Page 1 of 5 123 ... LastLast
Results 1 to 10 of 50

Thread: Thread for the Israel Megiddo outliers

  1. #1

    Thread for the Israel Megiddo outliers

    Finally got the southern levant dataset merged and ran qpAdm for Megiddo_MLBA_outlierfamily (I2200 & I2189) using Harney et al (2020 preprint) rotating strategy.

    1. The boy is definitely derived for R1a1a1, have checked it myself using the egenstrat file data. Dont think there will be more info to drill down further.


    2. Alalakh_MLBA does not work with any sources in qpAdm. Megiddo_MLBA works. This is not what G25 shows, so just noting this point.

    3. Testing 3 models with rotating strategy: Megiddo_MLBA + Rus_Catacomb &
    Megiddo_MLBA + Sintashta + Dzharkutan & Megiddo + Kubano_tersk_late

    a. Model 1: Right pops = Rightall (my usual eurasian right pops) + Sintashta + Dzharkutan

    Left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 73.7 +- 2.1
    Russia_Steppe_Catacomb: 26.3 +- 2.1
    p-value: 0.86
    result file: https://pastebin.com/90KpQVA5

    b: Model 2: Right pops = Rightall + Rus_catacomb

    left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 61.9 +- 3.4
    Russia_MLBA_Sintashta: 25.1 +- 3.4
    Uzbekistan_BA_Dzharkutan1: 13 +- 3.6
    p-value: 0.40
    Result file https://pastebin.com/58WPET9b

    c. Model 3: Rightpops = Rightall + Sintashta + Dzharkutan

    left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 68.4 +- 2.7
    Russia_North_Caucasus_MBA(KBD001 & 002): 31.6 +- 2.7
    pvalue: 0.51
    https://pastebin.com/K6bGRVBf

    Model 1 is simpler 2 way model and has higher probability than model 2 which is a 3 way model, although both are plausible. Model 3 is also simpler than model 2, and just slightly more probable.

    The PCA shows that these outliers fall on a straight cline between megiddo & north caucasus. And none of the BMAC outliers have the steppe heavy ancestry that is required as a source here.
    Such ancestry persists till 1100bce even in Ukraine, so we can be certain that such ancestry existed all through 2nd mill bce in russian south steppe although the region is extremely poorly sampled for this time period. The R1a is probably from this region as well, unless the kid turns out R-Y2/Y3 or R-L657. The connection of the north caucasus steppe with levant seems extensive with a lot of R1b's found in levant.

    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 2.2777% / 0.02277661
    69.6 Levant_Megiddo_MLBA
    30.4 UKR_Catacomb

    PCA plot
    Last edited by misnomer; 06-01-2020 at 07:45 AM.

  2. The Following 13 Users Say Thank You to misnomer For This Useful Post:

     Ariel90 (06-01-2020),  DMXX (06-12-2020),  grumpydaddybear (06-01-2020),  K33 (06-02-2020),  Kristiina (06-01-2020),  Megalophias (06-01-2020),  mukin (06-01-2020),  Onur Dincer (06-07-2020),  parasar (06-01-2020),  pegasus (06-01-2020),  pmokeefe (06-01-2020),  R.Rocca (06-01-2020),  Yaroslav (06-01-2020)

  3. #2
    Another simple two way model, with Armenia_MBA

    Right pops: Rightall + Sintashta + Dzharkutan

    left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 38.1 +-5.6
    Armenia_MBA.SG: 61.9 +- 5.6
    p-value: 0.5753

    Result https://pastebin.com/7nbL18LU

    A complicated model is not required at all

  4. The Following 3 Users Say Thank You to misnomer For This Useful Post:

     DMXX (06-12-2020),  grumpydaddybear (06-01-2020),  Yaroslav (06-01-2020)

  5. #3
    Registered Users
    Posts
    3,757
    Location
    Gonur Tepe

    Afghanistan Jammu and Kashmir United States of America Canada
    Quote Originally Posted by misnomer View Post
    Finally got the southern levant dataset merged and ran qpAdm for Megiddo_MLBA_outlierfamily (I2200 & I2189) using Harney et al (2020 preprint) rotating strategy.

    1. The boy is definitely derived for R1a1a1, have checked it myself using the egenstrat file data. Dont think there will be more info to drill down further.


    2. Alalakh_MLBA does not work with any sources in qpAdm. Megiddo_MLBA works. This is not what G25 shows, so just noting this point.

