Y-DNA

I recently had a Y-DNA test done by 23andme, which found my Y-DNA haplogroup to be R1b1b2a1a2f*. This needs a lot of interpretation to understand, and this is not very forthcoming from the 23andme website.

The Y chromosome is inherited, father to son, in an unbroken line stretching back more than 100 million years ago. Normally the son inherits a perfect copy of the father's Y chromosome, but occasionally a mutation occurs which then differentiates that son's DNA and that of his sons, from his father's, brothers', and other male relatives. One common form of mutation is a single nucleotide polymorphism or SNP, where a single base pair in the genome is altered. It is these which are principally of interest in Y-DNA testing, at least of the sort done by 23andme.

The Atlantic Celtic (R-L21) haplogroup

When a mutation occurred many centuries ago, and there are many male-line descendants, this is said to define a haplogroup or clade. For example, the R haplogroup is defined by the M207 SNP, believed to have arisen in a man living maybe 27 thousand years ago in southern or central Asia. Longer group names refer to sequences of mutations since that point, so R1b refers to three mutations: M207, M173 and M343.

The problem with this systematic notation is that as we gain new understanding of the human Y-DNA tree, and new branches are discovered, the haplogroup names change. The clade I belong to is defined by the L21 SNP, and was initially called R1b1b2a1b6 in the 2008 ISOGG tree, renamed R1b1b2a1a2f in the 2009 tree, R1b1a2a1a1b4 in 2011, R1b1a2a1a1b3 in 2012, and R1b1a2a1a2c in 2013, which is how it remains. It therefore seems that 23andme is using a 2009 or 2010 version of the tree, and from the haplotype tree on their website, it seems to be the former.

Because of this instability in names, it's more common to refer to groups by the parent haplogroup together with the most recent SNP name. My haplogroup would normally be described as R1b-L21 or just R-L21. Most estimates I've seen date the L21 mutation to the Bronze Age. This group is sometimes called the "Atlantic Celtic" branch due to its prevalence in Brittany, Cornwall, Wales, Scotland, and particularly Ireland where the majority of the population belong to it. Even in Hampshire, where my recent paternal ancestry is, it comprises 30–40% of the native population. In the Netherlands, Germany and Denmark, the incidence of L21 is less than 5%, and in Norway it is less than 15%, all of which suggests that my male-line ancestors are much more likely to be of Celtic origin than Anglo-Saxon or Viking.

See this map on Eupedia showing the distribution of R-L21.

English surnames generally arose a couple of centuries after the Norman invasion. My surname is Smith which suggests that the ancestor who first took that was in a English speaking at the time. Roughly 30 generations have elapsed between then and now, each of which is an opportunity for illegitimacy (specifically when a son inherits his mother's surname) or a so-called non-paternal event (when the assumed father was not the actual father). Studies that have looked at DNA for a rare surname, assumed to have a single origin, tend to find a 1–3% likelihood per generation of the son's surname not coming from his genetic father. Over 30 generations that works out at roughly even odds that the man from whom I inherited my surname was actually my patrilineal ancestor. Nevertheless, it is the best clue I have as to how I ended up with a L21 haplogroup. It suggests that I am not descended in the male line from a recent migrant from one of the Celtic nations, nor from an earlier Saxon or Viking invader, but rather from one the Celtic Britons who inhabited England before their arrival.

Within L21

On the face of it, this is all my 23andme result tells me. But is it? It assigned me a haplogroup of R1b1b2a1a2f*, and the * has meaning too. It denotes a paragroup, meaning I'm not in any of the subclades they tested. Unfortunately, it's not trivial to work out what they have tested. The 2009 ISOGG tree defined three subclades of R1b1b2a1a2f (R-L21):

• R1b1b2a1a2f1, defined by M37
• R1b1b2a1a2f2, defined by M222
• R1b1b2a1a2f3, defined by P66

All three are listed on 23andme, however 23andme do not list a defining SNP for R1b1b2a1a2f3, and I have been unable to locate anyone who has been assigned that haplogroup by 23andme. This suggests they may only look at M37 and M222, and assign to the paragroup anyone who tests negative to these SNPs.

