Saturday, February 10, 2018

DNA Testing and Ancestral Research

My folk comrade, Runebinder has recently brought to my attention concerns regarding the accuracy of DNA testing when applied to ancestral research and he is right to focus concern on this. DNA testing is still in its infancy and so must be treated with caution. When ancestral researchers such as myself use this tool it must be with care and should be used alongside and not instead of old fashioned paper (or Internet!) research.

I have spent the last 18 months getting to grips with researching my paternal ancestry. My maternal ancestry is currently a closed book, not being able to progress beyond my German grandparents. The reason for this is that only a tiny fraction of German records are digitalised and those which have not been are difficult to access without physically visiting town halls and churches as no central records exist. Added to this is no doubt a loss of records due to the allied bombing in WWII. The German authorities also place restrictions in the way of obtaining documents: one must PROVE  a biological connection first! The situation in the English speaking world is thankfully more enlightened and records are much easier to access and most have now been digitalised.

Both types of research have their limitations. As I have already pointed out DNA testing still has its many imperfections and inadequacies. There are three types of ancestral testing which are now available: Y chromosome (Y-DNA), mitochondrial DNA (mtDNA) and autosomal DNA (auDNA or atDNA) testing. The least accurate of these three types of testing is auDNA. This is relatively new when compared with the other types of testing and most companies struggle to distinguish between English (Anglo-Saxon) and continental Germanic DNA such as German and Dutch although they can distinguish Scandinavian DNA separately. Often they tend to lump together German/Dutch with French and term it 'North-West European'. The reason for this is the relative lack of data that they currently have available that they can draw upon. As time passes this situation will improve.

Regarding auDNA in theory an individual would expect to inherit 1/4 of his auDNA from each grandparent and 1/8 from each great grandparent. However this is a simplistic way of viewing the matter because in reality the transmission of auDNA is a complete lottery. Whilst one inherits 50% of one's DNA from each parent, the DNA from earlier generations such as grandparents and great grandparents starts to become 'lost'. This is why most companies only provide an analysis back 5 or 10 generations (3 x great or 8 x great grandparents). Y-DNA (paternal) and mtDNA (maternal) is much more reliable but this measures 'deep' ancestry from thousands of years ago. One can however spend more money and get more detailed tests which show mutations going back over a shorter period. My Y-DNA and mtDNA has been tested by two companies and both produce the same results although one set of tests which has been more recently conducted gives the subclades for the paternal and maternal haplotypes. Y-DNA and mtDNA are very useful for determining the racial origins of populations but not necessarily that of individuals-unlike auDNA.

As I have stated the transmission of auDNA which is generally the ancestral DNA people think of is (rightly) the subject of scrutiny and criticism and it is a lottery and as I have said it is also in its infancy. For this reason it may be helpful to try more than one company to test your auDNA and compare the differences (there are bound to be some) or you can download a tool and analyse your DNA yourself using for instance GEDMatch as the auDNA which is extracted by the different companies is exactly the same. What differs is their interpretation of it!

On the subject of siblings having different auDNA test results: this is normal and thus to be expected. As each parent transmits just 50% of their DNA to each child again it is a lottery as to which 50% is transmitted. It is possible in theory to have two siblings which have completely different DNA although this is statistically extremely unlikely! Nevertheless there will be significant differences. This is why siblings can either closely resemble each other physically and have similar psychologies and  behaviours or they may appear as polar opposites. Usually it is somewhere in between. Even in the case of twins and triplets they do not have identical DNA but similar DNA. Thus different auDNA test results will occur. That said auDNA must be treated with caution as it is not necessarily an accurate measurer of ethnicity. This is why I recommend analysing the DNA yourself by uploading your DNA and using the Internet analysis tools or using different companies who will analyse the same DNA but they will use their own and different algorithms. As time goes on their algorithms will converge (with the collection of more DNA data) and will produce similar results. What would be helpful is for full siblings to have their DNA tested by the same company and this will provide more information and uncover DNA that has not been transmitted to you. Half siblings and cousins could also undergo a similar evaluation. This is far better than analysing just one member of a family.

On the issue of auDNA and ethnicity, one could of course have a French grandparent and yet not inherit any DNA from that ancestor but one is still 1/4 French. There are thousands of people alive today (my partner and daughter being two of them!) who may be descendants of Charlamagne and yet have not inherited any of his DNA because it has become lost over time.

The two companies that I have thus far used are Oxford Ancestors (many years ago) for Y-DNA and mtDNA and more recently Living DNA for Y-DNA, mtDNA and auDNA. The latter company provided sub clades for the paternal and maternal DNA although I was able to work out my own paternal sub clade from the data already provided by Oxford Ancestors. The auDNA was rather mixed in terms of its accuracy. It accurately demonstrated my Lancashire and Welsh ancestry but was absolutely useless in finding any German ancestry despite having a German mother and two German grandparents! English and German DNA is just too similar to tell apart (at the moment). Thus the only types of testing that you can rely upon with 100% certainty are Y-DNA and mtDNA testing. For those with ancestry from the British Isles (Britain and Ireland) Living DNA is the best company for providing detailed regional and county wide analysis.

I have spoken about the inadequacies of DNA testing but conventional research also has its limitations. With much luck I have researched my paternal grandmother's ancestry using conventional records such as birth, baptismal, marriage, death, burial, census records and published wills and historical accounts back more than 500 years and beyond this if the assumptions regarding Alis Aughton's parentage are correct. However records differ according to where one's ancestors in the United Kingdom came from. If they are from North Meols in Lancashire as 1/4 of mine are then one is fortunate to have access to excellent transcribed digital records. The three sites that I use are familysearch.org, findmypast.uk and Lancashire OnLine Parish Clerk Project. The use of the familysearch.org and the Lancashire sites is entirely free. One can also build a family tree on Familysearch but be aware: these trees are public and any Tom, Dick or Harry can amend your records and many of these 'researchers' should not be allowed anywhere near a genealogical website! The findmypast.uk site is fee based but the maintenance of one's tree on there is free-and private! Familysearch rely on records from findmypast but they are not always accurately transcribed and the same goes for the records on findmypast but the latter site does allow you to view the original record so that you can check for transcription errors and find additional information not provided on the transcriptions. Surnames are often misspelled and up until recent times standard spellings did not exist and a surname could be spelled 4 or 5 different ways. The researcher must be aware of this possibility and not give up!