Ancestry DNA

[Orlaith17]

I'm wondering if anyone on here has taken part in the Ancestry DNA through Ancestry.co.uk? I noticed they were offering DNA testing using a swab test, and expressed an interest. I have had an e-mail offering me the opportunity to purchase a test kit for £99. This includes the cost of the testing, which apparently can reveal your ethnic mix from the past 500–1,000 years or more. It also apparently will allow you to connect to a database of approximately 700,000 people, discover distant relatives you never knew you had. Just curious to know if anyone has done this, and how accurate the results are likely to be, based on a swab taken from inside your cheek?

[Ina]

Hi Orlaith17,

I took the Ancestry test about a year ago when it was on special for only $99. I also did the Family Finder test on familytreedna. Truthfully I haven't found any close relatives other than a 4th cousin that I had already connected with on Ancestry, so the DNA confirmed that we really were related. I also discovered that I am related to our very own Sarah (one of our admins here on Talkingscot) but we are not sure where the connection is.

Ina

[SarahND]

Hello Orlaith17 and (cousin!) Ina,

I have not taken the Ancestry DNA test, only the one with FamilyTreeDNA, but here is what I understand are the strengths and weaknesses of each one:

To find relatives with the Ancestry test, you need to have a public tree up on Ancestry (is this correct, Ina?) They determine who your DNA matches are, then look for a match in yours and your match's trees and then you get a report. They have a new tool called "DNA circles" that groups together people with common matches and common ancestors in their trees. The assumption (and it is a huge one) is that if you match someone, the match will be on the line where your trees coincide. That means that both of you have to have done your trees correctly and neither one of you has a brick wall just in the place you match… Ancestry does not give you any tools to see where and on which chromosome a given match occurs, thus taking most of the control out of the hands of the user-- this is where they lose my business since, as a scientist, I want to know what is going on under the surface!

The strong point with AncestryDNA is that the people who test are all interested in genealogy and have at least put some sort of tree up that you can check. The weak point is that you can't see where in your DNA the match occurs and can't group matches by those criteria.

FamilyTreeDNA groups your matches entirely on the strength of matching DNA and gives you an estimate of how close the relationship might be. This seems to go pretty much in order of the length of the longest block of matching DNA. The matches page tells you how much total matching DNA you have, as well as the length of the longest block. For example, from my own known family, the longest block I share with my children is about 267cM, with my sisters somewhat less, with a first cousin 142cM, a half first cousin once removed 44cM and on down to the lowest threshold of match which they have set between 7 and 8cM (my most distant match shares a block of 7.69cM). Beyond that point they feel the match may no longer be significant. I have 91 pages of matches, 10 per page.

Whether or not there is a family tree attached to any give person's kit is up to them. Some are quite extensive, some minimal and others have put up nothing. Some people have posted a list of surnames, or the names and dates of their most distant paternal and maternal ancestors. Everyone is required to have a contact email address visible to matches. If you opt out of that, you can't see your matches either.

The really helpful thing, from my point of view, is that you can compare your matches in a "chromosome browser" and see if the place where they match you is the same as where another person matches you. You can see this in a visual display, or in a table of numbers and can download the information into a spreadsheet if you wish. You also have a tool to show you which matches you have in common with any given person, so can attempt to group them by checking if they match you in the same place, etc. There is a matrix tool where you can compare up to 10 people and see if they match each other, as well as you. This will help to tell you whether you may be all matching on the same line or on different lines.

The strength of FamilyTreeDNA is in the analysis tools-- it is more of a hands-on site. I don't like their current tree display, for those who have put up their trees, but that could change. It is up to you to try to figure out on which family line the match occurs.


The weakness in both companies from the European point of view is simply that most of the people who have tested are from North America. My family came to America in the 17th and 18th centuries, for the most part, so any matches I have with branches still in the "old country" are quite distant. And, as I said, I currently have 91 pages of matches. In contrast my husband, a European, has only 11 pages of matches and the list is growing much more slowly. As more people from the UK and Europe take the test, those numbers will change.


All companies are basically in the beta testing phase of determining ethnicity. It is really more of a curiosity/cocktail party conversational topic than any kind of proof at the moment. They need to get many more groups to test from more places in the world, before they can begin to separate them by DNA and determine which groups any given person belongs to.

To get back to your question, however, on the accuracy of a cheek swab-- no question that it is accurate, as long as you have given them a good sample. The test recognized right away my children, my siblings, my close cousins, etc. and regularly a name pops up in my matches that I recognize from long-ago correspondence on genealogy forums. It's encouraging to see that the paper-trail seems to be correct in most instances! And it does give one a good stab at breaking down a brick wall or two. By testing a first cousin I have just been able to narrow down where Ina and I must match and was just crafting an email on the topic

This is probably more information than you wanted, but I hope it has helped.

Best wishes,
Sarah

[Ina]

Very well explained (cousin) Sarah.

With Familytreedna I have 90 pages of matches and with Ancestry only 38 pages. I have also joined several Name groups (for example, FergusonDNA group). As Sarah mentioned most of the people being tested are from North America, so many of these groups are contacting folks in Scotland and actually paying (through donations) for them to be tested.

