Zana & Khwit of Abkhazia

[TanTin]

Zana and son
https://www.ebi.ac.uk/ena/browser/view/PRJEB45032

I have big big questions about Zana !
Did someone try to convert the bam files to plink?
 I see someone already provided G25 coordinates, but I don't trust that much.

So here is my version: I did myself a conversion from fastq/bam to plink and I tested the data for what should be Zana. 

There are 4 files:
SAMEA8804653
SAMEA8804654
SAMEA8804655
SAMEA8804656

The first one  Zana_hg19  - should be for Zana.
The coverage for this sample is more than excellent: 3x something. 


The authors of the study did all kind of test to determine she was of African origin.
However let start with some questions:

When I converted the file: to plink: I got 41k snips for Y chromosome.  41 k, imagine ! 
So this sample must be for a man.  There is no way to have 41 k snips and this to be a woman. 
Unless if this is the sample for the son of Zana: Khwit .
I haven't checked the Y chr for this sample,  I can do this as well but it takes some time.
Then I projected this sample and it went to completely different place, not much related to Africans. 
I may have some mistake with my scripts, merging etc..
Even if this is the data for Khwit:  it's not projected to where it is supposed to go. 

If you have already the converted files to plink format please share.

[TanTin]

@Pribislav , could you test the Y chr for Zana ?

https://www.ebi.ac.uk/ena/browser/view/PRJEB45032

I think there are 41 k snips provided for chr 24 for this sample.   SAMEA8804653

[AimSmall]

Target: abkhazia:Zana_final
Distance: 3.4772% / 0.03477159
45.0 Congo_Kindoki_Protohistoric
30.4 KEN_Kakapel_900BP
20.0 Congo_NgongoMbata_Protohistoric
4.0 Kenya_IA_Pastoral
0.6 MAR_EN

Target: abkhazia:Khwit_final
Distance: 1.4370% / 0.01436995
21.2 Congo_NgongoMbata_Protohistoric
12.6 Congo_Kindoki_Protohistoric
10.6 SWE_Gotland_VA
9.8 Sudan_EarlyChristian_R
9.6 Georgia_Kotias.SG
8.0 KEN_Kakapel_900BP
6.4 GEO_CHG
6.4 TUR_SE_Cayonu_PPN
5.6 RUS_Saltovo-Mayaki_low_res
4.6 SDN_MA_Kulubnarti
1.4 Russia_Shamanka_EBA.SG
1.4 Sweden_Viking.SG
1.2 China_Xinjian_IA_o2.SG
0.8 Russia_UstBelaya_Angara.SG
0.4 Russia_Yana_Medieval.SG


Didn't realize what these samples were until I was watching this video and a light went on.

DNA result of Zana of Abkhazia (Ape Woman) by Andrei DNA


Original run was this.... removed the USA_MD_Catoctin_Furnace_Africa samples to look closer at origins.

Target: abkhazia:Zana_final

Distance: 2.2243% / 0.02224332
60.6 USA_MD_Catoctin_Furnace_Africa
35.2 KEN_Kakapel_900BP
4.2 UGA_Munsa_500BP

Target: abkhazia:Khwit_final
Distance: 1.0952% / 0.01095245
32.2 USA_MD_Catoctin_Furnace_Africa
15.4 Georgia_Kotias.SG
11.6 KEN_Kakapel_900BP
10.4 Sudan_EarlyChristian_R
9.2 TUR_SE_Cayonu_PPN
8.2 SWE_Gotland_VA
4.0 GEO_CHG
2.6 Russia_LenaRiver_BA.SG
2.6 TUR_Arslantepe_LC
2.0 Sweden_Viking.SG
0.8 Russia_Siberia_Lena_EBA
0.6 Russia_Yana_Medieval.SG
0.2 Greater_Antilles_eastern_Ceramic
0.2 SDN_MA_Kulubnarti

[TanTin]

Didn't realize what these samples were until I was watching this video and a light went on.
DNA result of Zana of Abkhazia (Ape Woman) by Andrei DNA

In the video from Andrei he is explaining that Zana has some scores for Y-chromosome.

I did some tests for Y chr. and as per my initial estimation this sample is C-haplogroup, more precise: C2.

Then I did ran Y-lineage tracker and I got this result:

SampleID Haplogroup KeyHaplogroup Mutations LineageTrack
Map_with_BWA_on_data_81__mapped_reads_in_BAM_format_.bam K . * K(1/1)->IJK(1/6)->HIJK(1/1)->GHIJK(1/2)->F(4/68)->CF(1/3)->CT(5/199)->BT(8/375)->A1b(1/43)->A1(1/17)->A0T(1/30)->A00T(1/105)->A000T(1/8)->Y-Adam

(Sorry for OT, we may need to open new topic to discuss only for Zana)
(Actually on the screen from Andrei's video his software fund F and R calls for Y-chr .

