I've just had my Ancestry DNA results

[JessiT]

I'm new, looking for any help & suggestions. I'm a bit overwhelmed by my results. Ethnicity was no surprise, most UK and NW Europe. Is it normal to have over 8000 matches.

JessiT

[4b2]

I have 21 tests, which range between about 4,000 and 35,000 matches. There are a number of factors, the main one being where your ancestry is from - most tests have been taken by people in the Anglosphre.

It depends what you want to do. You can use DNA to prove your lines back, break down brick walls, discover offshoots born out of wedlock and more. As an example, my best friend's father took a DNA test and I quickly saw in his results lots of familiar matches. There was clearly matches that showed a descent from two of my ancestors. After triangulating with other tests I was able to determine that one of my ancestors born in 1803 had two children born out of wedlock to a married woman (by infidelity). Those birth ended in 1841 when he moved about 20 miles away. So, a little extra to put on his bio / rap sheet.

If you want to utilise matches to do the above, the rough guide is:

1) go though all your matches and add them into groups based on your sixteen gg-grandparents - you will find some listed as having a common ancestor (if you have a linked tree on Ancestry) - use those as a starting point to add people to groups - but then later used shared matches (under that tab) to group people who match with each other into common groups

2) also add people into groups based on shared matches when you can't see the link; currently Ancestry only gives you 24 groups, which is annoying as you need 48+ groups for all the matches - they will be adding 64 soon - at current I add a note to others like - #OVERFLOW - PAT - WELSH65#

Here is a brief video on the topic - https://invidious.privacydev.net/watch?v=81OsvG81SyU

One issue with Ancestry is that is only shows shared and triangulated matches >= 20cM. So if you have 20,000 matches, you'd need to go through all of them to find all the deeper matches. So I have developed a tool that just identifies the ones with shared matches, which are the most useful, as they show a flow of genes. It's important to only consider matches when you can see a common flow of genes and ancestry together. You will probably get some matches where Ancestry is flagging a common ancestor, but the slow of genes from shared matches does not overlap. You can also get some matches with no shared matches that have a common ancestor - in which case you can't be sure if the shared DNA correlates with the Ancestry.

I also have another tool that downloads all the trees of my matches and looks for common surnames, places and individuals. That is key, as you can have huge clusters. I have some with about 700 matches and it can be difficult to find a common thread in them.

This is useful for groups of matches for whom you don't know how they relate to you. You have created a group based on shared DNA. Now you need to pick out matches who share the same ancestors. Once you have that - that can give you a coordinate of people who are your ancestors or close relatives.

If you already have a good tree, these unknown groups could be a part of unknown ancestry, ancestry by a non-paternal event (e.g. infidelity), or descendants of one of your ancestors born via a non-paternal event.

Going down this route is very heavy work, but IMO the most rewarding work in genealogy. I still have multiple missing lines and lines I haven't confirmed with DNA. I am fairly sure I have three ancestors who were born via infidelity, which is around the average (according to studies). Of your roughly 128 ancestors for whom these DNA tests are useful, you'd expect about 2-3 to b born of infidelity. But generally you need quite a lot of tests to solve these mysteries.

[Ruskie]

Good suggestions from 4b2, but maybe a bit complex for a beginner to comprehend? 

Firstly, over 8.000 matches is to be expected.

Secondly, don’t take much notice of ethnicity estimates as they are only a very rough guide.

Do you have a subscription to Ancestry (or another paid site), and are you familiar with family history research in general? You will probably need some skills in this to try to find how you and your matches connect. You will probably find that many of your matches don’t know much about their family history so it might be down to you to research and find any connections.

If you have an extensive tree of your own, that will help greatly, and is probably vital. Also include siblings or your direct ancestors and maiden names, as many of your matches may come from “side shoots” rather than your direct line.

Start by looking at the matches you share the highest cMs with. A possible relationship is given eg “second cousin”, “fourth cousin” etc. Contact your matches - you might be lucky and get a reply, though I’ve found that most people do not reply. You might be even luckier and find that your match has a tree with names you recognize, or may be able to place you in their tree. Don’t take anyone’s tree as gospel, and it is worth doing your own research.

I’m sure others will have more suggestions to help you. Good luck.

[JessiT]

Thank you both for such very long replies.  I will be reading them both several times to fully take it all in. That is very helpful and considerate. one question, 4b2, why do you have 21 tests?

