--the answer would be no. It might be possible with Y chromosome testing of an unbroken patrilineal line because the Y doesn't undergo recombination when the gamete is formed; it changes only via mutation, and with a minuscule number of coding genes--about 70 out of around 59 million pairs of nucleotides--it has a lot of room to freely mutate without causing potential health issues. A boon to genealogists.

Autosomal DNA, conversely, has roughly 71 crossovers, or recombinations, with each generation...give or take: 45 in the mother's genome and 26 in the father's. The net is that we get 50% of our autosomes from dad and 50% from mom, but even by our grandparents the combinations are no longer even. We may receive 28% of our DNA from one grandparent, and 22% from another. By our great-grandparents, naturally, the disparity can be even greater. And so on with each generation.

This is vital to maintaining genetic diversity when faced with small populations, but it really messes up our ability to use autosomal DNA genealogy very far back. 

Well-known population geneticist Graham Coop did some research several years ago, and the results showed that by around our 8g-grandparents, there was roughly a 50/50 chance that we would receive none of the autosomal DNA from any given one of them. By 10g-grandparents, the odds were over 80% we wouldn't. And for any specific ancestor, the odds effectively zeroed out at 13g-grandparents.

Plus, it isn't just the ancestors we have to contend with. We have to compare our DNA results to living or recently deceased test takers. So we also have to be concerned with the number of birth events: the number of times the genomes have recombined during meiosis. If I'm your 3rd cousin we share 2g-grandparents. If we go strictly by an expectation of getting 50% of our DNA from each parent, and then 25% from each grandparent, we should each share about 6.25% with an in-common 2g-grandparent. That would be 4 birth events for each of us. But we have to go back down the line, too; so 3rd cousins are actually separated by 8 birth events and would--again by the inexact averages--share only 0.78% of their DNA. But at least there's about a 90% chance that they share some measurable DNA.

Research by Brenna Henn, et al., in 2012 and then reinforced in 2020 by Amy Williams at Cornell, showed that the sharing probability drops off precipitously with each generation. By 6th cousins--so a shared 5g-grandparent--there's only a 4% chance that any two cousins will share any measurable autosomal DNA. By 8th cousins, it's around 0.25%...or only 1 in every 399 cousins tested.

Pragmatically and with any expectation of accuracy, it's very difficult to use autosomal DNA for genealogy beyond the level of 5th cousins. If the lines were endogamous at any point in history--a significant population bottleneck, for example, or royals purposely inbreeding--the task becomes even more unlikely because the test takers will have multiple possible inheritance pathways for the typically singleton small segments that match. Beyond 5th cousins, 92% of matches will be on a single small segment; beyond 6th cousins 98% will be small singletons.

At an average of 28 years per generation, sex-averaged, and if the test-takers are 50 years old, then my opinion is that the maximum reach of autosomal DNA for genealogy, at least without the comparison of whole genome sequencing results, is roughly 220 years, so call it early 1800s. And by "for genealogy" I mean the ability to accurately associate DNA with a specific ancestor. WGS could open things up another generation, possibly, but the math and biology is very much working against going any farther unless exhumed ancestral remains are examined.

Way more explanation than you wanted. But I haven't been able to talk DNA in several days. I've been in withdrawal.  

Comments

Edison, thanks for the reply, how do i get my Y DNA, can it be extracted from Autosomal, thanks Bill

---

It was a wonderful explanation - I really did undestand all of it

---

Bill, Y-DNA testing is separate from autosomal. Check with your DNA testing company to see if they can do it with the sample you sent them. Otherwise you'll have to send a sample to a company that does Y-DNA testing.

---

Bill, there is a large Stewart YDNA group project through FamilyTree DNA. My brother is a direct Stewart descendant and they've helped us narrow down which branch we fit into. The more people who test, the more they can triangulate and link matches.

---

Bill, sorry if I'm the one who's confused here.  But you do understand that Y-DNA traces only the direct male line, right?  Are you thinking that the Allans are descended from the Fitzalans?

---

Hi, Bill. Sorry I'm late back to the party. Been working very odd hours.