Fortunately, it is possible (if very tedious) to discover whether I was tested for a given SNP, and if so whether I was positive or negative. First of all, I need to search SNPedia to convert the SNP name to a RefSNP id. This tells me that M222 = rs20321 and mutated G→A. If I choose to browse the raw data of my 23andme test, I can search for rs72625304 and find that I have G, therefore I tested negative for the M222 SNP. (This is not surprising as the M222 haplogroup is predominately an Irish one, sometimes suggested to correspond to the descendants of Niall Noígíallach, a semi-legendary High King of Ireland in about the 4th century, and founder of the Uí Néill family.)

Known male-line relatives

My earliest known patrilineal ancestor is my 4× great grandfather, John Smith (ancestor #64). Below are the male-line descents I traced from him so far. Unfortunately, they only include three living individuals: me, my father, and his cousin. It would be lovely to find a more distant male-line descendant who had had their DNA tested to confirm this.

---

John Smith. b c1779. m Frances Harding, 1814, Southampton, Hants. d 1834, Fawley, Hants.

• John Smith. b 1816, Fawley, Hants. d 1882, Fawley, Hants. Unmarried.
• Henry Smith. b 1818, Fawley, Hants. m Sarah Willis, 1835, Fawley, Hants. d 1902, Fawley, Hants.
  • George Smith. b 1835, Fawley, Hants. Vanishes from the record.
  • William Smith. b 1836, Fawley, Hants. m Mary Callen, 1873, Fawley, Hants.
    • Frederick Smith. b 1874, Fawley, Hants. m Florence Farman, 1907, Smallburgh, Norfolk.
      • William Robert Frederick Smith. b 1908, Fawley, Hants. Unable to trace further.
  • Thomas William Smith. b 1844, Fawley, Hants. m Sarah Eliza Croutear, 1868, South Baddesley, Hants. d 1914, Fawley, Hants.
    • Edward Smith. b 1873, Exbury, Hants. d 1959. Unmarried.
    • George Smith. b 1875, Exbury, Hants. m Edith Vane, 1899, North Stoneham, Hants. d 1952, Fawley, Hants.
      • George Archibald Smith. b 1901, Canford, Dorset. m Kathleen Helena Richardson, 1927, Leeds, Yorks. d 1973, Bowral, NSW, Australia. No children.
      • Lionel Vane Smith. b 1903, Kinson, Dorset. m Gwendoline Mary Moody, 1934, West End, Hants. d 1980, Fawley, Hants. One son and one grandson.
      • Albert Edward Smith. b 1905, Bealieu, Hants. m Kate Alice Hogben, 1926, East Ashford, Kent. d 1983, Fawley, Hants. One son.
    • Charles Smith. b 1877, Exbury, Hants. m Edith Mary Tickner, 1907, Weybridge, Surrey. Seems only to have one (or two?) daughters, but no sons.
    • John Smith. b 1877, Fawley, Hant. m Julia Mabel Trevett, 1909, Piddletown, Dorset. d 1966, Weymouth, Dorset. Only one daughter.
  • Alfred Henry Smith. b 1847, Fawley, Hants. m (1) unknown. m (2) Susannah Smith formerly Brett, 1883, Salford, Lancs. d 1892, Salford, Lancs.
    • William J Smith. b 1880, Salford, Lancs. Possibly a step-son of Alfred, by Susannah's first husband whose surname was also Smith. Too many similarly named people to trace reliably.
  • Edward Smith. b 1848, Fawley, Hants. Vanishes from the record.
  • James Smith. b 1852, Fawley, Hants. A gamekeeper in Kinson, Dorset in 1871, then vanishes from the record.
  • Frank Smith. b 1860, Fawley, Hants. m Mary Rose Whitehorn, 1879, Fawley, Hants.
    • Thomas Smith. b 1878, Southampton, Hants. Too many similarly named people to trace reliably.
    • Alfred Robert Smith. b 1884, Fawley, Hants. m Edith, 1910.
      • N Smith. b 1911, Donhead, Wilts. (Presumably either Wilfred Alexander C Smith or Oscar B Smith.) Have been unable to trace any further.
• George Smith. b 1821, Fawley, Hants. m Sarah, c1845. Living in Northwood, Isle of Wight in 1851, then vanishes from the record.
  • Henry Smith. b 1848, Lymington, Hants. Living in Northwood, Isle of Wight in 1851, then vanishes from the record.