    3. Testing 3 models with rotating strategy: Megiddo_MLBA + Rus_Catacomb &
    Megiddo_MLBA + Sintashta + Dzharkutan & Megiddo + Kubano_tersk_late

    a. Model 1: Right pops = Rightall (my usual eurasian right pops) + Sintashta + Dzharkutan

    Left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 73.7 +- 2.1
    Russia_Steppe_Catacomb: 26.3 +- 2.1
    p-value: 0.86
    result file: https://pastebin.com/90KpQVA5

    b: Model 2: Right pops = Rightall + Rus_catacomb

    left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 61.9 +- 3.4
    Russia_MLBA_Sintashta: 25.1 +- 3.4
    Uzbekistan_BA_Dzharkutan1: 13 +- 3.6
    p-value: 0.40
    Result file https://pastebin.com/58WPET9b

    c. Model 3: Rightpops = Rightall + Sintashta + Dzharkutan

    left pops:
    Megiddo_MLBA_outlierfamily

    Megiddo_MLBA: 68.4 +- 2.7
    Russia_North_Caucasus_MBA(KBD001 & 002): 31.6 +- 2.7
    pvalue: 0.51
    https://pastebin.com/K6bGRVBf

    Model 1 is simpler 2 way model and has higher probability than model 2 which is a 3 way model, although both are plausible. Model 3 is also simpler than model 2, and just slightly more probable.

    The PCA shows that these outliers fall on a straight cline between megiddo & north caucasus. And none of the BMAC outliers have the steppe heavy ancestry that is required as a source here.
    Such ancestry persists till 1100bce even in Ukraine, so we can be certain that such ancestry existed all through 2nd mill bce in russian south steppe although the region is extremely poorly sampled for this time period. The R1a is probably from this region as well, unless the kid turns out R-Y2/Y3 or R-L657. The connection of the north caucasus steppe with levant seems extensive with a lot of R1b's found in levant.

    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 2.2777% / 0.02277661
    69.6 Levant_Megiddo_MLBA
    30.4 UKR_Catacomb

    PCA plot
    . In your qpAdm you had Steppe Eneolithic as 30-35% Geoksyur or some Turan source its being masked in Yamnaya since Yamnaya itself is mainly Eneolithic Steppe. A direct Corded Ware/Yamnaya in the Levant would have massive implications akin to what happened in Europe in the EBA, but this never happened in the Eastern Mediterranean . The city states of the region in the MBA were shaped by military alliances based of expertise in chariot warfare. Those 3 outliers and the women are there at the right time and place. It looks like a red herring because of shared ancestral components , something which I stated here about ARM MBA. It was also the only working model for a long time with SPGT, though it made no sense, it passes qpAdm but does not reflect the archaeological and the right demic diffusions. I am also sure one would be able to get great models with Areni but that does not reflect actual population movements. Also, Sintashta is not a Central Steppe MLBA source, Sappali Tepe, Gonur, Teppe Hissar worked better than Dzharkutan. The O2 sample has significant Kumsay EBA ancestry as well.

    I did this model on Vaaho,

    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 2.2533% / 0.02253286
    66.4 ARM_MBA
    33.6 Levant_Megiddo_MLBA


    In context of that R1a ,the model prefers a Steppe MLBA source


    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 1.1177% / 0.01117731
    59.6 SYR_Ebla_EMBA
    23.4 RUS_Sintashta_MLBA
    13.2 UZB_Sappali_Tepe_BA
    3.8 KAZ_Kumsay_EBA

    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 1.5099% / 0.01509882
    61.4 SYR_Ebla_EMBA
    26.2 KAZ_Oy_Dzhaylau_MLBA
    12.4 UZB_Sappali_Tepe_BA


    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 1.8984% / 0.01898439
    57.2 Levant_Megiddo_MLBA
    27.4 KAZ_Oy_Dzhaylau_MLBA
    15.4 UZB_Sappali_Tepe_BA

    In the plot they don't form a cline with Catacomb at all, they are going towards Central Asian Indo Iranians , who are also fixed for R1a.




    https://anthrogenica.com/showthread....l=1#post671539
    Last edited by pegasus; 06-01-2020 at 12:33 PM.

