Tuesday 21 August 2012

My Ancestry autosomal DNA test
Part I: Consent forms and admixture analyses

This is the first of two articles reviewing my experiences with the new Ancestry.com autosomal DNA test. A second article will focus on the matching process.

Ancestry.com announced  the launch of their new autosomal DNA test on 3rd May this year. The test has been in the works for over a year and around 10,000 genealogists in the US were given the opportunity to receive free tests as part of the beta-testing program. At launch the test was restricted to Ancestry subscribers and they were able to buy the test at a special introductory price of  US $99. It is currently only possible to order the test from Ancestry's US site. The following information is provided on their FAQs page.
11. I live outside the U.S., when can I get the new test?
We hope to make the AncestryDNA test available outside the U.S. in the future. We do not have a date for this yet. Please note: for those who live outside of the U.S. and choose to purchase the test on the U.S. site, the use of the dna.ancestry.com site and the DNA products offered are subject to United States law and the AncestryDNA Terms and Conditions and AncestryDNA Privacy Statement.
It was not clear from this if Ancestry were making the test available internationally but I nevertheless signed up to register my interest and I eventually received an invitation to purchase a test on 17th June which I immediately accepted. The kit cost me a total of $108.95 -  $99 for the kit plus $9.95 for shipping. My credit card bill shows that I was charged a total of £71.35.

Once the kit arrives from Ancestry you need to activate the kit online. As part of the activation process you are required to accept Ancestry's Terms and Conditions and you are also encouraged to sign an online Consent Form to participate in a research project known as the Human Genetic Diversity Project. The full Consent Form can be found on the AncestryDNA website. Roberta Estes has already written of her concerns about the way this form is presented and her blog post can be read here. I share some of her concerns.  The check box for the Consent Form appears on the same page as the box for the Terms and Conditions and it is presented in such a way that it appears to be part of the standard legalese that one normally has to sign when taking any such test. It is not necessary to sign the Consent Form in order to take the test but I am sure there will be many people who won't bother reading the forms and will automatically tick both boxes without understanding that they have agreed to contribute to a research project and they will, therefore, not have provided the necessary informed consent. I did take the time to read the Consent Form and decided that I did not wish to participate in Ancestry's "research". I can find no information anywhere online about Ancestry's Human Genetic Diversity Project. It is not clear whether such a project is ever going to exist or if Ancestry are simply trying to keep all their options open. I feel that Ancestry are being somewhat deceptive in choosing a name for their "project" which is so similar to the Human Genome Diversity Project, a legitimate scientific project run by Stanford University. Ancestry is a commercial genealogy company and has never before been involved in research, and it is difficult to understand their motives.

While the Consent Form is optional, the Terms and Conditions are not. You have to agree to these if you want your kit to be processed. The Terms and Conditions are mostly standard legalese, and the full form can be seen here. I did note the following condition which is of potential concern:
By submitting DNA to AncestryDNA, you grant AncestryDNA a transferable license to use your DNA, and any DNA you submit for any person from whom you obtained legal authorization as described in this Agreement, and to use, host, sublicense and distribute the resulting analysis to the extent and in the form or context we deem appropriate on or through any media or medium and with any technology or devices now known or hereafter developed or discovered.
I decided on this occasion to accept the risk and get my test processed, though I can understand that this condition might deter some people from testing with Ancestry. I am not a lawyer but I suspect that such a condition might be in contravention of the European Union's Data Protection Directive and the UK Data Protection Act. This would explain Ancestry's insistence that non-US customers are bound by US laws.

As part of the activation process you are also required to enter your year of birth. I did not want to share my year of birth with Ancestry so I gave my year of birth as 1900, the earliest possible date that can be entered. I presume Ancestry are not expecting much demand for tests from people who are over 112 years old!

I activated my kit and returned my test on 11th July. I received a notification that my sample had arrived in America on 19th July. The test was processed very quickly and I received an e-mail on 5th August to tell me that my results had arrived. Unfortunately this coincided with an exceptionally busy time for me and it also coincided with the Olympics so I have only now started to explore my results.

Admixture analysis
The results page is very different from the 23andMe and Family Finder results pages. It seems that the primary focus of the AncestryDNA test is the admixture analysis. The screenshot below shows my homepage (I have blocked out the names of my matches for privacy reasons).

