Question about R1a

[rothaer]

Thomas Krahn now told me this:

Cladefinder essentially looks for the longest consistent path with the YFull tree from the given derived and ancestral SNP results. Positive SNPs increase the scoring and negative SNPs that are inconsistent with the tree position decrease the scoring. The path with the highest score is considered as the most likely position on the tree. In a small link at the bottom we also give the second best position from the scoring.
The details can be found in the function getPathScores:
https://github.com/.../blob/master/C...mmonMethods.py
There are a few unreliable SNPs in that area of R1a on the YFull tree that may throw off the scores due to bleeding over from highly similar X chromosome locations. Especially microarray data from low quality samples may have some significant noise which affects the scoring.

Btw. I find it surprising that negative tested SNPs that are inconsistent with the tree position do not exclude that position entirely.

[Jana]

Same thing happened with the MyHeritage raw data of my father when I ran it on yseq cladefinder. He got a false Z2125.
When I made him take the 23andMe test he turned out to be CTS3402.

Thanks a lot for this. Than I assume it's kind of safe bet my step father also belonged to CTS3402.

Thomas Krahn now told me this:

Cladefinder essentially looks for the longest consistent path with the YFull tree from the given derived and ancestral SNP results. Positive SNPs increase the scoring and negative SNPs that are inconsistent with the tree position decrease the scoring. The path with the highest score is considered as the most likely position on the tree. In a small link at the bottom we also give the second best position from the scoring.
The details can be found in the function getPathScores:
https://github.com/.../blob/master/C...mmonMethods.py
There are a few unreliable SNPs in that area of R1a on the YFull tree that may throw off the scores due to bleeding over from highly similar X chromosome locations. Especially microarray data from low quality samples may have some significant noise which affects the scoring.

Btw. I find it surprising that negative tested SNPs that are inconsistent with the tree position do not exclude that position entirely.

I find it very surprising too, it should exclude!!

[rothaer]

I heard of more such wrong Z93 > Z2125 predictions, that turned out to be Z280, but in the "worst" cases even M458.

I told this accumulation of specific cases to Thomas Krahn and he explained that the YSEQ predictor is completely (and seemingly automated) relying on YFull. He thinks actually the given SNP selection for Z280 (!) might be the problem. If there is some out of ancient DNA wrongly concluded SNPs among them and this is tested negative for a testee, then the score for the likelyhood of the Z280 branch is going far down and the Z93 branch with Z2125 can get the highest likelyhood score instead. He also told that there have been some wrong assumptions detected in the tree and it may also be hard to draw the applicable conclusions from ancient DNA, especially if you take the possibility of back and forth mutations into account. These things are not carved in stone, he said.

[Jana]

I heard of more such wrong Z93 > Z2125 predictions, that turned out to be Z280, but in the "worst" cases even M458.

I told this accumulation of specific cases to Thomas Krahn and he explained that the YSEQ predictor is completely (and seemingly automated) relying on YFull. He thinks actually the given SNP selection for Z280 (!) might be the problem. If there is some out of ancient DNA wrongly concluded SNPs among them and this is tested negative for a testee, then the score for the likelyhood of the Z280 branch is going far down and the Z93 branch with Z2125 gets the highest likelyhood score instead. He also told that there have been some wrong assumptions detected in the tree and it may also be hard to draw the applicable conclusions from ancient DNA, especially if you take the possibility of back and forth mutations into account. These things are not carved in stone, he said.

Is he planning to fix it? Is it even possible for him to do?

[reboun]

Yes. He might be of Romani origin (Indo-Iranic connection). This is also another possibility other than predictor’s being wrong.

thanks God it's a subclade of Eurasian Z93, on wikipedia I read that Roma people can have Z-93 in Croatia and Hungary
https://en.wikipedia.org/wiki/Haplogroup_R1a

[rothaer]

Is he planning to fix it? Is it even possible for him to do?

He actually said that this problem had to be adressed to YFull. From this I concluded that the YSEQ Cladefinder is automatically linked to the current YFull information. So no, he will not fix it as a separate "special solution" at the cladefinder and deviate from the YFull information.

[Jana]

Yes. He might be of Romani origin (Indo-Iranic connection). This is also another possibility other than predictor’s being wrong.

Piss off Durkish animal.

[reboun]

Piss off Durkish animal.

Okay, but you should seek psychological help.

[reboun]

Olina got mad after I told her stepfather might have Romani ancestry. I advised her to seek psychological help, she unfriended me. Interesting, very interesting

[Jana]

Olina got mad after I told her stepfather might have Romani ancestry. I advised her to seek psychological help, she unfriended me. Interesting, very interesting

Leave my thread, donkey. Thanks.

Target: Stearsolina_step_father
Distance: 1.9517% / 1.95170006 | R2P

68.9 Polish_Podlaskie
31.1 Italian_Tuscan

Distance to: Stearsolina_step_father

3.01680294 Hungarian_North
3.12929705 Croat_North
3.68586489 Slovenian
4.00923933 Hungarian_Alföld
4.19652237 Slovak
4.22803737 Hungarian
4.55680809 Hungarian_Transdanubia+Budapest
4.80397752 Croat
5.00893202 Poland_Lemko
5.06759312 Croat_East
5.17273622 Ukrainian_Carpathian
5.99738276 Bosniak_Bosnia
6.12831951 Czech
6.14322391 Croat_West
6.43241790 Ukrainian_Galicia
7.33137777 Croat_South
7.78598099 Moldova_Ukrainian
7.87043836 Moldova_North
8.03473708 Bosniak
8.05677355 Csángó-Ceangău
8.56291422 Hungarian_Transylvania+Székely
8.63184221 Polish_Southeast
8.85250247 South_Polish
8.93376740 Sorb_Lusatia
9.00382141 Romania_Moldavia_North