[kolompar]

(Sorry for OT,  we may need to open new topic to discuss only for Zana)

I would like a Zana thread too, maybe a moderator could move the few posts about her to a new thread, would be nice if @ilabv could post the genotype files too.

[TanTin]

I did the conversion from fastq to plink , however the results for my converted file are too far from what it should be.
So I suspect that the format of these fastq is not hg19 but hg18.

In addition: there are 3 fastq files for each sample, because they did 3x coverage. I did my conversion only for 1 fastq file and coverage is still very high.
In addition to fastq there is also bam file provided which should have all 3 fastq together. I will try to convert the bam file to see if that will make some difference.

[TanTin]

There was a mistake in my previous script. I fixed it, now Zana is projected in Africa as it should be.

https://ufile.io/sf3km5zp

[kolompar]

Thanks for the thread.
Couldn't get it to work in admixtools but whatever, G25 might be more useful anyway.

Scaled
abkhazia:Khwit_final,-0.243581,0.075149,-0.001131,-0.013243,-0.019388,-0.001673,-0.00846,0.013846,-0.022498,-0.028247,-0.007145,0.008992,-0.018583,-0.000138,0.01045,-0.013922,0.017602,-0.010008,0.001634,0.008254,0.001123,0.009274,0.005176,-0.009881,0.005987
abkhazia:Zana_final,-0.586189,0.063978,0.015839,-0.002907,0.000308,0.002789,-0.028436,0.028153,0.008385,-0.028247,-0.007632,0.015736,-0.010852,-0.003441,0.012079,-0.012994,0.008996,-0.013049,0.013073,-0.005628,-0.000499,0.008408,-0.007272,-0.004338,-0.001317
Raw
abkhazia:Khwit_final,-0.0214,0.0074,-0.0003,-0.0041,-0.0063,-0.0006,-0.0036,0.006,-0.011,-0.0155,-0.0044,0.006,-0.0125,-0.0001,0.0077,-0.0105,0.0135,-0.0079,0.0013,0.0066,0.0009,0.0075,0.0042,-0.0082,0.005
abkhazia:Zana_final,-0.0515,0.0063,0.0042,-0.0009,0.0001,0.001,-0.0121,0.0122,0.0041,-0.0155,-0.0047,0.0105,-0.0073,-0.0025,0.0089,-0.0098,0.0069,-0.0103,0.0104,-0.0045,-0.0004,0.0068,-0.0059,-0.0036,-0.0011

These are the top two-way models I get for Zana unscaled with all modern averages:
0.012 abkhazia:Zana_final = 62% Cameroon_Bakoko + 38% Sudan_Nuba_Koalib
0.012 abkhazia:Zana_final = 55% Cameroon_Bakoko + 45% Chad_Maba
0.012 abkhazia:Zana_final = 51% Sudanese + 49% Yoruba
0.012 abkhazia:Zana_final = 54% Chad_Dangaleat + 46% Cameroon_Bakoko
0.013 abkhazia:Zana_final = 51% Yoruba + 49% Dinka
0.013 abkhazia:Zana_final = 52% Yoruba + 48% Ethiopian_Anuak

I find the 50/50 Yoruba+East African the most plausible. Wikipedia says there was an African community in Abkhazia and they might have been from Ethiopia while West Africa was the most popular source for slaves, so it makes sense that Zana would be maybe a first generation mix of them.
https://en.wikipedia.org/wiki/Abkhazians...an_descent

Khwit works almost perfectly:
0.014 abkhazia:Khwit_final = 51% abkhazia:Zana_final + 49% Adygei
0.014 abkhazia:Khwit_final = 55% abkhazia:Zana_final + 45% Abkhasian_Gudauta
0.014 abkhazia:Khwit_final = 53% abkhazia:Zana_final + 47% Georgian_Ajar
0.014 abkhazia:Khwit_final = 52% Karachay + 48% abkhazia:Zana_final
0.014 abkhazia:Khwit_final = 54% abkhazia:Zana_final + 46% Abkhasian

[szin]

I wasn't even aware there were black people in Abkhazia, so I wondered how someone of East African descent could have ended up there until I realised there was a slave trade.

[Mulay 'Abdullah]

"Ape woman".....and all seems right

[Xenus]

"Ape woman".....and all seems right

It's a well known story, a curiosity taken up by European/American anthropologists and was passed on to writers/journalists of all levels of credibility. I don't remember the the name but one of the "true believers" behind the first genetic tests on her son tried to claim that her ancestors split from other Africans tens of thousands of years ago. 

It all makes good sense when you frame her story as part folklore and part good  old timey racism