My gg grandfather's brother immigrated to Canada in 1907 and through the FergusonDNA group, one of his descendants contacted me. I always knew they were somewhere in Canada but it was lovely to finally get to talk to some of them

Ina

[Orlaith17]

Thank you very much for the explanations, Sarah and Ina. I do have quite a few relatives with family trees on Ancestry, some closer, some more distant. Some I have connected with and know we are related and how through researching. So it's possible the test will connect me with people I already know I am related to. I'll give it some thought though as to whether it may turn up anything of interest. I haven't come across Family tree DNA. I'll Google and have a look there before deciding. Thanks again.

[SarahND]

You can always transfer your results to Family Tree DNA from Ancestry after the fact. I believe they have a deal where you can see your top 20 matches for free to get an idea of how it works, then "unlock" the complete set for a minimal cost $39 (abt £26), I think. That would get you the best of both worlds I have been tempted myself to try the other company too and maximize my match potential, but it's the requirement of putting up my tree on Ancestry that has me dragging my feet. If you already have one up though, you might think of testing there, then transferring the data to FTDNA.

All the best,
Sarah

[DavidMK]

I also have a question on DNA testing and perhaps one of our experts can offer some advice.

I have a friend who can trace her ancestors with some certainty to A. in the 1790s in Dumfries. My grtx3 grandfather W, was also from Dumfries and had the same surname as A. W passed down to his descendants that he was related to A , but how closely is not clear.

My question is, if my friend and I both had DNA testing done, would it provide definitive proof, or some probability of proof, that we are related to each other?

David

[Orlaith17]

That's an interesting question, David. Hopefully someone who has explored this DNA route can offer advice. I was more curious to try to find out our more distant genetic origins for no other reason than idle curiosity. My paternal grandparents' families both originated from Ireland. Relatives on gran's side of the family tend to have fair skin, blue eyes, brown or red hair. But granddad's side are different, dark haired, dark eyed and sallow skinned. I would say strong genes, as I can recognise at once a relative I have never met before as a relative simply by how they look, as we all tend to be similar in looks and colouring. My sister in law has speculated our Irish surname may actually have originated elsewhere....I'd love to find out.

[SarahND]

Orlaith17, it might be useful to look at this excellent chart on the ISOGG (International Society of Genetic Genealogy) Wiki: http://www.isogg.org/wiki/Autosomal_DNA ... ison_chart

If you scroll down fairly far you will find a row with the heading Overall accuracy and sophistication of the biogeographical ancestry analysis rated on a 1-10 scale with 10 being excellent and 1 being poor. You'll see that, of all the companies, 23andMe appears to have the best rating in terms of ethnic analysis, but "predictions in Europe are still not optimal." All companies are working at improving this, however, so wherever you test, your results will continue to be analyzed again each time they adjust the algorithm. Since your sample will be retained for at least 25 years, you will automatically get updates. This field is moving so quickly that surely more accurate results are only a few years down the road.

David,
I'm not an expert, but I have spent few years puzzling over my results and those of my other family members, so at least have some experience. Since the inheritance of DNA is random, you may or may not find a link between any two descendants of any given couple that far down the road. We don't actually inherit a straightforward mathematical percentage of DNA from our ancestors, e.g. 25% from each grandparent, 12.5% from each great grandparent, 6.25% from each great great grandparent, etc. It is more complicated than that. Some segments of the DNA are "sticky" and tend to get passed down in full for many generations. I have found distinct groups of people in my more distant matches who have all inherited the same string of DNA on the same chromosome, sometimes many generations back. If that is the case with your distant cousin, then you may well find a DNA link.

The major problem is that, with just two of you, there is no way of knowing what that link might reflect. The Autosomal DNA test can tell you that you have DNA in common, but cannot tell you where in your ancestry the common ancestor might be. If you know for certain (a virtual impossibility that far back) that this is the only line where you might be related and you do have shared DNA, then you can be pretty confident that you have found the link. Otherwise, you will know you are related, but can't be certain which line the relation is on.

If you don't show up as matches, it is not definitive proof that you aren't related, given the generational distance you are talking about. However, if you do match, the next step is to try to prove where that match is. This is where the tools at FamilyTreeDNA are very useful.

First, get a list of all those people who are matches in common between you and your friend.
Next, check on the chromosome browser to see which of those common matches share the same bits of DNA as the two of you do.
If those matches also display a family tree or list of surnames, check if the surname you are looking for is also in their tree, assuming they have researched back that far.
Email those who haven't posted a tree and see if they know anything about that line.

The more people you find who share the same string of DNA and also have the same ancestral couple in their tree, the more confident you can be of that connection.

So no quick answers, here! But, if you work on it, a combination of biology and old-fashioned genealogy can often give you the answer.

Good luck! If you do test, please let us know the results!

All the best,
Sarah

[DavidMK]

Thanks to orlaith17 and Sarah, for your replies to my query. Very interesting stuff, but I'm not sure if I will follow this up, at least for now

Thanks again
David