Recently my curiosity got the better of me and I've had my results almost two months now. Since then I've been on a steep but fascinating learning curve. I've used my raw data with a number of third-party tools and databases and I'm beginning to see some interesting stuff.

A number of companies now offer various DNA testing services and their databases are expanding enormously, week by week. Many folk are not necessarily interested in genealogy, but rather in the so-called 'ethnicity results' - a pretty inexact science that offers to tell you what percentage you are in relation to various geographical locations. While the results of these should be taken with a large pinch of salt, they can be interesting and used as a rough guide.

I have a number of results from my own data being uploaded to various testing sites, so let's have a look ... Here's Ancestry's ethnicity estimate for me: Great Britain 39%, Ireland/Scotland/Wales 33%, Europe West 16%, Scandinavia 6% and Iberian Peninsula 4%. You can dig down into these results and see that my Great Britain percentage is largely from the West Midlands and Yorkshire Pennines, and the Ireland/Scotland/Wales result is largely Northeast and Central Scotland. The Europe West area includes France, Belgium, Germany and Switzerland, among others, but also takes in a chunk of South-East England.

Ancestry's 'ethnicity estimate'

FamilyTreeDNA gives my overall origins at 100% European, breaking it up into 81% British Isles and 19% West and Central Europe - not far off the Ancestry results. DNALand assigns me 100% West Eurasian, of which 91% is Northwest European, 8% is Southwestern European, and 1.1% is 'ambiguous'.

One of the more interesting is LivingDNA who have a very good UK reference set to draw from. They put me at 100% Great Britain and Ireland and break that down as 59.8% Central England, 22.6% Aberdeenshire (this is a wide area, not just the county), 8.4% Southeast England, and then tiny amounts (<3%) from other UK areas.

One thing I was curious about before I decided to do the DNA test was whether any Asian, particularly Northwest Indian, would show up. The reason for this is that I have a rather strong branch of Romani Gypsies in my ancestry (whose ethnic origin goes back to this part of the world about 1500 years ago). However, while waiting for the results I did some reading and realised anything here probably wouldn't show up - the 22 pairs of autosomes that are analysed will have gone through so much recombination that not much can be detected from more than a few generations back.

So how accurate are the above ethnicity results? Luckily I have a good amount of research behind me so I decided to do my own 'ethnicity test' based on the genealogical record, rather than the genetic one.

To discern my genetic make-up I went back five generations to my sixty-four gggg-grandparents and looked at their birth counties. To start with the big picture, I'm 51.56% English and 48.44% Scottish. This reflects the fact that while my mother's ancestry is all English (back to the 1700s), my father's side is all Scottish - with the exception of one Englishman who got very briefly involved back in 1826.

Taking a regional view, that Scots 48.44% is all Mid Scotland - originating in the Tay and Forth areas of Perthshire, Angus and Fife. The English side contains 35.94% from the Midlands, 12.5% from the North West, and 1.56% each from East Anglia and Mid-North. You can see the breakdown at county level in the pie chart below, with Fife and Staffordshire taking the biggest slices.

My own 'ethnicity estimate' using genealogy rather than genetics

So the commercial ethnicity estimates are not quite correct at a detailed level, but they're not far off in broad strokes. Ancestry gives me roughly half each on Scotland and England, and the European mainland parts have to be taken as noise (my one French-born ancestor, around the time of Waterloo, had Scottish parents). The same goes for the others, though LivingDNA underestimated my Scots make-up by a fair chunk. One thing's for certain - I am unexotically very British.

Using the autosomal part of the DNA test for genealogy has already proved fruitful. With millions of people in the databases, your results can be compared and close and distant connections flagged up. With this I've been able to confirm a lot of my genealogical research genetically - which is a relief, especially for some of the more complicated relationships I've had to untangle (Gypsy ancestors, I'm looking at you!). I've even been using DNAPainter to start recording which bits of which chromosomes came from which ancestors (eg. a 24cM chunk of my maternal chromosome-14 from the Pritchards).

One match, rather astonishingly, suggested a DNA link with a known 8xg-grandparent, going back about 350 years - my match and I would be 9th cousins. I thought this would be well beyond the reach of autosomal DNA - and it might be, it's possible we could have a closer link on a separate, unrecorded branch of the family. But I read up on it, and it is also a fairly reasonable possibility.

While any chunks of DNA passed down from that long ago would be vanishingly small, it is also true that ancestors that far back will - if their lines survived into modern times - have thousands and thousands of descendants. So the chances of any one person having recognisable DNA from that long ago are tiny, but the huge number of possible carriers makes it likely it has survived intact somewhere (see Genetic Genealogy and the Single Segment).

It's still early days for my analysis of all these matches, and while I have yet to break down any of my personal research 'brick walls', a number of tantalising clues have been thrown up in a few places (the Worrilows from the little village of Haughton in the 1600s are definitely trying to get my attention!).

DNA seems to suggest a connection with the Worrilow family in Haughton - where my Lees come from. This 1625 parish record of a marriage between John Leighs and Jane Warrilow may be the connection.

Apart from our 22 autosomes, we also have either an X and Y chromosome (if we're male) or two X chromosomes (if we're female). And we have mitochondrial DNA - this comes only from our mother, while the Y comes only from our father. Analysis of these can tell us about the paternal edge of our family tree (which usually also includes our surname back into history), and the maternal edge of our ancestry - our mother's mother's mother's mother, etc.

Y-DNA can give you your male line haplogroup - for me it's R-L21 with a subclade of R-S3058. The R haplogroup is extremely common and has its origins in Central Asia, possibly around 27,000 years ago. About 18,000 years ago haplogroup R1b formed, mutating and moving into Europe. Another 'ancestor' of my haplogroup is R-M269, the most common Y-DNA lineage among European males. R-L21 is several steps below this, a signature of a Bronze Age people, the 'Atlantic Celts', and common today in the populations of Ireland, Scotland and Wales. Coming further down the line, my subclade of R-S3508 (also romantically known as R1b1a2a1a2c1g4) is approximately 3,800 years old.

I was actually able to analyse a little bit further and get an estimated sub-subclade of R-S190, a haplogroup whose members would share a common ancestor about 1,850 years ago and is a marker for a group known as the 'Little Scottish Cluster'. My earliest known Y-DNA ancestor is one James Ewan/Ewing, born around 1765, probably in Perthshire, so it's all pointing in generally the right direction. (Update: new test results from 23andMe have confirmed R-S190 as my haplogroup; update II: further testing has put me in the R-Z17999 subclade, well within the Little Scottish Cluster.)

Migration of my mtDNA haplogroups (female) and Y-DNA haplogroups (male)

My mitochondrial DNA (mtDNA) comes to me from a woman named Susanna - I don't yet know her surname (update: it's Rhodes), but she was probably born in Staffordshire circa 1780. I can say something about her maternal ancestors though, as they gave me my mtDNA haplogroup: J1b1 (subclade J1b1a1b).

J1b1 is most common in Britain and Ireland but is still quite a rare haplogroup, being found among only 1.2% of the English population. If you've ever read Brian Sykes' 'The Seven Daughters of Eve', the imagined matriarch of this group is known as Jasmine. She probably lived somewhere along the Euphrates in what is now Syria about 45,000 years ago, and is thought to be among the early adopters of agriculture. The subclade J1b1a1b is probably just over 4,000 years old.

Learning about DNA and analysing my genetic fingerprint has been fascinating in these early weeks of having the results, and I'm sure there's a lot more to come from it yet.