  6. The Following 3 Users Say Thank You to pegasus For This Useful Post:

     artemv (06-01-2020),  Coldmountains (06-01-2020),  Jatt1 (06-03-2020)

  7. #4
    Quote Originally Posted by pegasus View Post
    . In your qpAdm you had Steppe Eneolithic as 30-35% Geoksyur or some Turan source its being masked in Yamnaya since Yamnaya itself is mainly Eneolithic Steppe. A direct Corded Ware/Yamnaya in the Levant would have massive implications akin to what happened in Europe in the EBA, but this never happened in the Eastern Mediterranean . The city states of the region in the MBA were shaped by military alliances based of expertise in chariot warfare. Those 3 outliers and the women are there at the right time and place. It looks like a red herring because of shared ancestral components , something which I stated here about ARM MBA. It was also the only working model for a long time with SPGT, though it made no sense, it passes qpAdm but does not reflect the archaeological and the right demic diffusions. I am also sure one would be able to get great models with Areni but that does not reflect actual population movements. Also, Sintashta is not a Central Steppe MLBA source, Sappali Tepe, Gonur, Teppe Hissar worked better than Dzharkutan. The O2 sample has significant Kumsay EBA ancestry as well.

    I did this model on Vaaho,

    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 2.2533% / 0.02253286
    66.4 ARM_MBA
    33.6 Levant_Megiddo_MLBA


    In context of that R1a ,the model prefers a Steppe MLBA source


    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 1.1177% / 0.01117731
    59.6 SYR_Ebla_EMBA
    23.4 RUS_Sintashta_MLBA
    13.2 UZB_Sappali_Tepe_BA
    3.8 KAZ_Kumsay_EBA

    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 1.5099% / 0.01509882
    61.4 SYR_Ebla_EMBA
    26.2 KAZ_Oy_Dzhaylau_MLBA
    12.4 UZB_Sappali_Tepe_BA


    Target: Levant_Megiddo_MLBA_o1:I2200
    Distance: 1.8984% / 0.01898439
    57.2 Levant_Megiddo_MLBA
    27.4 KAZ_Oy_Dzhaylau_MLBA
    15.4 UZB_Sappali_Tepe_BA

    In the plot they don't form a cline with Catacomb at all, they are going towards Central Asian Indo Iranians , who are also fixed for R1a.




    https://anthrogenica.com/showthread....l=1#post671539
    There is 0 R1a in BMAC or any of its steppe mixed outliers out of 32 male samples till 1400bce. Megiddo outlier samples are around 1600bce.
    However Steppe at this time is full of R1a. So the logical conclusion given all these facts + the qpAdm models is that the R1a reached Megiddo through caucasus straight from piedmont steppe, unless the R1a in Megiddo is L657 or Y3. There are almost no samples from this steppe region between 2000-1000bce, so theres possibility that corded ware/srubnaya R1a came to this region in this time period.

    In this regard, the southern levant paper is absolutely correct when it says that Arm_MBA type source is required. A high pvalue 2 source model is always preferable over a 3-4 source model especially when armenia/caucasus is much more proximal to levant than bmac.

    Edit:

    Megiddo_mlba + Oy_Dzhaylau+Sapalli Tepe gives a lower p-value than even the model with sintashta.

    Right pops: Rightall + Rus-Catacomb

    Left
    Megiddo_MLBA
    Kazakhstan_MLBA_OyDzhaylau
    Uzbekistan_BA_SappaliTepe

    p-value: 0.19

    https://pastebin.com/U1Hvdibv
    Last edited by misnomer; 06-01-2020 at 01:46 PM.

  8. The Following User Says Thank You to misnomer For This Useful Post:

     Megalophias (06-01-2020)

  9. #5
    Registered Users
    Posts
    3,757
    Location
    Gonur Tepe

    Afghanistan Jammu and Kashmir United States of America Canada
    Quote Originally Posted by misnomer View Post
    There is 0 R1a in BMAC or any of its steppe mixed outliers out of 32 male samples till 1400bce. These samples are around 1600bce.
    However Steppe at this time is full of R1a. So the logical conclusion given all these facts + the qpAdm models is that the R1a reached Megiddo through caucasus straight from piedmont steppe, unless the R1a is L657 or Y3.

    In this regard, the southern levant paper is absolutely correct when it says that Arm_MBA type source is required. A high pvalue 2 source model is always preferable over a 3-4 source model especially when armenia/caucasus is much more proximal to levant than bmac.
    Well you have an R1a sample with both Steppe ancestry and BMAC ancestry and thats the same kind of ancestral packet you find in Steppe rich South Asians who lack ENA, so it must have entered as Ryu mentioned in the Bronze Age. The ancestral packet you see here is a literal mirror reflection of what you see in Steppe rich Indo Aryan groups. Also as Ryu stated earlier there is a gigantic shadow society of these Indo Iranian people living in the Tepes which have not been excavated, also there is no genomic data from Afghanistan from the BA, which was where the early Rig Vedic world was .

