PDA

View Full Version : How accurate is this M458 map?



Brent.B
07-22-2015, 08:06 PM
This map has been around for a while now... but I was curious to see what people here thought of it? I'm surprised by such high frequencies in Pomerania.

5292

Volat
07-22-2015, 09:05 PM
Such maps are seldom accurate. Underhill et al (2014) tested for different SNPs under R1a1 in over 100 populations . See table 4 : http://www.nature.com/ejhg/journal/v23/n1/extref/ejhg201450x5.xls

Tomenable
08-05-2015, 09:14 AM
^ Eupedia has better maps than Underhill, at least when it comes to Z93 (e.g. Underhill's map shows low frequency in ethnic Pashtun areas):

https://en.wikipedia.org/wiki/Y-DNA_haplogroups_in_South_Asian_populations

IIRC, Underhill's sample confirmed high frequency of R1a in Pashtuns. So he probably didn't know where Pashtuns live when making the map...

Generalissimo
08-05-2015, 09:33 AM
Such maps are seldom accurate. Underhill et al (2014) tested for different SNPs under R1a1 in over 100 populations . See table 4 : http://www.nature.com/ejhg/journal/v23/n1/extref/ejhg201450x5.xls

Underhill 2009 and Underhill 2014 were both very entertaining papers, but not very informative.

Coldmountains
08-05-2015, 09:39 AM
^ Eupedia has better maps than Underhill, at least when it comes to Z93 (e.g. Underhill's map shows low frequency in ethnic Pashtun areas):

https://en.wikipedia.org/wiki/Y-DNA_haplogroups_in_South_Asian_populations

IIRC, Underhill's sample confirmed high frequency of R1a in Pashtuns. So he probably didn't know where Pashtuns live when making the map...

Yes his maps for Z93 look ridiculous. Most Pashtun populations have 50+% R1a-Z93 and many Pashtun populations even 70% or even slightly more. R1a is lower in Tajikistan (Duschanbe just 19% but on average close to 40%) except of some local peaks and in Kirghizstan not all regions are so rich in Z93. Z93 is lower among Afghan Tajiks than among Afghan Pashtuns in all studies I have seen so far and most of the Tajiks samples so far where from the north. Most studies show R1a percentages around 35% for Afghan Tajiks (but they have more R1b than Pashtuns). But Tajiks in the east seem to have R1a levels close to Pashtuns while Tajiks in the west are relatively low in it and Tajiks in the north show often very different percentages (sometimes 50% sometimes just 20%) To be honest most studies about R1a are crap and basically all maps I have seen so far are extremely inaccurate because they are based on too few samples. But yes the eupedia map is the best I have seen so far for Z93. But maps for European R1a subclades look better and are relatively accurate

Volat
08-05-2015, 11:22 AM
Underhill 2009 and Underhill 2014 were both very entertaining papers, but not very informative.

There are over 30 authors in 2014 paper. So, who is more informative on R1a1 at the moment?

Generalissimo
08-05-2015, 11:35 AM
There are over 30 authors in 2014 paper. So, who is more informative on R1a1 at the moment?

There's not a single decent academic paper on R1a, so let's stop pretending like there is, shall we?

Volat
08-05-2015, 11:40 AM
There's not a single decent academic paper on R1a, so let's stop pretending like there is, shall we?

The OP of the thread is interested in the distribution of M458. Are you suggesting these 32 authors could not have tested for M458 counting the frequency of the mutation in their samples?

Generalissimo
08-05-2015, 11:57 AM
The OP of the thread is interested in the distribution of M458. Are you suggesting these 32 authors could not have tested for M458 counting the frequency of the mutation in their samples?

Yes, I'm saying there are serious problems with the study, so it doesn't look like a competent effort to me and its results should be viewed with caution.

Volat
08-05-2015, 12:03 PM
Yes, I'm saying there are serious problems with the study, so it doesn't look like a competent effort to me and its results should be viewed with caution.

What kind of problems?

Generalissimo
08-05-2015, 12:35 PM
What kind of problems?

