Connecting the dots; DNA match prediction and genelaogy

Idunn

Junior Member
Messages
8
Reaction score
0
Points
0
I am quite new to DNA-testing, but no beginner in genealogy. So far the match predictions from ftdna have been a bit off compared to the actual connections. All my matches are in the distant cousin bracket.

Here are some examples:

Relation A
Shared total: 55,01 cM
Longest segment: 27, 32 cM
Number of segments: 13
Predicted 3rd cousin -> Actually 6th cousin
_____
Relation B
Shared total: 46,03 cM
Longest segment: 18,49 cM
Number of segments: 12
Predicted 3rd cousin -> Actually 5th cousin
_____
Relation C
Shared total: 55,07 cM
Longest segment 15,08 cM
Number of segments: 21
Predicted 4th cousin -> Actually 6th cousin
__
Relation D
Shared total: 33,94 cM
Longest segment: 7, 72 cM
Number of segments: 12
Predicted: Remote -> Actually 9C1R
____
My absolute closest match is:

Relation X:
Predicted 3rd cousin
Shared total: 61,01 cM
Longest segment: 29, 16 cM
Number of segments: 12

Both the person behind the matching DNA and I are experienced genealogists, and we can not find the common link at all! Should such a result not consitute a very likely relation of at least 6th cousin? :unsure: Is it possible to get a match like this and not be related?
 
Even with 5th cousin one has 32 grand, grand... mothers and fathers. I don't believe you'll find a person who knows them all, even in modern world. 4th cousin requires to know 16 of them, and it is still very tough, not mentioning possibility of bustard kids. Most likely you don't know all connections.
 
Even with 5th cousin one has 32 grand, grand... mothers and fathers. I don't believe you'll find a person who knows them all, even in modern world. 4th cousin requires to know 16 of them, and it is still very tough, not mentioning possibility of bustard kids. Most likely you don't know all connections.

I think I am fairly knowledgable about the uncertainties of genealogy, it´s the uncertanties of genetics I am less familiar with. :rolleyes:
 
And you are wrong to assume one can´t trace all 32. It isn´t at all uncommon over here, at least with people who have made an effort. It depends ofcourse on many factors.
 
I'm one of those people who have an in-depth family tree spanning 12 generations with off shoots of maternal family trees. There are over 8,000 distantly related people discovered so far, and more are being added all the time. We forget about great, great, great, grand dads first wife who wasn't recorded and had 2 children, but our records show his 2nd wife as his first. There's no way to track bastard children and adoptions. There are a lot of hidden secrets looked in our families closets. I would trust DNA more than sketchy relationships and church records.
 
And you are wrong to assume one can´t trace all 32. It isn´t at all uncommon over here, at least with people who have made an effort. It depends ofcourse on many factors.

FWIW, I know the names (even maiden names for the women) of all 32 of my 3rd-great grandparents with high confidence. Two of them were adopted with unknown original parentage, though, so I can't quite get to all 64 4th-great grandparents. Obviously, there are some people with things like recent adoptions and estrangements in their family tree that keep them from doing so, but tracing 32 3rd-great grandparents isn't too uncommon in many countries.
 
There's no way to track bastard children and adoptions. There are a lot of hidden secrets looked in our families closets. I would trust DNA more than sketchy relationships and church records.
Second that.
 
FWIW, I know the names (even maiden names for the women) of all 32 of my 3rd-great grandparents with high confidence. Two of them were adopted with unknown original parentage, though, so I can't quite get to all 64 4th-great grandparents. Obviously, there are some people with things like recent adoptions and estrangements in their family tree that keep them from doing so, but tracing 32 3rd-great grandparents isn't too uncommon in many countries.
I would say that this is very commanding to trace so many ancestors. You need peaceful land, no wars, no big fires ravaging the cities and countrysides, or huge scale population resettlements. After WW2 it is hard to find old records in central and eastern Europe. To make matter worse, some documents were falsified and family names changed to hide racial identity and save lives. My favorit singer learned family secret, in age of 30 years, that he was of Jewish descent.
I guess US and Norway were lucky in this department.
 
Yes, ofcourse the level of difficulty varies greatly depending on which geographical region you are researcing. I too have a family line that meets "a brickwall" because all church records burnt in that particular area. When you do genealogy there will always be mysteries, which is why it is so fascinating. Hopefully DNA-testing will be a tool to uncover old secrets despite the lack of documentation. :rolleyes:

My question was not about genealogy per se, bet how one can interpret the autosomal results. In my case there is an unexpected one. From what I have learned so far the test can show the presence of rather old DNA, and the predictions from FTDNA seems too optimistic about the closeness of relationships.

Am I the only one who has this experience?
 
I am quite new to DNA-testing, but no beginner in genealogy. So far the match predictions from ftdna have been a bit off compared to the actual connections. All my matches are in the distant cousin bracket.

