A longfather of old

I have a question.

My 23 and Me results said that my ancestry composition is:

78.4% British and Irish

19.9% French and German

1.7% Northwest European

but my Living DNA results are different, based on the download of my raw genetic data from 23 and Me:

97.5% Britain and Ireland

1.4% Northwest Germanic

1.7% Northwest European

I don't know which one is correct (and why they would differ)?

I think and others here are far more knowlegable than I, will be able to confirm or deny this, but it can depend on how large the samples used to arrive at the estimates are, obviously the larger the sample the more acurate the results are likely to be. DNA research is moving at such a pace that many studies aren't published before they're superceded by newer research. Some countries are better represented in genealogical research than others, America seems to be very well researched as so many Americans know their recent ancestry is from elsewhere usually of European decent. Some countries, I believe France is one of them outright bans this sort of genealogical research in the name of Fench unity and also against comercialisation of genetic material, a DNA test is usually only carried out in France if court ordered, like for a paternity test.

According to My Heritage, I'm 87% Southern English, with 7% Iberian and the rest made up of random Northwestern European. My Mum, has no Iberian DNA and I can't find anyone on my dads side of the family going back nearly 200 years who seems to be a recent imigrant. It's confusing and not an easy search when one has a fairly common name and seems to have spent a large part of thier time in the East End of London, thats if I got that connection right and I'm not sure I did. Unless you can go and physically search archives, church records and make any sense of military ones which seem designed to obfuscate. It can be pretty difficult and not as many of them are digitised as you're led to believe, some of the digitised records are of poor quality and difficult to read, others may well be accurate, but what I wondered looking at mine, was how much of what my peasant ancestors mumbled at an educated registrar of priest was recorded properly? Some of it I can make out where the mistakes are as I know what the local accent was and the stresses on vowels were a bit different to "educated" English. For example I guessed where Teresa May grew up because I recognised some of her pronunciation, such as Mondee instead of Monday, that particular way of pronouncing "ay" as "ee" is quite common in parts of Oxfordshire and Berkshire.

I think and others here are far more knowlegable than I, will be able to confirm or deny this, but it can depend on how large the samples used to arrive at the estimates are, obviously the larger the sample the more acurate the results are likely to be. DNA research is moving at such a pace that many studies aren't published before they're superceded by newer research. Some countries are better represented in genealogical research than others, America seems to be very well researched as so many Americans know their recent ancestry is from elsewhere usually of European decent. Some countries, I believe France is one of them outright bans this sort of genealogical research in the name of Fench unity and also against comercialisation of genetic material, a DNA test is usually only carried out in France if court ordered, like for a paternity test.

According to My Heritage, I'm 87% Southern English, with 7% Iberian and the rest made up of random Northwestern European. My Mum, has no Iberian DNA and I can't find anyone on my dads side of the family going back nearly 200 years who seems to be a recent imigrant. It's confusing and not an easy search when one has a fairly common name and seems to have spent a large part of thier time in the East End of London, thats if I got that connection right and I'm not sure I did. Unless you can go and physically search archives, church records and make any sense of military ones which seem designed to obfuscate. It can be pretty difficult and not as many of them are digitised as you're led to believe, some of the digitised records are of poor quality and difficult to read, others may well be accurate, but what I wondered looking at mine, was how much of what my peasant ancestors mumbled at an educated registrar of priest was recorded properly? Some of it I can make out where the mistakes are as I know what the local accent was and the stresses on vowels were a bit different to "educated" English. For example I guessed where Teresa May grew up because I recognised some of her pronunciation, such as Mondee instead of Monday, that particular way of pronouncing "ay" as "ee" is quite common in parts of Oxfordshire and Berkshire.

LivingDNA do something similar to match people to ‘Classical Populations’, Greek, Roman and Egyptian. I am distantly related to populations from all three. It isn’t that surprising as people were migrating in numbers. My Ancient Greek match is from NE Spain, the Roman from NE England and the Egyptian hasn’t been narrowed down other than being from Egypt.