For instance, the authors failed to include two well known and very important R1a subclades in their analysis: the Northwest European-specific R1a-CTS4385 and the East and Central European-specific R1a-Z280. As a result, the former is lumped with R1a-M417* and the latter with R1a-Z282*. In fact, Z280 is shown to be above Z282 in the topology of R1a-M420 (see Figure 1 here (http://www.nature.com/ejhg/journal/v23/n1/fig_tab/ejhg201450f1.html#figure-title)), which is plain wrong.

lgmayka
08-05-2015, 03:24 PM
For instance, the authors failed to include two well known and very important R1a subclades in their analysis: the Northwest European-specific R1a-CTS4385 and the East and Central European-specific R1a-Z280.
To be fair, the authors gave a specific justification (http://www.nature.com/ejhg/journal/v23/n1/full/ejhg201450a.html) for omitting Z280:
---
The marker Z280 was not used as it maps to duplicated ampliconic tracts.
---

Although YFull still calls the clade R-Z280 (http://yfull.com/tree/R-Z280/), YFull rates its three equivalent SNPs differently:
Z687 - 5 stars
Z91 - 4 stars
Z280 - 3 stars

Yseq offers an R-Z280 SNP panel for $88 (http://www.yseq.net/product_info.php?products_id=10169), but never actually tests Z280 because of its difficulty. Yseq tests Z91 instead.

parasar
08-05-2015, 03:45 PM
For instance, the authors failed to include two well known and very important R1a subclades in their analysis: the Northwest European-specific R1a-CTS4385 and the East and Central European-specific R1a-Z280. As a result, the former is lumped with R1a-M417* and the latter with R1a-Z282*. In fact, Z280 is shown to be above Z282 in the topology of R1a-M420 (see Figure 1 here (http://www.nature.com/ejhg/journal/v23/n1/fig_tab/ejhg201450f1.html#figure-title)), which is plain wrong.

I would not consider these to be significant problems.

It is a shortcoming that they did not fully sequence any sample from the CTS4385 line and therefore missed that branch, but the STRs are available to confirm the main L664 portion anyway.
Their 'fully sequenced' R1a samples:
P3 14.9x Ashkenazi R1a-M582
Burusho001 43.5x Pakistani: Burusho R1a-M560
Makrani002 227.8x Pakistan: Makrani R1a-Z2125
HGDP00243 3.6x Pakistani: Pashtun R1a-L657
Baloch001 194.4x Pakistani: Balochi R1a-L657
Cambodia191 7.2x Cambodian R1a-L657
20-230 173.7x European R1a-L1029
P0 9.1x European R1a-M558


Regarding Z280 and Z282, their fully sequenced samples do not give them the hierarchy as no fully sequenced R1a sample is Z282+Z280-.
Plus they also note: "The marker Z280 was not used as it maps to duplicated ampliconic tracts."

Their Z93 map, I would agree has problems due to the data-points used and the contouring.
http://www.nature.com/ejhg/journal/v23/n1/images/ejhg201450f3.jpg

For both M458 and Z93 (M17 in the subcontinent to the Altai), I think the maps in Underhill's prior paper are better.
M458:
http://www.nature.com/ejhg/journal/v18/n4/images/ejhg2009194f2.jpg

Z93 (essentially all M17 in the subcontinent to the Altai):
http://www.nature.com/ejhg/journal/v18/n4/images/ejhg2009194f1.jpg

Tomenable
08-05-2015, 08:23 PM
This map seems to be not bad (but IMO Eupedia's maps are still better, even though not peer-reviewed):

http://i49.tinypic.com/1177epk.png

BTW - I'm not saying that R1b or any other hg maps are much better. It is hard to accurately map frequency distributions based on not-so-large samples collected from relatively few distant locations. It's probably better to take a look at primary data for each sampled population or ethnic group, rather than at maps.

Tomenable
08-05-2015, 08:40 PM
This map has been around for a while now... but I was curious to see what people here thought of it? I'm surprised by such high frequencies in Pomerania.

5292

This and other similar maps were made by Maciamo from Eupedia. Maybe ask him about his sources? I also noticed some inaccuracies in some of his map, but they are generally good. They are usually well-sourced - he compiles his data from many publications (so he certainly used Underhill's data, but also data from other sources). One of main flaws of his maps is that - at least it seems so to me - he tends to "invent" frequencies in areas for which there is no sufficient data. Maybe this is what he did with Pomerania?