Here are some examples:

Relation A
Shared total: 55,01 cM
Longest segment: 27, 32 cM
Number of segments: 13
Predicted 3rd cousin -> Actually 6th cousin
_____
Relation B
Shared total: 46,03 cM
Longest segment: 18,49 cM
Number of segments: 12
Predicted 3rd cousin -> Actually 5th cousin
_____
Relation C
Shared total: 55,07 cM
Longest segment 15,08 cM
Number of segments: 21
Predicted 4th cousin -> Actually 6th cousin
__
Relation D
Shared total: 33,94 cM
Longest segment: 7, 72 cM
Number of segments: 12
Predicted: Remote -> Actually 9C1R
____
My absolute closest match is:

Relation X:
Predicted 3rd cousin
Shared total: 61,01 cM
Longest segment: 29, 16 cM
Number of segments: 12

Both the person behind the matching DNA and I are experienced genealogists, and we can not find the common link at all! Should such a result not consitute a very likely relation of at least 6th cousin? :unsure: Is it possible to get a match like this and not be related?

My confirmed match predictions in the last week.
I recently had a 25,31 cM shared with a person designated 5th cousin.
I emailed this person, we shared family trees and found we had a common line dating 1820 ( my maternal line with others paternal line.....the offsprings ) so for me the 5th cousin was correct even though it was a much smaller cM than what you have.
 
I think it depends where are you from.In some places everybody knows everyone and everything.Even the neighbours :D

I guess it's trickier in New World countries.
 
My confirmed match predictions in the last week.
I recently had a 25,31 cM shared with a person designated 5th cousin.
I emailed this person, we shared family trees and found we had a common line dating 1820 ( my maternal line with others paternal line.....the offsprings ) so for me the 5th cousin was correct even though it was a much smaller cM than what you have.

Interesting!
 
I have read somewhere that if a population have lived a long time in the same area, they are more likely to have shared DNA that gives a false impression of closeness, because of the long history of marriage within the group. The parents are from the same old stock so to speak, which surely is the case with my grandparents. However in this particular unexpected case there are two people from two very different groups that share this snippet of DNA. The other person have one single branch from the same geographical area we have, but the MRCA in that particular branch, if the research is correct, was born ca. 1590. That makes us 9C1R. (I found this connecton last night after I started the thread)

I´m thinking that 61 cM shared is too much to stem from the before mentioned MRCA. Perhaps the longest segment of 29 cM is from a fairly resent connection and the total is this large because of the old connection? The second largest segment we have in common is 8 cM. It can´t all be "ancient" DNA, can it?

Thoughts or experiences anyone?
 
FWIW, I know the names (even maiden names for the women) of all 32 of my 3rd-great grandparents with high confidence. Two of them were adopted with unknown original parentage, though, so I can't quite get to all 64 4th-great grandparents. Obviously, there are some people with things like recent adoptions and estrangements in their family tree that keep them from doing so, but tracing 32 3rd-great grandparents isn't too uncommon in many countries.

Aah, the benefits of some Italian laws....woman must retain their birth surnames after marriage........far easier for me to find my lines once I got that process in my head.
 
I have read somewhere that if a population have lived a long time in the same area, they are more likely to have shared DNA that gives a false impression of closeness, because of the long history of marriage within the group.
That's right, it will give you stronger signals for being cousins. Statistically it might show you as being 4th cousin, but in reality it was 6th. Other than that it is not a false because all group is related.
 
So, for predictive purposes would you say it would be ok to add 2 generations on average if you know the two samples to stem from the same "genetic area"?

I see that my reasoning regarding older or newer connectons makes no sense.:confused: If what you inherit is completely random, there is ofcourse no way of knowing if a large segment is passed down from one or several ancestors.

From the ISOGG page "Identical by Descent" (which I can´t post a link to yet, due to few posts) I see that only the segments larger than 5-7 cMs matter and is considered IBD. All the small segments can be ignored. Both the amount of SNPs and the number of IBD segments are used in calculating relationship distance.
 
So, for predictive purposes would you say it would be ok to add 2 generations on average if you know the two samples to stem from the same "genetic area"?

I see that my reasoning regarding older or newer connectons makes no sense.:confused: If what you inherit is completely random, there is ofcourse no way of knowing if a large segment is passed down from one or several ancestors.

From the ISOGG page "Identical by Descent" (which I can´t post a link to yet, due to few posts) I see that only the segments larger than 5-7 cMs matter and is considered IBD. All the small segments can be ignored. Both the amount of SNPs and the number of IBD segments are used in calculating relationship distance.

you might find this helpful


cM (centiMorgan) measures the lenght of shared DNA. More is usally a more recent ancestor:
http://www.isogg.org/wiki/Identical_By_Descent_segment
http://www.isogg.org/wiki/Centimorgan
https://en.wikipedia.org/wiki/Centimorgan

from my project guy
shared DNA under 8cM is uninteresting. 10 to 12 cM is the size where it begins to become interesting, even more if the matches have the same ancestral origin (in your case NE Italy). But the relation could be still over 500 years old
 

This thread has been viewed 10637 times.

Back
Top