Shared Matches at Ancestry

[Pauline]

I may be missing the obvious, but I am sometimes puzzled by the results I see when viewing shared matches at Ancestry.

For example, I have a DNA match with A and when I click to see our shared matches, I get a list showing B, C, D, E and F. Looking at shared matches with each of B - F in turn introduces 2 additional matches G and H. That I can understand (more or less!).

What confuses me is that later I may be looking at another match, say Z, and top of our list of shared matches is A.

So why does Z not show up in the list of matches I share with A?

 

[Bob Spiers]

Pauline,
The thought springs to mind it is all to do with match 'weightings'. Just as cousins have at least one line that is applicable only to themselves (and the greater the separation the greater the variances) so 'weightings' will vary individual to individual. So a qualifying match 'weight' to warrant notification to you may not qualify for another (and vice versa). Perhaps this accounts for the seeming anomaly you report?

But, as I say, it's just a thought.

 

[peter]

The only shared matches that Ancestry displays are those with 4th cousins or closer. It can be confusing if you don't realise that.

 

[Pauline]

Thanks, Peter. I guess that means I was missing the obvious - I did browse around the help pages but obviously missed that. Anyway, it all makes more sense now.

On the whole the shared matches feature is a useful one, and can sometimes help in pinpointing where the link might be. However sometimes, even where a good number of the shared matches have trees, I am as mystified after inspecting them all as I was before.

 

[Bob Spiers]

Just thought I'd mention at this point have now received my Ancestry results and my underwhelmed -as was- is now an overwhelmed as is! Less interested in ethnicity but no surprise 36% Great Britain, and my main Genetic Community West Midlands (very likely) -too true blue! However even in this short time have had two US contacts very reminiscent of MH (3 guesses) where someone believes I may be related to a 18th Century 'Parker' family from Lancashire - on the basis of similar based in Warwickshire. The other US contact even more bizarre but will get round to answering them, as is my way. Clearly both have picked up on my genetic data being added to the pool.

However I am happy with the 4 shared Ancestry links, especially as I know of 2 of the four by previous Ancestry contact and remaining two look interesting. Will follow up these before settling down to investigating the '67 4th cousins or closer'. Perhaps now I will be better able to understand the comments of others, or that might be pushing things a bit.

 

[Pauline]

I have a further question about shared matches at Ancestry that I am trying to get my head around.

If you have a match with person A , and person B shows up as the only shared match, and it appears from their trees that A and B are 3rd cousins, should you be focussing on A and B's shared ancestral lines in order to try and work how you relate to each of them?

 

[peter]

I have a further question about shared matches at Ancestry that I am trying to get my head around.

If you have a match with person A , and person B shows up as the only shared match, and it appears from their trees that A and B are 3rd cousins, should you be focussing on A and B's shared ancestral lines in order to try and work how you relate to each of them?

If you are related to both A and B, and they are related to each other, then it makes sense to start by looking at the lines they share - indeed, that's the main reason you'd want to identify shared matches in the first place.

However, it is also perfectly possible - though less likely - that you are related to both of them through different lines, ones that they don't share. After all, 3rd cousins only share one-eighth of their tree. It's also possible that they are related more than once.

At GEDmatch or FTDNA you would look at the shared segments for clues - if all three of you share the same segment of DNA then it suggests that you are all related through the same line. Unfortunately at Ancestry you don't have this option - you need to ask your cousins to upload their results to GEDmatch or FTDNA (or ideally both).

 

[Pauline]

Thanks, Peter. That clarifies the sort of lines I was thinking along. Ancestry do make it clear in their information that shared matches won't necessarily all share a common ancestor, but so far with all the shared matches I've investigated, at least those with whom I can pinpoint a definite connection, we do all share a common ancestor.

I guess things like how close the relationship is, and where folk lived (large towns or small rural villages) will affect the likelihood that you and your shared matches all share a common ancestor, or if you relate to some of them in different lines.

I've uploaded my DNA to GEDmatch and FTDNA, though I think at the latter you need to pay to investigate where your shared segments are. However, not everyone is willing to upload their DNA elsewhere - many seem reluctant even to respond to messages, which is a shame.

 

[Bob Spiers]

Having had a couple of months to check out my Ancestry DNA matches I have reached a few conclusions that I would now like to air. First praise for those matches where Ancestry hints are shared and for the most were already known to me through normal research means. Also others with High ratings where with a some effort - sometimes considerable -I was able to establish the connection and undertook some rewarding exchanges of messages and emails. Then of course those -still rated High and some Good - where I could not for the life of me establish a connection, and neither could the other Tree owner. Nevertheless it was well worth taking the time to investigate.

Then we come to the down to the less than good, and I have two comments on this score. The first will I am sure be echoed by others; the matches that have "No family Tree" link. If -and it is a big if - I ever get round to following through with these I may do as Ancestry suggest and check if they have an Ancestry Public tree; bearing in mine (as Ancestry warn) the 'Home person' may not be the same person as the DNA match. But overall I consider them a waste of space and I doubt any effort on my part will be rewarded. So for now I pass them by.