JT

[Lisa in California]

Hi JessiT, welcome to RootsChat.  When I received my first results (from Ancestry), I believe I had under 1,000 matches.  Quite possibly incorrectly, but I attributed the “low” number to the fact that all of my ancestors settled in Canada between c1815 and 1870.  (My American friends’ DNA results were much higher.)

I have built up a fairly well researched tree, but then I started a long, long time ago.  If you are just starting your tree, don’t be concerned with the amount of time others have spent on their trees.  So much is available on the internet now that in one month you could possibly have as much information as I had after five years of searching. 

After contacting matches and not receiving too many replies, I started contacting matches who had limited details for surnames in their trees.  I had fairly good success with this approach, either just by chance due to “nicer” people, or because they could see that we could help each other.  Note: I generally contact folks who, for example, have an Ovens ancestor who lived in Co. Fermanagh, Ireland, c1800s.  I don’t believe I’ve ever contacted a match just to see how we might be related.

I’ve also had success (hearing back from people) when I find someone who has a grandparent or parent with the same surname that I have in my tree, especially if the surname isn’t common.  [Of course, not only does the surname match, but I also look for the same geographical location, time frame, etc.]   Added: if I contact a match who has a Mumford parent or grandparent, they tend to reply rather than someone who has a Mumford ggggrandparent.

You possibly already know, but by contacting people, or even just through research, you may stumble across surprising results.  Are you willing to accept unexpected results?  In my own research, I found out that an ancestor gave up a baby prior to her marriage.  I was thrilled to communicate with the descendant of the baby as she had always wondered about the baby’s parents and together we figured this out (at least the mother’s name).

I hope the above makes sense - it’s late in California and I’m trying my best to articulate my thoughts.    Lisa

[ikas]

So I have developed a tool that just identifies the ones with shared matches, which are the most useful, as they show a flow of genes.

I also have another tool that downloads all the trees of my matches and looks for common surnames, places and individuals.

Would you be willing to share details of those tools you mentioned?

[Biggles50]

Now the fun begins.

What you will find is that your number of reported matches will increase.

It is often that a test can produce over 20,000 DNA matches but it is those who are reported as 4th Cousin or closer who are the ones to concentrate on first.

If you have a Family Tree on Ancestry, one method is to build it back to each of your 3xGGP’s and the siblings of each side, generation by generation.  Bringing each sibling as far forward as you can.

Within Ancestry DNA matches you can enable the Common Ancestor filter which activates Thrulines, a tree pathway between you are the chosen match.

A word of caution, please do not take these suggested pathways as 100% correct, work through each person veryfing each and everyone of them.

I normally suggest that Maternal and Paternal Cousins also take a DNA test, this will help in validating Grandparents, same thing with Second Cousins if you known of any or if you have DNA matches who turn out to be at this relationship level.

Going back too far as a start with DNA is not the way.  Initially the aim with DNA should be to prove lineage to each Great Grandparent and to their lines beyond.

Start with the high cM matches and work down.

Use the Ancestry Grouping feature to assign the match to a Group, Ancestry has just changed their system and now I suggest having a Group for each Great Grandparent.  You can place a DNA match in multiple Groups, it makes filtering a useful feature.

Ancestry now has their Pro Tools, which is an extra subscription which until one gets used to dealing with DNA matches may be left to develop until knowledge is gained.

As for Ethnicity, or Admixture as it is otherwise referred to take the results with a very large packet of Saxa Salt.  They are a very rough guide that may or may not be somewhere near or short of actuals.  Over time you will find the results vary, and if you were to upload your Ancestry DNA to other sites (which you easily can) then the Ethnicity given may well look like a totally different person.

Meanwhile if you wish to learn more about DNA then do buy Blaine T. Bettinger’s book all about DNA, it is not the most technical but it is easily the most readable.

Futurelearn also have a free DNA course that is worth signing up for.

Finally a BBC Sounds series “The Gift” is a series of six, thirty minute epidoes covering different aspects of DNA.

Any specific questions simply post them in Roots Chat

[4b2]

Thank you both for such very long replies.  I will be reading them both several times to fully take it all in. That is very helpful and considerate. one question, 4b2, why do you have 21 tests?