When I click on the button to see my full results I am taken to this page:
As can be seen, my Ancestry "genetic ethnicity" percentages are: 58% Central European, 25% British Isles, 13% Eastern European and 4% unknown.

My own family history research tells a somewhat different story. I know the names and birth places of 15 of my 16 great-great-grandparents and they are all English. In this generation I have one illegitimate line which has prevented from me finding out the name of the remaining ancestor. The birthplaces of these 15 great-great-grandparents are:: Burrington, Devon; Bristol (2); Thornbury, Gloucestershire; Clapham, London; Colchester, Essex; Sandon, Hertfordshire; Limehouse, London; Bermondsey, London; Merriott, Somerset; Sydenham, Kent; Sydmonton, Hampshire; Kintbury, Berkshire; Westminster, London; Sherston, Wiltshire.

I know the names of 27 of my 32 great-great-great-grandparents, but I only know the birth places of 21 of these ancestors. All of my known ancestors are from the British Isles. These are the birth places where known: Ashreigney, Devon; Mariansleigh, Devon; Thornbury, Gloucestershire; Bristol; Great Yeldham, Essex; Preston, Hertfordshire; Sandon, Hertfordshire; Scotland (place not known); Hackney, London; Laverstoke, Hampshire; County Kerry, Ireland; Merriott, Somerset; Rickmansworth, Hertfordshire; Shoreditch, London; Ecchinswell, Hampshire; Welford, Berkshire; Kintbury, Berkshire; Salford, Bedfordshire; Holborn, London; Leighterton, Gloucestershire; Purton, Wiltshire.

My research into the ancestry of my 32 great-great-great-grandparents has been hampered by a further illegitimacy and the difficulties of researching more common surnames in large cities such as Bristol and London. This process becomes even more difficult if the ancestors die before the 1851 census so that there is no information on the place of birth. The London research should in future be easier now that so much data is being made available online on both Ancestry and Findmypast. Nevertheless, despite the gaps in my research, there is no reason to suppose that any of my ancestors in the last few hundred years came from anywhere other than the British Isles, and there is certainly nothing in my documented ancestry to account for the large percentage of Eastern European ancestry.

Frustratingly, Ancestry does not provide any information on the reference populations that are used to calculate these admixture percentages and without this information the results are somewhat meaningless. There are few samples from the British Isles in the public databases, and it may be that my low percentage of British Isles admixture is just a reflection of the paucity of data from the British Isles. A lot of the American testees have commented that Ancestry is finding unexpectedly large percentages of Scandinavian DNA. CeCe Moore has already commented on this issue in her review of AncestryDNA's admixture tool. Scandinavian DNA would be less surprising in someone of British ancestry than Eastern European DNA.

I have also had my DNA tested with Family Tree DNA's Family Finder test which provides admixture percentages with the Population Finder tool. My parents and my husband have also been tested with FTDNA. My husband is 100% English. His ancestors are mostly from Devon and Cambridgeshire, though he also has some ancestry from Somerset, Sussex, Herefordshire, Hampshire, Huntingdonshire and Surrey. The FTDNA tests provide a useful comparison with my Ancestry test. These are the Population Finder results from FTDNA for my family members (the percentages in brackets are the margins of error):

Europe (Western European) - French, Orcadian, Spanish 86.23%  (±10.03%)
Europe - Tuscan, Finnish, Romanian, Russian, Sardinian 13.77%  (±10.03%)

My dad
Europe (Western European) - Orcadian 93.00%  (±2.77%)
Middle East - Palestinian, Adygei, Bedouin, Bedouin South, Druze, Iranian, Jewish, Mozabite 7.00% (±2.77%)

My mum
Europe (Western European) - French, Orcadian, Spanish 88.97%  (±8.80%)
Europe - Tuscan, Finnish, Romanian, Sardinian 11.03%  (±8.80%)

My husband
Europe (Western European) Orcadian 93.57%  (±1.88%)
Middle East - Palestinian, Adygei, Bedouin, Bedouin South, Druze, Iranian, Jewish 6.43%  (±1.88%)

Family Tree DNA, unlike Ancestry, provide a very comprehensive set of FAQs about their Population Finder tool. They include in the FAQs details of the reference populations that are used for the analyses. As can be seen from the FAQs, the Orcadians (from the Orkney Islands) are the only population from the British Isles used for Population Finder. As discussed in a previous blog post the People of the British Isles Project has already found marked differences within the population of the British Isles. The Orcadians were found to be genetically quite distinct from the rest of the country, and are therefore not necessarily a good proxy for the British population. However, in the absence of other data from the British Isles FTDNA is simply providing details of our closest matches with other populations. A lot of people with 100% British ancestry are getting the strange Middle Eastern percentage in the Population Finder test for reasons which are unclear.

