by Jorma Jormakka, Feb. 15, 2020
I found this very interesting article by Kalevi Wiik from the Web:
http://www.jfpalmen.nl/files/Wiik_Haplogroup_R1a.pdf
The main thesis in the article is that there are two types of R1a men: the Scandinavian type and the General type. This claim seems reasonable based on the data presented. However, I would add that there can be populations that have independently developed similar mutations and if populations are studied as groups, such independent mutations can easily be interpreted as admixed populations. I will show that this reservation to Wiik’s thesis can be justified. I will look at the markers he mentions in the paper using as a test person one Polish (already deceased) man. I have his 67 marker Y-DNA results from the FamilyTreeDNA (and the permission to use them, but I mention only some marker values).
The haplogroup predictor of FamilyTreeDNA, and the Nevgen haplogroup predictor in https://www.nevgen.org/ give R1a1a (R-M198) as the predicted value. As R1a1a is a very large haplogroup, this result does not tell much. Wiik’s excellent tables are very useful in obtaining more information by looking at individual SNPs. Furthermore, from his tables I got a more correct result of his ancient origins than I had earlier got from the Polish R1a study group in the FamilyTreeDNA, though they looked at the same SNPs.
The most useful SNPs in this case are as follows: (SNP name = the value for the test person.)
DYS19=14. This is a rare value: usually DYS19 is 15-17. From page 29 in Wiik’s paper we se that DYS19=14 occurs most often in Poland in Pol (G). As the test person is Polish, we can just based on this result predict Pol (G).
DYS481=24. This value is quite rare in Poland. Page 53 in Wiik’s text shows that the likely place is Pol (G).
YCAIIb=24. Again, this is a rare value. Page 43 in Wiik tells that in Poland this value for YCAIIb is relatively common only in Pol (G).
DYS391=11. Wiik’s page 33 shows that this value does occur in Poland, especially, it occurs also in Pol (G).
DYS447=23. Page 41 in Wiik gives three parts of Poland: Pol (P), Pol (N) and Pol (G). We can rule out Pol (N) because DYS439=10. On page 35 Wiik tells that DYS439=10 does not occur in Pol (N).
DYS537=11. According to page 47 in Wiik, this value is very common in Pol (N), but Pol (N) is not possible because of DYS439=10. DYS537=11 does occur in Pol (G) often enough.
DYS442=14. This value is common in Poland except for Pol (E) and Pol (B) according to Wiik’s page 45. Especially, it does occur in Pol (G).
DYS565=13. This is a typical value in all Poland, see Wiik page 55. Pol (G) has somewhat less of DYS565=13 than other areas, but enough so that Pol (G) cannot be rejected.
DYS520=21 shows that the test person is not typical Baltic R1a of types Ky, Ku or Kt, page 12 in the document. From the same page we can make a number of observations. The East Slavic R1a of type E has DYS447=25. As the test person has DYS447=23, he is not a typical member of type E. Likewise, the test person, with DYS388=12, is not a typical member of North-Western R1a with DYS388=10. The test of 67 markers does not include DYS399. Instead, DYS393=13 for the test person, which agrees with a typical Old European R1a of type C.
Page 15 in Wiik gives two populations named Central European 1a and 1b. The former has DYS464 as 12-12-15-15-16 (possibly an error as DYS464 has only a,b,c,d) and 12-15-15-16 respectively. The test person with DYS464 as 13-15-15-16 is not a typical member of either population.
We have still a few markers left:
DYS594=10 is a typical value in all Poland, Wiik’s page 51. Especially Pol (G) has 100% of this marker value.
DYS406=11. The value excludes Pol (B) (page 49). Many areas of Poland have this value, including Pol (G).
DYS389I=13. Page 39 in Wiik shows that the value is typical in all Poland, including Pol (G).
DYS439=10. The value excludes Pol (N). Elsewhere in Poland, including Pol (G), this value occurs, page 35.
All markers mentioned by Wiik support the conclusion that the test person belongs to Pol (G) and this is the most likely result as the family of the test person has lived in the area of Pol (G) in central Poland in all known history of the family.
It is interesting to notice that Wiik’s tables gave a better result than the Polish group in FamilyTreeDNA, who proposed that the origins of the person are in Pomerania. Such a proposal is natural in the light of the theory, also supported by Kalevi Wiik, that there are two R1a populations: one with the center in Scandinavia and radiating to Pomerania, England, and to places where Vikings went, and another R1a population in Slavic areas. In the general case this theory is reasonable: there are these two populations, but this theory does not explan the origin of the population where the test person belongs (that is, Pol (G).
Let us see the markers by which Wiik divides these two populations (page 58):
Marker value General type Scandinavian type Test person from Pol (G)
DYS19 16-17 15 14
DYS391 10 11 11
YCAIIb 23 21 24
DYS442 13-14 12 14
DYS594 10 11 10
DYS565 13 12 13
The last three markers of the test person are as in the General type, while the first two are as in the Scandinavian type (DYS19 can have values 13-15 in the Scandinavian type). The value of YCAIIb is close to the value in the General type. As these markers are in a single Y-chromosome (in the so called “junk” DNA which is not affected by natural selection), we must conclude that the type originally has been of the General type and mutations have made it resemble the Scandinavian type in some markers. Thus, we would not expect paternal descent from the Scandinavian type for Pol (G).
We can still estimate if the test person belonged to the R1a1 branch R1a-Z93 or to the European branch R-Z283. Ashkenazi Levite Modular haplotype can be found from page 18 in
Calculating the differences in the STR data we find 11 one point differences, 5 two-point differences, 2 three-point differences and one 4 point difference between the Ashkenazi Levite Modular haplotype and the STR data of the test person. As mutations mainly happen one at a time, we can estimate 31 mutations. In 67 marker test one mutation is about 210 years. We get the time deviation of about 6500 years. The Levite haplotype is believed to have developed from one man who lived 1500 years before present. He belonged to R1a-Z93. R1a-Z93 is about 4000-4500 years old. Assuming that the test person belongs to a tree that separated from R1a-Z93 before R1a-Z93 was formed, the difference to the Levite modular should be 8000-1500 to 9000-1500 years, i.e., 6500 to 7500 years, which matches very well with 6500 years difference. If the test person belonged to R1a-Z93 for some time, say 2000 years, and then separated, we should get the difference 4500 to 5500 years. That is too little for the 6500 years. Most probably the test person does not belong to R1a-Z93 and that implies that he very probably is in the R-Z283 branch. We see that investigating marker values can give a better haplogroup prediction than what the predictors offer.
References:
Kalevi Wiik: Haplogroup R1a Maps and Tables, Feb. 2012, freely available at: