PDA

View Full Version : What does the P312/U106 clade ratio tell us?



Cofgene
07-01-2017, 09:53 AM
Iain McDonald pushed out the following plot several days ago to several of us to consider. I'm putting it up here without his comments to see where the interpretation conversation goes. When the P312 ages from his analysis become available this type of comparison between any two haplogroups of similar ages can be set up within Excel. The ratio at a specific time is acquired by binning the number of clades present from the output of the age analysis. Data slice and dice time.

17327

What do you see in the plot? :P

RobertCasey
07-01-2017, 12:38 PM
First - since this plot is obviously based on some mathematical model, the smoothness of the curves suggests that either the math is correct (validated by the well behaved characteristics of the curve) or it could be the assumptions behind the model producing nice looking results.

Second - The actual chart shows that P312 was much more dominant at one time and then was slowly replaced by U106 for many centuries followed by a modest increase in P312 during the last part of the chart. I assume that this must be based on TMRCA dates associated with older haplogroups and the numbers of current testers stuck in these older haplogroups. The large number of YSNP branch equivalents in the 1,500 to 2,500 year range shows that L21 really struggled during this era. So this plot is probably biased on YSNP dating in some form. Since, in general, the population continued to grow, if L21 was challenged to grow during these genetic bottlenecks, other haplogroups such as U106 may have been more prolific during L21's struggle to dominate.

You would really have to publish the source material behind such an analysis to get a proper review. Unfortunately, for the genetic genealogy community, few of us have the skills to challenge any such paper or have the time/interest to challenge such a chart. My interests are closer to the genealogical time frame but I am always interested in such charts as they make you think outside of the box when you see such charts. I prefer to see more emphasis on more relevant topics such as YSNP prediction using YSTR signatures, charting of genetic genealogy information once YSNP testing becomes significantly available, TMRCA estimates in the last 1,500 to 2,500 years vs. very old estimates and analysis of NGS 400 YSTR results that are available but are not being used (we need mutation rates and methodologies to replace FTDNA's current hybrid approach for mutation rates).

rms2
07-01-2017, 01:09 PM
Right offhand, it looks like die Völkerwanderung period of the middle of the 1st millennium A.D. was much better for U106 than it was for P312. That could be a reflection of the advances made during that time by the Germanic peoples at the expense of Celts and Italics.

Not everyone will like that idea, since it tends to support the notion that U106 is primarily associated with Germanics and P312 with Celts and Italics, so perhaps those who disagree can come up with some other interpretation.

MitchellSince1893
07-01-2017, 04:31 PM
My first observation, which is on par with "No Duh" is it shows the explosive growth of P312 during the Bell Beaker Period.

The following Tumulus culture was not as prolific as BB, combined with the U106 growth in the Nordic Bronze age, starts the downward slope of P312 to U106, which reaches it's low point when the Germanic tribes are overrunning the Roman empire and the arrival of the Anglo-Saxons in Britain.

As to why P312 starts to increase again after 500 AD, maybe it was a combination of Slavic expansion into Germanic areas, possibly the Reconquest of Iberia (just guessing on that one), and Irish population growth towards the end, as well as P312 growth in the new world (e.g. DF27 in Latin America)

Nothing else huge really stands out to me as to why P312 would increase compared to U106 in places like Britain, Germany, France, Low Countries during this period.

However, just some wild a$$ guess (WAG) is: As the upper classes tended to have a population growth advantage compared to the lower classes, maybe there was a period of increased growth of P312 dominated upper class families compared to their U106 counter parts. Also maybe U106 dominated nobility were disproportionately on the losing end of upper class struggles for dominance in their respective regions. e.g. the Norman nobility was more P312 prevalent than the previous Anglo-Saxon nobility. I'm grabbing at straws...just throwing it out there.

Nibelung
07-01-2017, 04:44 PM
Renaissance?

The explosion in Italy may have sent cracks everywhere in the Late Medieval Germanic European superstructure, although the English Renaissance beginning with the Tudors may simply be coincidence.

Cofgene
07-02-2017, 09:31 AM
First - since this plot is obviously based on some mathematical model, the smoothness of the curves suggests that either the math is correct (validated by the well behaved characteristics of the curve) or it could be the assumptions behind the model producing nice looking results.


Not a model. These are actual counts of the number of currently identified haplogroups present at that point and ratioed.

Cofgene
07-02-2017, 09:35 AM
...
As to why P312 starts to increase again after 500 AD, maybe it was a combination of Slavic expansion into Germanic areas, possibly the Reconquest of Iberia (just guessing on that one), and Irish population growth towards the end, as well as P312 growth in the new world (e.g. DF27 in Latin America)


Partly that but it could also contain sampling bias for recent British Isle clades due to under representation of continental results.

RobertCasey
07-02-2017, 01:35 PM
Not a model. These are actual counts of the number of currently identified haplogroups present at that point and ratioed.

So these are based on TMRCA (YSTR and YSNP models ?). However, there is a major assumption that all haplogroups are the same in being prolific. U106 is a much more bushy haplotree with many more early branches - but smaller quantities associated with each haplogroups. L21 has a lot more starbursts of expansion and many of these haplogroups are much larger than U106 haplogroups. Have you thought about introducing counts for haplogroups to see actual number of testers associated with each haplogroup (and its descendants) ?

kevinduffy
07-02-2017, 05:19 PM
Partly that but it could also contain sampling bias for recent British Isle clades due to under representation of continental results.

Is that what Iain McDonald thinks?

Dewsloth
07-02-2017, 05:52 PM
17327

What do you see in the plot? :P

A drunk N.

Cofgene
07-03-2017, 12:37 AM
Is that what Iain McDonald thinks?

No. Just my thought. Conceptually testing bias should get more pronounced closer to current time due to what we know about composition of regions such as the British Isle. If that same composition was present across continental Europe one wouldn't need to consider the bias present in the counts of more recent haplogroups.

rms2
07-03-2017, 01:58 PM
Partly that but it could also contain sampling bias for recent British Isle clades due to under representation of continental results.

Isn't U106 pretty well represented in the most heavily populated parts of the British Isles?

P312, especially L21, is dominant in the Celtic Fringe countries, but their populations are much smaller than that of England, especially SE England, where U106 has its highest frequencies outside the Netherlands.

Under-representation of the Continent is also likely to favor U106 rather than P312.