Then I come on to the more controversial topic of locked Trees (i.e Private Trees) which despite Peter's many (did I say many) warnings are in my view only a tad better than the No Tree brigade. Put another way I would rather take time out to view a Public Tree with 60 or less people showing than a few thousand contained within a Private Trees, despite a High rating!

Perhaps, just perhaps, when I have time on my hands (very rare) I might be intrigued enough to message a Private Tree owner. But if their responses are anything like the handful of Private Tree owners I have messaged over time (with one notable exception), I won't hold my breath awaiting a reply.

As I have said may times, all my Trees -including of course my main Tree used for the DNA link - are Public Trees and when I do contact other Public Tree owners I seldom find them lacking in their responses. Of course not all their Trees are 'pukka' but more often than not I commend their attempts, however misguided, and cautiously offer advice which they can take or leave.

So despite my two 'niggles' my final conclusion is that it was worthwhile taking an Ancestry DNA test and I am now venturing further with 'GEDmatch' and see what I can make of that. As this is a new venture I will be sure to read what Peter has to say in his Newsletters.

 

[peter]

Bob, if you follow the strategies I outlined in my recent Masterclass it won't make very much difference whether your cousins' trees are public or private - that's why those strategies are so effective.

As far as the cousins who have no trees are concerned, remember that we were all beginners once. If the amount of shared DNA is sufficiently high that they're likely to be 3rd cousins or closer, why not help them get started with their research? There's a good chance that you'll be able to figure out how the two of you are related.

There will be some, of course, whose sole motivation in taking the test was to find out about their ethnicity, and who have no real interest in going any further. But you won't know until you ask.

 

[Susan]

I have recently received my DNA results and like Bob I have found well over half my matches have no tree, including all three of my 3rd cousin matches. I have messaged two of them but after 2 weeks not had any reply. I recognised the name of the third - we 'met' years ago on Genes Reunited and have met in person on several occasions. It is nice to know that we have not made any mistakes and really are 3rd cousins

I also have a lot of matches where the tree has less than 10 people. Most of these joined Ancestry in 2017 and last logged in during August. Again, none of these have replied to messages. I wonder if some people expected their DNA results to name their ancestors?


One thing that I find interesting is DNA circles. Ancestry has found one circle for me, centred on my great great great grandmother, but no mention is made of her husband, yet we all have the same 3 x great grandfather in our trees. Is this normal for DNA circles?

 

[peter]

I wonder if some people expected their DNA results to name their ancestors?

It's hard to know what peoples' expectations were unless you ask them. The way that the tests are marketed is certainly misleading (in the way it focuses on ethnicity) but I don't think I've see anything which implies that you'd find out the names.

 

[jorghes]

One thing that I find interesting is DNA circles. Ancestry has found one circle for me, centred on my great great great grandmother, but no mention is made of her husband, yet we all have the same 3 x great grandfather in our trees. Is this normal for DNA circles?

As far as I can understand, DNA circles are meant to be created when people share DNA and a similar ancestor in their tree.
I did have 4 circles - 1 each around a husband/wife pair of 4th great grandparents (the wife’s circle has since vanished) most of those in the circle I had already identified as being descendants of that pair of 4th great grandparents - so that one more or less makes sense.
The other one was slightly more confusing, as it was around a second set of 4th great grandparents who were also maternal first cousins. Currently the 2 DNA Circles are focused around my 4th great grandparents, but I think they really should be focused around my 6th great grandparents from the pedigree collapse as that’s where I think the DNA link is really coming from!

So I'm not completely sure how they're working at the moment, other than I think they're a work in progress.

Although it can highlight some of the glaring errors that are in some people's trees, as it will show you where it draws the information about your ancestor.

 

[PhilGee]

The only shared matches that Ancestry displays are those with 4th cousins or closer. It can be confusing if you don't realise that.

Perhaps it will sink in at last! I keep forgetting this, despite the number of times you have to repeat it.

I have a two/three "shared person" group; two are father/daughter at 4th cousin (4C)/4C1R* level with the daughter not appearing as shared with the third person. I now realise I'm looking too far back for the connection, the only common surname/place suggests a possible 6C though the single segment is 33cM whereas the father/daughter are 23.1/22.3cM. There is an extensive (6624 person) tree attached to the DNA results and the two sets of locations diverge below the 6C level. This clearly suggests a "join" below the known Common Ancestors that have been found with the father/daughter - exposing a flaw in my "direct ancestors and siblings" tree attached to the DNA

I wish I didn't enjoy the puzzle solving so much, then I would be quicker to contact people (and lose the fun element)

Phil

* not to be confused with the spreadsheet notation

 

[peter]

I have a two/three "shared person" group; two are father/daughter at 4th cousin (4C)/4C1R* level with the daughter not appearing as shared with the third person. I now realise I'm looking too far back for the connection, the only common surname/place suggests a possible 6C though the single segment is 33cM whereas the father/daughter are 23.1/22.3cM.

