Do shared matches NOT all share the same DNA segments?

[melba_schmelba]

Following on from my obscurely titled thread, I thought I'd make a clearer one in the hope someone can answer this. Yesterday I came across this example, that challenged my belief, that all shared matches shared between two users, MUST share some DNA.

"I have Andrew P. who matches me on 2 segments at 31 centimorgans. He shows 4 shared matches between me and him, 3 on one segment A B C and 1 on two segments, D. But if I click through to A and B, they have a HUGE number of shared matches, including Andrew P, with one, two or three segment matches. But why were these HUGE number of shared matches also not showing on Andrew P's shared matches?"


The only way I can explain it, is imagining it this way-

I share segments with Andrew P on chromosome 1 between 30 and 45, and chromosome 4 between 55 and 71. But with user 'A' and 'B' I share segments on chromosome 1 between 39 and 59, and all of the other shared matches with A and B share segments with me on chromosome 1 between 46 and 76, so they match A and B, but they don't match Andrew. So Andrew is on A and B's shared match lists with me, but he shares no segments with most of the other shared matches?
  Presumably this is the reason why people would like a chromosome browser so they can see exactly what is going on with Ancestry matches and see which shared matches do actually share segments and which don't.

[ConfusedMason]

"Shared Matches" on Ancestry won't show any names with less than 20cm DNA in common with you or him.

He probably shares less than 20 with those people

[melba_schmelba]

"Shared Matches" on Ancestry won't show any names with less than 20cm DNA in common with you or him.

He probably shares less than 20 with those people

Hi Confused. Thanks for your input. But what determines whether someone shows on a shared match list is their total centimorgan connection to YOU being over 20 centimorgans, their link to the other user you are looking at the shared matches of may be as low as 6 centimorgans on one segment. So if these other people on my shared matches with A & B did actually match Andrew P at all they would be listed on my shared matches with him.
   I am now fairly certain that the overlapping segment situation is the only logical explanation for this example.

[melba_schmelba]

Actually now I’m not sure at all . Taking the fact that what is displayed regarding centimorgans and segments in shared match lists relates to US and not to the person we are sharing matches with - in the above example, it may well be that many of the shared matches between me and A or B share multiple segments with A and B but only one with me. So I had falsely assumed that the people showing one segment on the shared match lists with A and B should match Andrew P when that is not the case because we simply do not know the amount of segments or centimorgans shared between our shared matches, only from each user to us.
   So overlapping segments is not required for the above example, just different segments matching different people, but we can’t see the detail of that because Ancestry does not have a chromosome browser .

[Janethepain]

My head hurts, just thinking about this! I was actually about to ask a similar sort of question, but at a lower complexity level, but now I think I'll let things settle, and come back to it in a day or two!!

Janex

[melba_schmelba]

My head hurts, just thinking about this! I was actually about to ask a similar sort of question, but at a lower complexity level, but now I think I'll let things settle, and come back to it in a day or two!!

Janex

Hi Jane  . I think my confusion had all come from the wrong assumption that when you look at shared matches and it says

Shared DNA: x cM across y segments

I had assumed it meant - share between me, the match A at the top and that person B. But it doesn't. It just means how much I share with that one person B, it has nothing to do with how much I match the match A or how much that match A matches this other person B in the shared match list. So I well match to match A at only one segment. But person B might match them at 5 segments, we have no way of knowing, unless we ask match A to check, or we are the manager of match A's DNA account.

[Janethepain]

I think that for me the problem, is that, as a child you learned something over time, starting with the basics, slowly learning the ever more complicated details, constantly reinforcing the correct information, as you go on. It's like when I learned French, I must have studied it, on and off for about 12 years, and though I never use it, and am now 44 years post school, I can still pick up a book, or magazine and make some sense of it, though my memory of vocabulary is poor!!

In Genetic Genealogy, there is no 'gg for dummies' - you jump right in, reading about highly complicated topics and principles, and you do understand - but someone ask you about a closely related topic, and you realise how narrow your area of understanding is!  I am forever suddenly questioning myself, re my understanding of bits of information, that I assumed were right, but - well maybe not!!

Anyway, that's enough of my 'mature woman's confidence issues'!!  On to this specific subject!

So, the basics, as you say are that you share, say 100 cM , over say 9 segments.  Depending on the company, that may be only segments over a minimum size, and it may be the cM's making up those segments.  Or it may be the total cM's, whatever the size of any of the segments you have in common.  So you share those amounts with him/her, and they share the same amounts with you. You would be matching for the cM's & segments given, if you could compare your DNA on a chromasome browser, the one Ancestry doesn't have!!

Your joint matches are a different kettle of fish. What I think is that you both have to have matches with any 'joint' match, but I don't know whether these have to be the same matches, or if you could have a match on Ch 1, and the other party matches the third individual on CH22, would that still constitute a joint match?

One of the reasons fore my sore head, is that I have just discovered a low level match, who by comparison with a'gold standard' match on my paternal side, I had allocated to my paternal side - easy yeh!!  and yet 90 minutes later, I find him matched to a 'gold standard' match on my maternal side.  I would say that, though I am from the great Irish diaspora, on both sides, I have never found any evidence that they had 'met' before, until both sets of Grandparents moved to my parents and my birthplace, just before my parents were born.

