Wikitree Cousins but no DNA match on Gedmatch

Question

I have found that many wikitree cousins I have are found not to be supported by DNA results. Many of them come back as zero matches on Gedmatch.

Any comments?

Answer 1

The relationship finder is only as good as the genealogy research done to create the connections (mistakes in, mistakes out).

Also, the DNA segments inherited can be different from cousin to cousin, hence the reason to use traditional & DNA research in combination in order to determine relationships.

Peter Roberts & some others here that have in depth knowledge on the subject can best address your question.

Answer 2

How distant are these cousins? ISOGG has a chart that shows the average amount of DNA shared between relatives. After about 4th cousins the chance of a match is pretty low. http://isogg.org/wiki/Autosomal_DNA_statistics

Comments

Even a few 2nd cousins will not match, and half of your 4th cousins will likely not match you on autosomal DNA:

https://www.familytreedna.com/learn/autosomal-ancestry/universal-dna-matching/probability-relative-share-enough-dna-family-finder-detect/

This is another reason why Y-DNA and mtDNA testing is important.  Too many genealogists use autosomal DNA as an alternative test which is a mistake.  It should be used as an additional test.

---

Yes, the intact (or near-intact) transmission of yDNA and mtDNA makes them exceptionally good tools for tracing ancestry, but newcomers to genetic genealogy also need to understand that each of these tools traces only one ancestral line. At the 4th cousin level, where the sharing of auDNA starts to drop off severely, the most recent common ancestors are 3G grandparents. We each have 32 3G grandparents (fewer if there were cousin marriages), and our yDNA and mtDNA lineages trace to exactly two of those 32 ancestors -- but we are likely to have auDNA from all 32 of them.

In principle, I could use yDNA and mtDNA from cousins of various degrees to develop yDNA and mtDNA trails for those other 30 ancestors, but that does require that I:

1. Locate living cousins in the appropriate positions on my ancestry chart (for example, to determine my paternal grandfather's yDNA, I would want to test a son of my mother's brother, or a male child of that son)
2. Get those cousins to test and share the data with me, and
3. Use auDNA matching to confirm that they really are biologically related to the expected extent.

Of course, it's easier if you have parents, grandparents, and aunts and uncles who are still living, or if large broods of children are customary in your family. But many of the people who get into genetic genealogy don't have those luxuries.

---

Over 30 of my cousins of various degrees have taken a Y-DNA test or mtDNA test for each of our shared ancestors:

http://www.wikitree.com/wiki/Roberts-7085#Y_Haplogroups_of_my_Ancestors_in_WikiTree

http://www.wikitree.com/wiki/Roberts-7085#mtDNA_Haplogroups_of_my_Ancestors_in_WikiTree

Except for some close relatives, I did not pay for their test or have to track them down.  WikiTree helped me find them.

---

Your record is impressive, Peter, but it is due in part to having a lot of relatively large families in your ancestry. (You have a heckuva lot more cousins than I do. There are a number of families in my ancestry in which only one child went on to have children.)

Answer 3

Peter is so right about using the Autosomal DNA test. It involves recombination, which means that when looking at say, you, and even a siblings DNA there will be/could be noticeable differences because of the randomness of recombination. If you want to see some explanations about this, Blaine Bettinger's Blog has some great examples.

Google Blaine Bettinger and atDNA.

Mags

Answer 4

In at least one instance, I've found that somewhat distant cousins (e.g., 5th cousins, as determined by genealogy) matched with me on DNA segments somewhat shorter than the Gedmatch matching thresholds. By itself, a match of something like 5.6 cM and 1026 SNPs isn't particularly good evidence of "identical by descent," but it at least hints that the genealogical connection is likely to be correct.

Comments

More than 90% of segment "matches" between 5 and 7 cM are not real.

Please see:

http://www.wikitree.com/g2g/199470/are-you-using-audna-segments-which-are-less-than-7-cm?show=199470#q199470

Say you ARE related to a 4th cousin (on paper) but you can't find a segment match larger than 7 cM.  So you lower your threshold to 5 cM.  It is likely that "matching segment" is not real.  It is a false artifact of the fuzzy matching process used to find matches since the standard auDNA testing can't tell the difference between auDNA from one's father and auDNA from one's mother.

Phasing your chromosomes using GEDmatch's Phasing utility can solve this.

http://www.isogg.org/wiki/Phasing

However you usually need to test yourself and both your parents or one parent and a lot of close relatives.  The "phasing" done by AncestryDNA is not the same.

---

I'm not suggesting that a match of 5.6 cM and 1026 SNPs should be used in triangulation. However, when a person is wondering whether the lack of reported DNA matches indicates an absence of a relationship, a match of that degree may indicate "don't give up yet!"

Also, when I see that a genealogically determined 5th cousin and I have two matches like that, it encourages me to look harder to find more family members to compare data with.

PS - Please note that people who don't have living parents can get rather tired of being informed that we need to get our parents tested.