开云体育

Hi Ed,

?

Rather than re-analysing all the ancient DNA samples in detail, I've been relying on published data for broad assignments and Ray's spreadsheet for detailed calls. Published data tends to work off very incomplete trees like ISOGG's 2019/2020 tree. Ray's spreadsheet basically represents a U106-based subset of community efforts to update these calls to Family Tree DNA's haplotree. There's a lot of work gone into this, none of which I've done myself, so I can't comment on individual cases and calls. I can speak generally, however.

?

Ancient DNA is much more degraded than the carefully preserved modern samples we supply to labs. Its sequences break and its individual molecules degrade. This means that what is recovered has a much lower number of reads and a much lower read quality than tests like BigY. If you applied quality thresholds from a modern test (like BigY) to ancient DNA, you would get extremely few calls - perhaps even none - so the quality at which we have to accept a SNP as positive or negative much be decreased. All this means that only a fraction of the SNPs we would call in a BigY test can be recovered from any ancient individual, and those SNPs are more likely to be false positives or negatives.

?

The fraction of SNPs that are called is highly variable as well. Some individuals' DNA might be recovered well, others very poorly. In some cases only broad haplogroups like R1b can be determined; in others a very specific haplogroup can be assigned. Where specific haplogroups are assigned, this is often based on only two or three SNPs in a chain - e.g. M269, L151 and S6189, with the intervening SNPs remaining null. In this case, S6189 might be a false read, or a duplicate of an existing mutation. If the reads for S6189 are nevertheless convincing enough, it might be reasonable to call the sample as being in R-M269>L151>(P312)>S6189.

?

However, if the only SNPs you are attempting to call are on the ISOGG tree, then this terminates at R-Z156, not including the R-S3311 branch at all. If the sample can be called S3311+, S3995+ and A10645+, for instance, then it's much more likely that the sample is R-M269>L151>(U106)>S3311>S3995>A10645, because there is a much more robust set of SNPs down this chain of haplogroups.

?

Something along these lines probably happened for COL239, where the incompleteness of the ISOGG tree led to the original publication over-extending haplogroup calls beyond what could reasonably be justified, and where a more complete tree gives a true picture of where the sample actually lies. This suggests that it's probably worth periodically (~annually) going over all these thousands of ancient DNA samples and seeing whether the haplogroups they are called under are still valid. I'm not sure if FTDNA is actively doing this or not, but we might find that many of these calls that are just labelled "Z381" or "U106" can actually be placed into some of the minor clades that are occasionally discovered even today.

?

Cheers,

?

Iain.

开云体育

开云体育

Three interesting U106+ individuals here:

?

CGG_2_022616 - early medieval (circa 700 AD) Italian, R-L48>L47>Z159

CGG_2_023745 - Bronze Age Spanish (circa 1550 BC, Cuesta del Negro, 37.33N 3.23W), R-BY30097>S18632

CGG_2_023808 - Bronze Age Spanish (circa 1950 BC, Motilla Del Azuer, 39.04N 3.50W), R-Y3444 ~= R-BY30097>FTT8>FGC396

?

The Italian is likely explainable through post-Roman or Roman-era migrations, but the Spanish pair are much earlier. They are the first evidence we see of early R-U106 migrations to south-west Europe, at a surprisingly early date. These clearly are less-successful migrations, as these haplogroups didn't really take off in Spain. It's likely that these particular R-S18632 and R-FGC396 lines died out.

?

In different circumstances, they might have become equivalent to R-P312>DF27. Getting to Spain by about 1950 BC probably meant riding along with the Bell Beaker resurgence that took R-DF27 in particular into Iberia, hence this is the first real evidence we have that some R-U106 actually took part in the Bell Beaker migrations, even if they weren't some of the main drivers behind them.

?

CGG_2_023808 sits towards the end of the predicted range of ages for R-FGC396. This could mean that he is actually pre-FGC396, and only positive for some of the SNPs that make up the current list for haplogroup R-FGC396.

?

Cheers,

?

Iain.

Hi Shane,

?

R-FGC3861 doesn't have any ancient DNA until about 200 AD, so we can't use ancient DNA to understand the first few thousand years of its history. We instead will have to infer it from the other haplogroups. Stay tuned.

?

Cheers,

?

Iain.

开云体育

Dear all,

?

I've updated my . There is now a new section on the spread of R-U106 based on ancient DNA results (mostly written a week ago before life got busy here!).

?

The bottom line is that there seem to have been three different branches of R-U106 expansion: one with R-Z18 going north into the Nordic Bronze Age about 2300 BC; one with R-Z156 spreading west, south, and south-east from about 2000 BC; and one with R-Z301, which is surprisingly absent from the ancient DNA record, but which probably ended up in modern Germany or somewhere nearby. The primary catalyst for R-U106 expansion into much of Europe (especially the British Isles) appears to be the post-Roman Germanic migrations, though it's possible that a smaller number had previously migrated within the Roman Empire. This analysis will not identify if two populations with near-equal amounts of R-U106 mix (as might be possible with later Viking and Norman invasions).

?

In the coming days, I'll be braving storm-force winds on the ferry south from Orkney before hopefully finally managing to get into the Christmas spirit. Replies may be slow and short (partly because I am also typing one-handed after a minor injury), but your thoughts are welcome. Otherwise, consider this an early Christmas present for you all.

?

Best wishes,

?

Iain.

开云体育