Thanks, that’s quite fascinating, it’s a whole topic I’ve never really thought about. Tempted to look into my ancestry 

The Roman male was one of 127 individuals from 29 archaeological sites in and around Rome. They had their genome sequenced by researchers in 2019. The analysis by 23andMe shows that we share a segment of DNA making us very distantly related. I have been matched to 14 historical individuals by 23andMe, including an individual who was probably a victim of the St Brice’s Day massacre in Dorset c.1002CE, an Eastern European prince and a late Iron Age individual. 

Curious point around the Roman male, how did you find out such a specific connection? 

Thank you Martin. My father and his parents and his grandparents were from Lancashire, and the results from LivingDNA have probably fairly accurately reflected this with Lancashire 24.9%, together with Yorkshire and Cumbria making up a total of 50.1% Northern English. All the males in my father’s line were from Lancashire, at least as far back as 1620, and all the male line marriages have been to women living in Lancashire. I am beginning to wonder if inbreeding was prevalent until more recent history, although I know marriages between 1st cousins happened more back then.

LivingDNA has assigned the remaining DNA to Ireland and Southern England, with only 4.4% Northern Ireland and South West Scotland, while 4.5% Scandinavia. My maternal grandmother and grandfather were Irish, but now it looks like there was more English than Irish ancestry on that side of the family. 

I would like to have some Italian ancestry, but unsurprisingly, nothing has shown up except the late Roman male who suffered a violent death and was buried in Rome. 

The companies tend to update their databases, so looking at your results every six months or so may resolve things to some extent.

I only have experience of Living DNA and 23 and Me. A UK-based company would probably be better at distinguishing between British and near-Continental ancestry, than one based elsewhere. As for me Living DNA gave me 100% ancestry from the British Isles, 23 and Me gave me 99.8% of the same ancestry.

For ancestral geographic origins, companies are looking at the last 500 years, so are looking at relatively long stretches of DNA for comparisons. The further back in time an origin is the smaller are the lengths of DNA that will match. This is because random chromosomal mixing (chiasma) effectively chops up and diminishes the individual contributions of ancestors over time. This is why I have no Scandinavian ancestry, but have many matches to ancient DNA from Viking Age Swedes, Danes and Icelanders. My Norse ancestry is probably from 1,000+ years ago, so to find it quite small stretches of DNA identity need to be investigated. My Neandertal DNA (less than 2%) is around 75,000 years old so it mostly consists of very, very small lengths of DNA, very well scattered across my chromosomes.

Martin, apologies for using you as a personal information service, do you know how good LivingDNA is at picking up French ancestry? I was led to believe by my Grandparents that one side of the family was French. Would Ancestry UK be better for that?

That's good to know.

Many thank yous. 

Living DNA are far better than 23 and Me at picking out fine structure in the populations of the British Isles. They came within 3% of my predicted Irish ancestry, from what I know from my family tree. Because of the random nature of chromosomal chiasma, predicting percentages from the origins of gt. grandparents and beyond is inherently inaccurate.

Because  23 and Me is weighted towards sorting the often diverse ancestry of Americans and Living DNA is UK-based, I would tend to believe Living DNA is probably more accurate in discriminating between UK ancestry and ancestry from elsewhere in Northwest Europe. If you had some more exotic ancestry, 23 and Me would be more reliable in picking this up.

Another variable is that areas of the British Isles have populations with greater similarity to the near Continent than others. In one study, the people of the area around York had about 60% Northwest German and Danish like ancestry when the average for England as a whole was 38%.

Thank you. 

I've lost some faith in the process as I either have a fifth French/German genes or I don't.

The major variation is on the size and composition of the databases that companies use. They tend to create reference databases from different regions by using the genomes of people who have all grandparents from the same area. This is not infallible as earlier generations in any family might be more geographically diverse. Another large variable is where the companies set their boundaries between populations.. There are also variations in what areas of the genome are investigated - you do not get full genome coverage. Areas of the genome called hypervariable regions are looked at and different companies look at different regions of DNA.

I'm guessing that Living DNA has different databases to 23 and Me, so maybe it takes your raw data and analyses it differently? It was Martin who said about Living DNA.

TheCatWoman said:

but it can depend on how large the samples used to arrive at the estimates are, obviously the larger the sample the more acurate the results are likely to be.

In this case it was the same sample as I uploaded the raw genetic data from 23 and Me to Living DNA which someone on here (Martin?) had previously done which gave me the idea.

Thanks though and what you write is interesting.