When there's a single shared segment it could be further back than the length of the segment implies, and this is particularly likely if the shared segment is on the X-chromosome. My brother has a 44cM match with a cousin at Ancestry, though I don't. It currently looks as if our common ancestors were born at least 300 years ago, and I strongly suspect that the match is on the X-chromosome since my brother and I only share a small part of the X.

 

[PhilGee]

Peter,
I should have mentioned that, on GEDmatch, the segment is 41.4cM on Chr 19. Based on your brother's match and what I currently believe to be correct (I have yet to disprove it), the only common surname and place cannot connect before about 1700 - at 7C level.
Phil

 

[Bob Spiers]

As someone occasionally accused (not necessarily on the Forum) of blinding with science and lapsing into jargon when discussing computers or the use of same, I now find myself totally and utterly bewildered reading Phil's posting and Peter's response. I simply do not have a clue what they are talking about and - unless I am very much mistaken - there will be others similarly perplexed. I just about recognise their references to GED match segment numberings, but trying to comprehend their interpretations is quite beyond me and I tend to lose the will to live.

I remind of the KISS (keep it simple stupid) principle and suggest a 'lingua franca' approach rather than an argotic one will be best. Then perhaps it might become 'fun' for others. Just a thought!

 

[peter]

Bob, you ought to know by now that we all have 22 pairs of autosomal chromosomes, plus - in our case, being male - an X and a Y. And if you don't know what the abbreviation cM stands for you can always look it up.

If you read our posts again you'll see that we are making deductions about how far back the common ancestor might be based on our family trees - nothing complicated in that, surely? Perhaps you didn't realise that 1700 in Phil's post is a date? Or that 7C means 7th cousin? (He did explain the latter terminology in his previous post.)

Discussions of DNA won't be fun however they're described - it's jolly hard work, which is why I repeatedly emphasise in my newsletters that anyone who finds completing their My Ancestors page too tedious or complicated, but decides to buy a DNA test - assuming that because it's expensive it must be easier - is in for a really big shock.

What you read in my newsletters and on this forum about DNA is simplicity itself compared to most of the genetic genealogy blogs, books, and articles. That's why DNA testing companies focus on ethnicity when marketing their tests - it may be bogus, but it's something that anyone can understand.

 

[PhilGee]

Thanks Peter, that's better than I could have put it.

I would add that I try to follow advice received many, many ... years ago from a Royal Armoured Corps* Col. regarding abbreviations etc., but often fail. The basis of the approach is, at first use, to give the phrase in full immediately followed by the abbreviation within parentheses; thereafter, the abbreviation can be used as it has been fully defined within the document (which, in this case, is the thread). There are some cases where this can be relaxed, such as abbreviations believed to be in general use** - as in the first line of this paragraph - or common within the subject matter, such as cM..

In a previous post, after defining the 4C notation and using the same structure as father/daughter, I decided the context and subject would make it obvious the 1R was the "once removed" extension of 4C.

Phil

* As there is no further reference to this corps, there was no need to append "(RAC)"
** In the case of "Col.", through the use of UK military records if not within the language of the reader

 

[Bob Spiers]

Discussions of DNA won't be fun however they're described - it's jolly hard work, which is why I repeatedly emphasise in my newsletters that anyone who finds completing their My Ancestors page too tedious or complicated, but decides to buy a DNA test - assuming that because it's expensive it must be easier - is in for a really big shock.

I completely agree with all you say in this paragraph, especially your closing comment. You may recall I only ventured into DNA on a 'suck it and see' basis but having sucked, I am a fair way off seeing.

Bob, you ought to know by now that we all have 22 pairs of autosomal chromosomes, plus - in our case, being male - an X and a Y. And if you don't know what the abbreviation cM stands for you can always look it up.

As a DNA 'newbie' I made a point of reading your Newsletters and your Master Classes. I even bravely attempted the GEDmach "Learn More" articles (and watched the YouTube video links). I know about Centimorgans and Markers and was not overly fazed by Phil's use of abbreviated references without first using the long form. (I loved in particular his '4C/1R' (and his asterisked comment not to confuse it with a spreadsheet reference which I presume he means 4th column, 1st row).

What you read in my newsletters and on this forum about DNA is simplicity itself compared to most of the genetic genealogy blogs, books, and articles. That's why DNA testing companies focus on ethnicity when marketing their tests - it may be bogus, but it's something that anyone can understand.

Having sat through the GEDmatch articles I fully believe this and commend your attempts to simplify things, but sadly DNA does not really lend itself to such. It is like trying to simplify Einstein's E=m c2 (sorry I could not elevate the 2). As you say, no wonder DNA testing companies focused on ethnicity!