Ask him:

http://www.eupedia.com/forum/threads/30729-New-map-of-haplogroup-R1a-M458-(Y-DNA)


I have created a new map of R1a-M458, a lineage which I associate with the Corded Ware expansion and which peaks in West Slavic countries today. The Underhill et al. 2014 paper provided a rough distribution map of R1a-M458, but I wasn't satisfied by its accuracy (it is auto-generated, not hand-made like mine). For example I have drawn the exact data for the Caucasus and the Volga-Ural region. I have also recalculated some percentages based on my Y-DNA frequencies. It does not affect a lot of regions, but in some cases the gap is considerable, and Underhill et al. had far fewer samples. E.g. for Hungary I have over 800 samples totalling 29.5% of R1a, while Underhill et al. had 18.6% of R1a out of 113 samples. (...)

==================

Here is some data for R1a as a whole from Kayser 2005; Immel 2005; and Roding 2007:

Graz (Austria) - 42,9%

And some cities in Germany:

Dresden - 32,6%
Rostock - 32,4%
Halle at the Saale - 30%
Leipzig - 27,1%
Berlin - 22,3%
Magdeburg - 21%
Greifswald - 19,2%
Hamburg - 16,8%
Mainz - 8,4%
Muenster - 7,8%

Two of these cities - Rostock (32,4%) & Greifswald (19,2%) are in Mecklenburg-Cispomerania.

But for example Rebala 2012 got only 13,7% R1a in his Mecklenburg sample of 131. I'm not sure where exactly in Mecklenburg was his sample from, but his sample included only pre-war inhabitants. It seems that after WW2 frequency of R1a in present-day Germany increased due to the influx of "R1a-heavy" refugees from East-Central and Eastern Europe. But, on the other hand, the frequency distribution of R1a in Germany correlates also with Slavic-inhabited areas in the Middle Ages, and with frequency of Slavic toponyms:

Concerning the region of Mecklenburg-Hither Pomerania (Mecklenburg-Vorpommern):

http://www.anthrogenica.com/showthread.php?4974-Ethnic-patterns-of-toponyms-and-DNA-do-they-correlate


(...) I think that there is also a correlation between percent of toponyms of Slavic origin among all placenames and percent of certain haplogroups among inhabitants of those places in North-Eastern Germany.

For example in the region around Greifswald percent of Slavic toponyms is only 11,88% versus 88,12% Germanic toponyms, and percent of R1a among inhabitants is lower than in areas with higher percent of Slavic toponyms (for instance, in the city of Greifswald only 19,2% of males have R1a, while in the city of Rostock 32,4% have R1a - and area around Rostock has indeed a much higher frequency of Slavic toponyms than area around Greifswald).

Y-DNA patterns near Rostock and Greifswald cited above are from Kayser 2005; Immel 2005; Roding 2007.

Proportions of Slavic to Germanic placenames (toponyms) in several areas of North-East Germany:

area around Greifswald - 11,88% / 88,12%
area around Grimmen - 36,58% / 63,42%
area around Franzburg - 32,51% / 67,49%
the island of Rügen - 79,21% / 20,79%

Source: http://koszalin7.pl/st/pom/pomorze_104.html

Out of those four places, Rügen has the highest % of Slavic toponyms. I would expect a lot of R1a there. In various parts of the islands of Rügen and Usedom, Slavic toponyms range from 75% to 88% of the total.

The Slavic origin of Usedom's population indicated by toponymy, is confirmed by archaeology and anthropology:

http://www.diss.fu-berlin.de/diss/receive/FUDISS_thesis_000000019471?lang=en


The archaeologically based assumption of a mainly Slavic population cannot be rejected with anthropological means.

A map showing Slavic placenames in that region (they are not evenly distributed throughout the land):

http://www.wizlaw.de/assets/images/Rujana-Karte2.jpg
(...)

=========================

Edit:

Check also this:

http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1001555


Europe-Wide Patterns of Relatedness

Individuals usually share the highest number of IBD blocks with others from the same population, with some exceptions. For example, individuals in the United Kingdom share more IBD blocks on average, and hence more close genetic ancestors, with individuals from Ireland than with other individuals from the United Kingdom (1.26 versus 1.09 blocks at least 1 cM per pair, Mann-Whitney p<10−10), and Germans share similarly more with Polish than with other Germans (1.24 versus 1.05, p = 5.7×10−6), a pattern which could be due to recent asymmetric migration from a smaller population into a larger population. (...)

What does it actually tell us ??? - that not all Germans are part-Bavarian, but most are part-Polish ???