So, anybody with 'Genetic Genealogy for Dummies' - I'll pay a fair price!!

[melba_schmelba]

I think that for me the problem, is that, as a child you learned something over time, starting with the basics, slowly learning the ever more complicated details, constantly reinforcing the correct information, as you go on. It's like when I learned French, I must have studied it, on and off for about 12 years, and though I never use it, and am now 44 years post school, I can still pick up a book, or magazine and make some sense of it, though my memory of vocabulary is poor!!

In Genetic Genealogy, there is no 'gg for dummies' - you jump right in, reading about highly complicated topics and principles, and you do understand - but someone ask you about a closely related topic, and you realise how narrow your area of understanding is!  I am forever suddenly questioning myself, re my understanding of bits of information, that I assumed were right, but - well maybe not!!

Anyway, that's enough of my 'mature woman's confidence issues'!!  On to this specific subject!

So, the basics, as you say are that you share, say 100 cM , over say 9 segments.  Depending on the company, that may be only segments over a minimum size, and it may be the cM's making up those segments.  Or it may be the total cM's, whatever the size of any of the segments you have in common.  So you share those amounts with him/her, and they share the same amounts with you. You would be matching for the cM's & segments given, if you could compare your DNA on a chromasome browser, the one Ancestry doesn't have!!

Your joint matches are a different kettle of fish. What I think is that you both have to have matches with any 'joint' match, but I don't know whether these have to be the same matches, or if you could have a match on Ch 1, and the other party matches the third individual on CH22, would that still constitute a joint match?

One of the reasons fore my sore head, is that I have just discovered a low level match, who by comparison with a'gold standard' match on my paternal side, I had allocated to my paternal side - easy yeh!!  and yet 90 minutes later, I find him matched to a 'gold standard' match on my maternal side.  I would say that, though I am from the great Irish diaspora, on both sides, I have never found any evidence that they had 'met' before, until both sets of Grandparents moved to my parents and my birthplace, just before my parents were born.

So, anybody with 'Genetic Genealogy for Dummies' - I'll pay a fair price!!

You are right Jane, this genetic genealogy is a bit of a steep learning curve, the problem is many of us think we have learnt enough, then we realise we didn't get it at all . We do have what I think we must consider our resident DNA expert here squawki11 who commented on my other thread
https://www.rootschat.com/forum/index.php?topic=808668.0

They stated that shared matches 'do not necessarily triangulate' which in layman's terms I think basically means the shared matches don't have to be related to each other at all.
   I was just looking at an example last night of shared matches with a fairly close relative where I share a lot of close matches and in that list some much more distant matches. I realised, actually I have always known that shared matches are not all related, logically, because I know some people on that shared list match my close relatives on one particular side, and some other of those more distant matches connect on a completely different side which I know are not related. So clearly there is a logical disconnect in my brain here  , as I continued to collect people in a shared match list as likely all related (i.e. descended from a recent common ancestor) when that may well not be the case .
   I have also discovered that, in one example I recently found, even if, say three people in a shared match list match another person at a low level, say 6-20 cM, even just above that, you still really can't be sure that they all share the same DNA segments unless you get all their DNA on gedmatch genesis and use the one to one match tool from each user to another. That is because when you get to the 6-20 cM level, you could be talking about one out of 128 ancestors in common, even up to one in 1024 ancestors in common for very low cM, so the likelihood that a DNA link from A to B, A to C, A to D will all be on different segments and come from different ancestors greatly increases as the fact you are all connected simply becomes a statistical coincidence due to the fact there are only a certain amount of breeding people in the UK at any one time - obviously that likelihood increases as you have ancestors who lived in similar areas close to each other. In fact, I suppose that is why Ancestry imposes the above 20 cM limit for shared matches .

[melba_schmelba]

Have been doing some tinkering with GEDMATCH Genesis this evening in a quest to see whether their 'People who match both or 1 of 2 kits' tool behaves in a similar way to the way we have now determined Ancestry shared matches works.

1. I picked a fairly small match from my match list of one segment.
2. I then put that kit number, and mine, into the  'People who match both or 1 of 2 kits' tool using the standard 10 cM minimum shared segment sizes.
3. I opened another tab with the 'One-to-One Autosomal DNA Comparison' tool.

4. I input my kit number and the other kit number I had chosen in 1. and noted down the chromosome and position on that chromosome on which we match.
5. I picked out a random kit number from the results in 2. and compared it to mine in the 'One-to-One Autosomal DNA Comparison' tool. I noted whether we matched on the same segment as I matched the first kit number in 4.
6. Repeated 20 times, even going down to a lower segment size.

THE RESULT

Not one of the other kits that both me and this distant match matched matched on the same segment as the first kit  . Six of these matches, using the lower segment size of 8 cM were over 20 cM, but still, none of the matches were on the same segment. So this seems a fairly strong indication that for all shared matches, especially the lower ones under 40 cM, we should verify that matches we think might share a common ancestor actually do by using these tools on GEDMATCH, or similar (now I think paid ones) on myheritage or familytreedna especially in the absence of reliable genealogical data.