Myheritage dna

[melba_schmelba]

I have lots of puzzling matches on MyHeritage. I didn't test with them, I tested with FTDNA, but my results from there have been uploaded to MyHeritage.

Try this one for size. One match is on my mother's side and is a proven 4th cousin by paper-trail research. This man matches me on 32.6cM.

I also have two matches on my father's side who share some of my 3 x great-grandparents - proven by paper-trail research years ago - and these people match me on 35cM and 33.4cM respectively.

So far, so good. I know exactly who these people are. BUT - there is a Swedish woman who matches me on 35.7cM, a Swedish man who matches me on 32.6cM and a Norwegian man who matches me on 30.8cM. Believe me, I have traced all my ancestors back into at least the early 1700s and sometimes much earlier, and I have NO Scandinavian ancestors whatsoever in recorded historical time.

So how do you explain that?

Harry

This has come up before, all the DNA profiles I have on MyHeritage also have the same problem. I think there is something quite wrong going on in the MyHeritage algorithms - somehow it is looking at segments that are actually just common to the whole of NW Europe and taking them to mean a closer relationship than is actually the case.

[hdw]

So far, so good. I know exactly who these people are. BUT - there is a Swedish woman who matches me on 35.7cM, a Swedish man who matches me on 32.6cM and a Norwegian man who matches me on 30.8cM. Believe me, I have traced all my ancestors back into at least the early 1700s and sometimes much earlier, and I have NO Scandinavian ancestors whatsoever in recorded historical time.


Harry

Are these people matching on a single segment of 30cM, or several segments that add up to 30cM?  If it's the latter then I think it is way way back or a false match.  The imputation at MH may be a factor.

Had a match in Iceland pop up yesterday at MH who triangulates with some known cousins on a 13cM segment.  This shared segment can only have come down to us from Perth/Fife or Dorset/Devon.  But it is very close to a pile-up region.

I've just been checking out all my Scandinavian (mostly Swedish) matches on both FTDNAFamilyFinder and MyHeritage, and there is a definite pattern of shared segments of chromosome. And how about this - I have a FamilyFinder match called Anne-Marie Solveig Jacobsson who matches me on 45cM, and another match on MyHeritage called Eva-Lott Nilsson who matches me on 35.7cM, and all three of us match on chromosomes 1,2,5,17 and 18!

I have other matches on
chr.5,15, 18 and 20
chr.2,5,12 and 15 (a Norwegian)
and there are others I haven't checked yet.

Harry

[LornaHen]

I've just been checking out all my Scandinavian (mostly Swedish) matches on both FTDNAFamilyFinder and MyHeritage, and there is a definite pattern of shared segments of chromosome. And how about this - I have a FamilyFinder match called Anne-Marie Solveig Jacobsson who matches me on 45cM, and another match on MyHeritage called Eva-Lott Nilsson who matches me on 35.7cM, and all three of us match on chromosomes 1,2,5,17 and 18!

I have other matches on
chr.5,15, 18 and 20
chr.2,5,12 and 15 (a Norwegian)
and there are others I haven't checked yet.

Harry

Harry,
The total cMs on FTDNA are overstated as they include ALL segments right down to 1cM in the total, so to compare apples with apples at the total shared level you need to subtract the segments less than say 7cMs from that total.
You also need to look at the individual segments and their placement on the chromosomes, along with the shared matches rather than stating the total and an overall chromosome.
eg if a segment, not the total shared, is less than 10cMs* it may indeed be a false positive and not worth worrying about, even if it is real as it will very likely be a LONG LONG way back in time.
* some say 15cMs to be sure, but I generally work at 12 as mostly being real segments

Concentrate on any of the matches that have the little triangulation symbols to the right of the match on a shared match list, and within those, the ones that share at least a 12cM segment.
Which size could still be along way back.

Also you may know your own ancestors a long way back, but do you know all the descendants of the furthest back couples?


Lorna

[hurworth]

I agree with Lorna.

The other possibility is that some time back some of your relatives settled in Scandinavia.  We have a Y-DNA match who descends from a soldier who was an 'Englander' who settled in Sweden in the 1600s.  His sons had a dispute over his will which provided documentation of his origins.  In another line we have documentation of a Scottish relative who was in the army and also were granted land in Sweden.

 

[melba_schmelba]

I agree with Lorna.

The other possibility is that some time back some of your relatives settled in Scandinavia.  We have a Y-DNA match who descends from a soldier who was an 'Englander' who settled in Sweden in the 1600s.  His sons had a dispute over his will which provided documentation of his origins.  In another line we have documentation of a Scottish relative who was in the army and also were granted land in Sweden.

I don't think any Norwegian or Swedish close relatives are involved here. It is due to the fact MyHeritage have not yet correctly worked out a way to filter out 'Identical By Descent' segments, which may indicate a very distant (1000+ year) common ancestry, which Ancestry's Timber system has managed to do

https://isogg.org/wiki/Identical_by_descent

http://blogs.ancestry.com/techroots/filtering-dna-matches-at-ancestrydna-with-timber/

See also from 20:40 in this Rootstech 2019 talk with Ancestry DNA's head of science Barry Starr

https://www.rootstech.org/video/ask-a-scientist-from-ancestrydna

[hurworth]

I don't think any Norwegian or Swedish close relatives are involved here. It is due to the fact MyHeritage have not yet correctly worked out a way to filter out 'Identical By Descent' segments, which may indicate a very distant (1000+ year) common ancestry, which Ancestry's Timber system has managed to do

A lot of Harry's ancestry is from Fife seaports, so there could have been interaction with Scandinavia. 

But I'm not sure how large the largest segments are - if he's sharing 30cM with people over 5 segments then most of those segments would be very small, and IBS.

[melba_schmelba]

I don't think any Norwegian or Swedish close relatives are involved here. It is due to the fact MyHeritage have not yet correctly worked out a way to filter out 'Identical By Descent' segments, which may indicate a very distant (1000+ year) common ancestry, which Ancestry's Timber system has managed to do

A lot of Harry's ancestry is from Fife seaports, so there could have been interaction with Scandinavia. 

But I'm not sure how large the largest segments are - if he's sharing 30cM with people over 5 segments then most of those segments would be very small, and IBS.

But I also have these and have no even likely Scandinavian connections - and these only appear on MyHeritage not Ancestry or GEDMATCH, so I think it is a MyHeritage specific quirk.

[Mart 'n' Al]

"... interaction with Scandinavia." What a lovely way to put it.  As a teenager I used to dream about the blonde girl in ABBA. Does that count?

Martin

[hurworth]

But I also have these and have no even likely Scandinavian connections - and these only appear on MyHeritage not Ancestry or GEDMATCH, so I think it is a MyHeritage specific quirk.

In an earlier post in this thread I mentioned imputation, which I think is a factor in this.  One of the kits I manage gets some Scandinavian matches at certain segments, BUT these segments were already getting Scandinavian matches on these segments at FamilyTreeDNA before MyHeritage was offering DNA testing and FtDNA doesn't use imputation.

I suspect that for this person these are mostly IBS matches but a small number are valid, but from many generations back.