To Julie first: I didn't blink at the possible FitzAlan connection. I assumed Bill had some reason to think it possible, and "Fitz" was a Norman patronymic convention that was introduced to England along with William the Conqueror. 'Course you probably already knew that. From the French fils, or son. In some cases it was even used as a matronymic. Very confusing. Don't you wish we could go back and tell all the patro- matronymic users what a mess they were making for future genealogists?

Had to look this up: It was Alan fitz Flaad--who came to England with Henry I upon his succession to the crown--that first started the FitzAlan surname. I don't pretend to understand the linkage to the House of Stewart. That's all 11th and 12th century; I've never had a reason to go there.

Bill: For starters, you can upload your MyHeritage raw data here: https://cladefinder.yseq.net/. Our autosomal DNA results are not really a substitute for the various types of for-purpose yDNA tests, but what this can give you is an accurate estimate of your haplogroup...probably at a middling level down the hierarchical haplotree, but that can be helpful. It can't be used as a form of positive evidence, but it can be used as negating evidence. In other words, haplogroups that are a millennium old can't suddenly change along a family line.

As Becky mentioned, there are two Stewart group projects at Family Tree DNA looking at the surname's yDNA. One is a general surname project and has almost 2,000 members: https://www.familytreedna.com/groups/stewart/about. The other has over 1,100 members and is titled "Stewart (royal)": https://www.familytreedna.com/groups/stuart/about.

That latter project offers some interesting information--which I can't confirm, so I'm trusting--that makes uploading your MyHeritage data to the YSEQ Clade Finder potentially useful:

"Testers may join this project if they are believed to descend from the family of the Hereditary High Stewards of Scotland, often known as the Royal Stewarts. Male testers will belong to the L746/S310 subclade of the R-M269 haplogroup, and they will test positive for a number of specific SNP values, including L744, L745 and L746."

I don't know which particular markers (or SNPs, single nucleotide polymorphisms) are examined by MyHeritage during their autosomal microarray testing. I could research that, but it would take some time. However, the advantage you have is that the project has delineated a haplogroup subclade for the royal Stewarts.

Of note is that the FTDNA project is using the nomenclature not from FTDNA, but from YFull. Go figure. All the naming tedium gets complex, but on the plus side, the YSEQ Clade Finder uses the YFull database, so you should get apples-to-apples labeling there. Here's that subclade at YFull: https://www.yfull.com/tree/R-L744/.

Up at the top of that tree-like chart you'll see a whole string of tabs starting with "Home" and ending with "R-S775." These are, in hierarchical order, the parent clades of R-744/746/S310.

So... Even if MyHeritage didn't test any of the specific SNPs mentioned by the FTDNA project, they may have tested one upstream. Starting in order of most recent, those would be S775, CTS2501, DF13, L21, S461, and P312. P312 is five branches deeper than the extremely common R-M269, and I'm fairly confident that the MyHeritage data would get at least that deep in the tree.

If your MyHeritage data show a difference diverging from any of those branches (and you may have to look at the synonymous SNP names at the branches to make full sense of it), it looks like you can be reasonably confident that your patrilineal line is not from the royal Stewarts. Thus the negating evidence. Can save you a lot of time and money if that's the hypothesis you want to pursue. For example, I'm R-P312, but my next branch down the tree is DF27 not S461. So I can tell right there that I don't match with the Stewarts.

Well, P312 formed sometime around 4,800-5,000 years ago, so I would have a common ancestor with the Stewarts then. Probably doesn't put in line for the throne, though. 

If you want any second opinions about the results the MyHeritage data provide, just post again here and I'll see if I can be of any help. If you don't get negating evidence from the procedure--meaning you may still be a royal Stewart--we can also talk about other testing options. The biggie (no pun intended) is the Big Y-700 at FTDNA, but depending upon your results there may be a much less expensive second step before you decide to spring the $450 for the full sequencing.

Oh... And thanks, Jean. "I really did understand all of it" is not my typical reaction. And thanks to the notorious DAH for the best-answer star!