How to use WikiTree and GEDmatch to find a mistake in your ancestry?

Question

As you browse through your family tree back a few generations you may see other auDNA testers show up (under DNA Connections) on those profiles.  If that other tester is a 2nd cousin or closer to you, then you WILL share auDNA with each other. If you do a one-to-one comparison in GEDmatch and you share no auDNA then there is a mistake in your (or their;-) ancestry.  This is because all known relatives who are 2nd cousins or closer share a detectable about of autosomal DNA.  See http://isogg.org/wiki/Autosomal_DNA_statistics#mediaviewer/File:Shared_cM_Project_v2_updated.png

Comments

or one of you is adopted? :)

---

But note that WikiTree doesn't distinguish between full- and half-cousin relationships, so if there are half-siblings in the family, some of your "second cousins" may be more distant than WikiTree indicates.

---

If one is adopted then their relationship status should be marked as "Non-Biological." https://www.wikitree.com/wiki/Non-Biological

A tester's DNA information in WikiTree is not associated with their non-biological relatives.

---

I believe Dennis was thinking in terms of a previously undisclosed adoption. While that's something that would completely invalidate a family tree, a person might not consider that to be a "mistake" in researching their ancestry.   :-)

---

One interesting variable affecting these comparisons is the effect of the inclusion/exclusion of X chromosome on the statistics.  I have a good DNA match with 4th cousin on lines of descent that alternate through female and male generations, so it is predicted to make a 25% contribution to one of my X chromosomes, and (I think) at least a 12.5% contribution to hers. My 53-cM match to this cousin includes an X-chromosome segment. The size of the match is on the high end for a 4th cousin, but if we didn't happen to share the X-chromosome inheritance, the match (one segment on another chromosome) is only 26 cM (more typical for a 4th cousin).

Blaine Bettinger's data input form doesn't ask whether the X chromosome is included in the comparative data, and it doesn't ask whether the pair of people would be expected to have inherited X chromosome material from their MRCA. Addition of that information could significantly refine these predictions of percentages shared.

ADDED: Regarding that 4th cousin, I can compare her data with data for four people on my line of descent from the MRCA (three of us are 4th cousins with her and one is a 4th cousin once removed). All four of us have an X-chromosome match with her, but only two of us (both full 4th cousins) also match on another chromosome. In this family, at least, the common inheritance of the X chromosome is a major factor in the statistics of matching.

---

Yes, The auDNA statistics chart is only for auDNA (not X-DNA).  Thanks.

---

I don't believe that's true. The upper end of the range for parents and children on that chart appears to be for a child who received an X chromosome from that parent.

The range reported in the chart for a parent-child match is 3266 to 3720 cM.

In my family, 23andMe (which Includes X chromosomes in its default reports) reports a match of 3538 between a father and male child (no X chromosome) and  3717 cM between a mother and a child (this includes the X chromosome, and it's essentially the same as the 3720 maximum value shown on the chart).

---

I suggest to ask the chart's creator https://www.wikitree.com/wiki/Bettinger-1

The chart is from a page on autosomal DNA statistics (which as you know is different from X-DNA) http://isogg.org/wiki/Autosomal_DNA_statistics

---

Thank you Peter. Excellent point as always! Thanks for sharing.

---

One of the things discovered from DNA testing, is that women have as many extramarital partners as the men.  In all the research surveys it was determined that 30% of the men had at least one extramarital partner, but for women that number was 8%.  But through DNA testing it has been proven that number is more like 30%.  The women just lied on the surveys, and lied to there partner, keeping there offspring of there extramarital partners a secret.

---

OMG, Too much information!  :)

---

Most GedMatch numbers I compare to on WikiTree are not matches at all. The ones I do match with are always several generations distant. I've never paid it much attention, but now I'm wondering about the possibility of error. Incidentally, I'm very careful to not publish anything I if I am not confident of my research.

I'm also wondering how my experiences compares with those of others.

Answer 1

Project Database Error now flags Template AuDNA_Match as an error

What is the status on doing DNA genealogy research inside WikiTree?

See 

• G2G
• Template:AuDNA_Match  has been in development mode the last year...

I think this template was excellent and one reason why WikiTree should be used and a common Family Tree is needed. 

I am getting less and less convinced that a common family tree is the best tool for quality genealogy.... but DNA genealogy must have something like this and also better integration with tools like GEDMATCH and FTDNA ....

Comments

Thanks Magnus. I can imagine using this template (or one similar) for each of my auDNA matches who are in GEDmatch and WikiTree.  However if I had dozens of them on my profile then I believe my profile might look like the Michelin Man;-)  My thinking is that having the information in the template would generate a more uniform display (less unusual citations by different users).

---

Why should they be visible for readers? 

What WIkiTree needs is more structured data and a Query generator ==> you can ask question....

1. Find all people who share a segment 4-5 with me in chromosome 12 and have a common ancestor or share the same sources as one of the people in my family tree

 

---

You and I might share a shared segment on chromosome 12 from 4,567,910 to 10,111,234 (which looks like cccatatcg.....ttactgacccc).

John and Mary might share a shared segment on chromosome 12 from 4,500,000 to 10,200,000 (which has nothing in common with you and I).

Same for Jack and Jill.

We need to rely on a site like GEDmatch to do the comparison of the raw data.

---

But this is a problem with more dimensions and more unknown. GEDMATCH has just person x is linked to person y on chromosome/segment z

To get the IRL connection you need to add some more "known" to the equation.... and then WIkiTree needs to do some homework...

1. WikiTree has if we have done our homework
   1. What locations we have ancestors from
   2. What sources we have to proof that
   3. What other people we have connection to that have done DNA tests

Someone needs to define what we need and what implications that have. Then someone should agree what changes should be done inside WikiTree....

or 

My feeling is that the easiest way to implement something like this "in Wikitree" is to do like Aleš did with all badly researched profiles and bad formatted profiles....   

---

You may be saying something similar to what DNAgedcom's Autosomal DNA Segment Analysizer does with In Common With files from Family Finder (Family Tree DNA). https://dna-explained.com/2014/01/09/introducing-the-autosomal-dna-segment-analyzer/

---

@Peter

No 

DNAgedcom's Autosomal DNA Segment Analyser has no genealogy research data.... they are just playing around with the DNA test data.... 

You need to combine all information sources you have and DNA is one, research in church books is another, locations found of people is a third..... 

My guess this is done better outside WikiTree as WikiTree lack a good development process as we still are not sure if its ok to use  Template:AuDNA_Match or not

Maybe Aleš found the way forward of his giant project to "clean" WikiTree and get rid of total nonsense profiles and bad formatting.... by doing all the work outside WikiTree

I have stopped adding Swedish Parish Categories in WikiTree and are now doing it at Wikidata and then WikiTree if people are interested can use that data.... 

Feels like the DNA part maybe also should be done outside WikiTree but use what's in WikiTree, GEDmatch, Ancestry and ..... but then we need to find an Aleš person who can set up a website and have the right tools and skills...

Answer 2

Thanks to Peter for sharing this easy-to-read graphic. I'm going to bookmark it!