JT

We don't inherit the same amount of DNA that are used in these tests from each ancestor. I don't know too much about the mechanics, but you inherit a random amount of DNA from each parent within a range.

So if you have a chain of ancestors that happened to be at the low range of inheritance for each generation then you quickly loose the DNA that can be used to match with other people who inherit DNA. So in terms of matching 4th cousins - you can inherit enough DNA to match with actual 4th cousins, you can inherit DNA from shared ancestors but not the same DNA, or just not enough to match with any living descendants.

As an example, I have three tests from my maternal aunts and uncles. Each of them has a group of matches at around the fourth cousin level that another doesn't. Having the three of them gives a more complete picture of the DNA and matches my grandmother would have had.

So by having lots of tests I get more matches, including groups of matches that would be absent in the tests of my aunts and uncles (I am not tested).

In terms of the line that I have ten tests for, of descendants of my gt-gt-gt-grandparents, rather than having about 150 DNA matches from that line from one aunt, I have more like 1,000 matches on that line from the 10 tests.

Most important is to get tests from previous generations, since they have more DNA from ancestors. If I just had my own test I'd only have about 10% of the connections that I've been able to make.

Extra note on something that was mentioned above - that many matches have basic trees, e.g.



In these cases, open up the profiles of dead-end ancestors and then press the search button at the top right.



Then navigate to public trees and in about 90% of cases you'll find you can continue the line. Doing this has been able to reveal a lot of links.

Obviously the public trees contain a considerable number of errors - but we don't have the time to research the trees of our 1,000s of matches.

I also use Ancestry to flesh out the trees of matches from MyHeritage a bit. Since MyHeritage is not very useful for that.

One of my most important DNA breakthroughs was in researching the tree of a match who was about a 100cM match. He overlapped with two groups of matches which suggested to me he was a half second cousin by my great-great-grandmother who disappeared from my tree aged 20 after abandoning a child born out of wedlock. All  had was his name. He did not respond to a message. There were four possible candidates for his name living in the UK. So I decided to flesh out the trees for all four and see what I found. I started with one whose birth index entry had a mother's maiden name - since the shared ancestry is Welsh. Five minutes later I pulled up the 1911 census entry for one of this man's great-grandmothers. The first thing my eyes landed on were the place of birth Aberhavesp, Montgomeryshire in the place of birth column. The place of birth of my missing great-great-grandmother. I pull the image to the left and see the name - Annie Elizabeth - the name of said missing great-great-grandmother. 14 years she had remained incomplete in my tree. And without the help of DNA and a bit of sleuthing she would have remained. Her year of birth was a few years out in the censuses - probably owing to having married a younger man - which meant she had evaded my considerable combing through records for possible candidates.

[4b2]

To add to that. If you want to work out the identities of children born out of wedlock, or infidelity, you will tend to need to be able to triangulate shared matches with multiple tests.

So if you have a test for 2nd cousins of each set of great-grandparents, you can use those tests to better triangulate where your matches fit into your tree. e.g. one have a group of matches with common ancestry and you don't know how they relate to you. If those matches are also found among the matches of a 2nd cousin, then you can know those matches relate to a certain set of great-grandparents.

Repeat with 3rd cousins and you can narrow it down even more.

The drawback is, as mentioned, some of your relatives won't have inherited enough DNA to make these connections. This is why you want more and more tests if you really want to plumb the depths of DNA.

The main reason I do is that I have three ancestors born out of wedlock in the period 1861 to present. Despite huge work I have only been able to solve one of those. Of the other two - I know much of the ancestry that fits a couple of generations back into those gaps, but not how they all fit together.

Then another ancestor was born in India, and the records to trace the line will likely be in an unknown church somewhere over there. I have been able to reconstruct that line solely with DNA, but only for one parent of the said ancestor.

Between them these said events left me with 34.375% of my ancestry missing for 14 years. A lot of work with DNA has recovered 15.625%. This is why some people will need many many tests. Your ability to solve such mysteries is based on if close relatives also test.

As one example, I have a match on MyHeritage, who I have been able to ascertain is a descendant of my great-grandfather via illegitimacy. This great-grandfather himself was born out of wedlock. So that woman and her relatives are the only coordinate I could have to isolate shared matches on my unknown paternal line. However, MyHeritage does not have as many tests as Ancestry. But I was not able to offer her a free test with Ancestry.