Recommending DNA tests

Last week I met one of my cousins, N., for the first time. He is interested in family history and, among other things, we talked about autosomal DNA testing.

A genealogical DNA test examines the nucleotides – the building blocks of DNA and RNA – at specific locations on a person’s DNA.

An autosomal DNA test looks at the 22 pairs of chromosomes that are not sex chromosomes (the X chromosome and the Y chromosome are sex chromosomes). These 22 chromosomes are known as autosomes. Autosomal DNA tests can be used to find cousins, and this often helps geneological research.

N.’s brother has already taken a test with AncestryDNA. We talked about whether it would it be better to take a test with another company and how much DNA might he and his brother share.

My recommendation is to test at AncestryDNA. N. and his brother will share close cousin matches but will also share different matches with more distant cousins.

AncestryDNA has by far the largest number of users. In May 2019 it had 15 million kits in its database.

The second largest database is 23andMe, with more than 10 million people in April 2019. It is an expensive test for Australians, however, as shipping costs are $40 US per kit. As a result there are fewer Australian participants. I have not tested with 23andMe.

FamilyTreeDNA (FTDNA) is one of the smaller databases. It is estimated to have about 1 million kits in its database. It is not marketed as aggressively as the other companies. However, with FTDNA, you can test not only autosomal DNA but also Y-DNA and maternal DNA, and so you are able to trace your direct paternal line and your direct maternal line further generations back. You can also transfer your DNA data from other
testing companies to join the database.

MyHeritage is a newer testing service. It allows people to transfer their DNA data into the MyHeritage database. In May 2019 it had about 3 million kits in May 2019.

Another service is GEDMatch. It is not a testing site but you can transfer your DNA data there and compare data and find matches with people who have tested at other companies. In May 2019 GEDMatch had about 1.2 million participants.

AncestryDNA does not allow users to transfer data into its database. To participate you have to test at its site.

Given this variety of testing and analysis services, I recommend testing at AncestryDNA and uploading your DNA data to FTDNA, MyHeritage and GEDMatch. A small fee at each of these separate sites can unlock some more advanced analysis tools.

If N. is interested in pursuing his paternal and maternal genealogy he would probably also be interested in testing at FamilyTreeDNA and taking their two additional tests for Y-DNA and Mitochondrial DNA.

Your siblings have inherited part of your parents’ DNA that you don’t carry. There is some overlap with your DNA. You would expect siblings to share about 50% of their DNA. In fact there is some variation. Siblings have been found to share 2209 to 3384 centimorgans of 6800 possible centimorgans, from 33% to just under 50%.

Schematic representation of the transmission of the autosomes of the grandparents to three siblings. Retrieved from Wikimedia Commons, diagram created by Jordi picart – Own work, licensed CC BY-SA 4.0

It has been estimated that up to 10% of third cousins and up to 50% of 4th cousins will not share enough DNA to show up as match. (Third cousins share great great grandparents and 4th cousins share 3rd great grandparents.)

Having your siblings’ DNA analysed will increase your chances of finding cousin matches. This will probably help you to make progress with your family history.

Sources

• International Society of Genetic Genealogy Wiki
  • Genealogical DNA test
  • Autosomal DNA statistics
  • Cousin statistics
• Larkin, Leah. “Genealogical Database Growth Slows.” The DNA Geek, Leah Larkin, 22 June 2019, thednageek.com/genealogical-database-growth-slows/.
• Bettinger, Blaine. “August 2017 Update to the Shared CM Project.” The Genetic Genealogist, Blaine Bettinger, 26 Aug. 2017, thegeneticgenealogist.com/2017/08/26/august-2017-update-to-the-shared-cm-project/.
• DNA Painter: The Shared cM Project 3.0 tool v4

DNA: variations in DNA matches between companies

My father has a second cousin once removed, ‘GH’. I am GH’s 2nd cousin twice removed.

GH’s daughter administers his DNA kits. She asked me an interesting question. On the DNA evidence, is this the same person:

• Ancestry – AnneYoungAU on Ancestry (shares 33 cM with my father)
• Family Tree DNA – Christine Anne Young (shares 93 cM with my father)
• GEDmatch – Anne Young A947648 (shares 63.5 cM with my father)
• My Heritage – Anne Young (shares 56.3 cM with my father)

If they are the same person, how can the results be so different?

Also, on the DNA evidence, are these the same person:

• Ancestry – RRC001 (shares 110 cM with my father)
• Family Tree DNA – R. Rafe Champion De Crespigny (shares 134 cM with my father)
• GEDmatch – RD A587626 (shares 113 cM with my father)
• My Heritage – Richard Rafe Champion De Crespigny (shares 98.3 cM with my father)

DNA aside, of course, we know that they are the same two people. The first is me, the second is my father.

The reason that the amount of shared DNA seems to vary so widely has to do with different assumptions and techniques used to calculate genealogically significant DNA. As a species we share DNA with other forms of life, but much of this is irrelevant genealogically. Discarding the parts we share as living beings and concentrating on what we share as family relatives introduces different emphases. The result is apparent differences between DNA analysis by different companies.

AncestryDNA explains it in these terms:

“If you choose to upload your AncestryDNA raw DNA results to another website, they will look different for a number of reasons. Other companies do not use the same algorithms, database or methods to translate the data. Only AncestryDNA has access to the unique information available on Ancestry, including the family trees and records to help power the accuracy of the results. In addition, the proprietary algorithms that we use to calculate results are based on documented family trees and a one-of-a-kind, comprehensive database of DNA samples from around the world.”

Curious about the detail, I used DNAPainter, an online chromosome mapping tool, to investigate a little further.

I was interested to see the variation in what segments my father and I shared with GH so I painted the matches from MyHeritage, FamilyTreeDNA and GedMatch at DNAPainter using the default parameter excluding segments under 7 centimorgans and also experimenting with including all segments greater than 1 centimorgan (the lowest threshold allowed).

Reviewing the data for segments of 7 centimorgans or more I found that MyHeritage did not include matching segments on chromosome 1. The segment matched with GH corresponds to a known pile up area and is indicated with grey shading on the DNAPainter diagram. It probably indicates shared DNA without genealogical significance.

DNA profile comparing matches for GH with Rafe and Anne: 7 centimorgan threshold

Reviewing the data for all segments greater than 1 centimorgan I found that Family Tree DNA and GedMatch included these segments whereas MyHeritage did not. I shared some small segments with GH but my father did not indicating that I did not inherit these segments from the most common recent ancestor GH, my father and I share. In general 7 centimorgans is considered necessary before a segment is possibly inherited from a common ancestor.

DNA profile comparing matches for GH with Rafe and Anne: 1 centimorgan threshold

The MyHeritage estimates of shared DNA make sense in view of the 7 centimorgan threshold and by not including the shared DNA at the pileup region on chromosome 1. Unfortunately because the detail of the match at AncestryDNA is not revealed by that company I cannot comment on what data they chose to include or exclude for their match.

All this is a reminder that DNA matching, though a technique of great precision, makes certain assumptions, and operates within recognised limits. Its apparent accuracy will occasionally seem to vary. Be careful how you interpret DNA results. You are indeed related to the fly  you just swatted- you murdered a distant cousin – and you are related to yourself, since your DNA, though analysed differently by different companies is all yours. Somewhere between you and blowflies are your cousins and other family members, not all of whom you’d want to claim for immediate relatives.

Further reading

• “International Society of Genetic Genealogy Wiki ISOGG Wiki.” Identical by Descent, International Society of Genetic Genealogy, 23 Nov. 2018, isogg.org/wiki/Identical_by_descent.