I have also tested with 23andMe. They do not as yet provided detailed admixture breakdowns and according to their test I am 100% European.  23andMe, like FTDNA, do at least provide details of the reference populations they use. They also have a new feature which provides an estimate of your Neanderthal admixture. According to 23andMe an estimated 2.5% of my DNA is from Neanderthals, placing me on the 39th centile.

It is clear that these admixture analysis tests are still in their infancy and are at present not particularly informative, especially if you have already done a lot of family history research. Both the Ancestry test and the FTDNA Population Finder are still in beta-testing. It is somewhat surprising that Ancestry have chosen to focus so heavily on the admixture part of their test when the results clearly have very little practical application and are potentially very misleading. I am certainly not going to start looking for ancestors from Eastern Europe or the Middle East! I believe these percentages are only appearing because of the lack of British reference populations in the public databases. As Blaine Bettinger has pointed out we have two family trees: a genetic family tree and a genealogical family tree. At present our research is directed by our genealogical family tree. However, admixture tests should eventually become more accurate as more reference samples become available. The data from the People of the British Isles Project should make a big difference to admixture predictions for those of us with ancestry from the UK and Ireland. It might even one day be possible to take a test which will tell you what percentage of your DNA is from Devon or Cornwall. Ancestry, 23andMe and Family Tree DNA have all promised to update their admixture percentages as new information becomes available so, even if the tests are not particularly useful at present, they could still have some value in the longer run.  And of course if you are adopted or don't know anything about your ancestry for whatever reason then some information is always better than none. For anyone interested in taking a DNA test primarily for the admixture percentages the new Geno 2.0 test from the Genographic Project will be the test of choice, as it includes more ancestry-informative markers than the other tests. The autosomal DNA tests from FTDNA, 23andMe and Ancestry have a much more practical application for finding matches with genetic cousins. I will look at the matching process for the Ancestry test in my next blog post.

Update 25 August 2012
Ancestry have now advised that the confusing adoptee relationship prediction was the result of a lab error. CeCe Moore has further information on her blog.

© 2012 Debbie Kennett


Genealem said...

That is very disturbing. We are seeing a lot of problems with AncestryDNA's autosomal test. Of course, there were issues with their other tests in the past as well. They couldn't get the haplogroups correct many times. No major upgrades were offered. Readers should see my blog for other blogger regarding various issues: http://genealem-geneticgenealogy.blogspot.com/2012/08/autosomal-testing-genetic-genealogys.html

Keep up the great blog, Debbie!


Debbie Kennett said...

Thanks Emily. I've come across a few cases of people getting incorrect haplogroup assignments at Ancestry with both Y-DNA and mtDNA tests. It's not surprising when Ancestry don't have any facility to do the proper SNP (marker) testing to confirm the haplogroups. They also don't offer any advanced tests so a lot of people end up having to retest elsewhere if they want to order extra Y-STR markers or get the full mitochondrial genome sequenced. Ancestry no longer seem to be actively marketing their Y-DNA and mtDNA tests.

Tim said...

Interesting post. I just recieved my Ancestry DNA this weekend and was also a little surprised it showed 93% British Isles and 7% unknown, though I know at least 3 of my great-great grandparents came from France or Belgium. However, I also know that 10 of my great-great grandparents were born in Ireland. I was actually shocked I had no Scandinavian ancestry as I know their genes were spread far and wide in medieval times. As others do, I hope the Ancestry results become more specific over time.

Debbie Kennett said...

These analyses are still in the very early stages. A lot of my American matches are getting much higher percentages of "British" than me even though, unlike me, they have ancestors from a variety of different countries. Lots of people are in fact reporting unexpectedly high amounts of Scandinavian. I'm sure the results will eventually become more accurate in time as more reference populations become available.

Joy said...

I have an extensive family tree that goes way back. I thought my family was British, but come to find out, they are mostly from France arriving in England with William the Conqueror in 1066 from Normandy. I have traced some lines back as far as 600 AD and it seems the upper classes took over England as the ruling classes. I just happened to be an American with most lines going back before the Rev. War in America that instead of English are mostly French. Joy