    Previously you said if it was R1a , "I would win", I did, but thats not the point. I was actually an advocate for a Yamnaya proxy for Indo Iranian groups but thats not who they were and I was wrong and moved on. R1a did not arrive from the Caucasus, this is understood. Also nit-picking clades does not change the status quo at all. R1a this deep in South Eurasia can only be linked with Steppe MLBA groups in Central Asia.

    ARM MBA has shared components that skews how it actually went down. Also, those Megiddo outliers don't sit on a cline between Megiddo MLBA and ARM MBA either and none of them are R1a nor would any group in Urmia be.

    Lets just say that if the Megiddo R turns out to be R1b, I win, and if its R1a you win.

  10. The Following 2 Users Say Thank You to pegasus For This Useful Post:

     Coldmountains (06-01-2020),  Jatt1 (06-03-2020)

  11. #6
    Moderator
    Posts
    1,399
    Sex
    Y-DNA (P)
    R1a-Z93
    mtDNA (M)
    H28/H1b5 (Wife)

    Quote Originally Posted by misnomer View Post
    There is 0 R1a in BMAC or any of its steppe mixed outliers out of 32 male samples till 1400bce. Megiddo outlier samples are around 1600bce.
    However Steppe at this time is full of R1a. So the logical conclusion given all these facts + the qpAdm models is that the R1a reached Megiddo through caucasus straight from piedmont steppe, unless the R1a in Megiddo is L657 or Y3. There are almost no samples from this steppe region between 2000-1000bce, so theres possibility that corded ware/srubnaya R1a came to this region in this time period.

    In this regard, the southern levant paper is absolutely correct when it says that Arm_MBA type source is required. A high pvalue 2 source model is always preferable over a 3-4 source model especially when armenia/caucasus is much more proximal to levant than bmac.

    Edit:

    Megiddo_mlba + Oy_Dzhaylau+Sapalli Tepe gives a lower p-value than even the model with sintashta.

    Right pops: Rightall + Rus-Catacomb

    Left
    Megiddo_MLBA
    Kazakhstan_MLBA_OyDzhaylau
    Uzbekistan_BA_SappaliTepe

    p-value: 0.19

    https://pastebin.com/U1Hvdibv
    well there is a big problem with this theory here. R1a was absent in Yamnaya, Catacomb and any ancient Armenian/South Caucasian samples. Modern day R1a among Armenians belongs to Iranic clades and is of rather recent origin. His R1a definetly arrived from Central Asia there is no point in denying it. He has complex ancestry and it is not easy to model him, but even if he had the best fit with only Catacomb and Yamnaya this would say nothing about the origin of his R1a. Many of this models have vastly different results when a bit different source pops are used. It is not even relevant here if he is Z93+ or L657+. R1a in this region not existed prior to Indo-Iranians. There is untill yet not any evidence for a Pre-Indo-Iranian existence of R1a-M417 in West Asia. So even if he turns out to be Z93- it is still much more likely that he belonged to a rare R1a lineage among Indo-Iranians like the R1a-Z283 in Srubnaya than to some R1a clade from Yamnaya, Ancient Armenians or Hittites.
    Last edited by Coldmountains; 06-01-2020 at 02:01 PM.

  12. The Following 5 Users Say Thank You to Coldmountains For This Useful Post:

     Alain (06-01-2020),  artemv (06-01-2020),  Jatt1 (06-03-2020),  pegasus (06-01-2020),  Sapporo (06-01-2020)

  13. #7
    Global Moderator
    Posts
    696
    Sex
    Location
    EU
    Ethnicity
    Finnish
    Y-DNA (P)
    Father N1c
    mtDNA (M)
    I5a

    To my knowledge, Sintashta has only yielded R1a1a1b2a2a-Z2125 and upstream. Z93 has many branches with a TMRCA of 4000-3000 ybp. In my opinion it is simplistic to claim that all Z93 must have passed through Sintashta if Sintashta only yields a limited set of branches, possibly only one branch, Z2125 and upstream.

    These branches seem not have been detected in ancient samples to date: R1a1a1b2a3-Y40 (TMRCA 4300), R1a1a1b2c-YP1451 (TMRCA 3800), R1a1a1b2g-Y34351 (TMRCA 3600).

    R1a1a1b2h-YP5586/YP5585 (TMRCA 4600) has been detected in Roopkund and Srubna:
    Srubno-alakulskaya Kazburun Bashkorstan kzb002 1875-1665 BC R1a1a1b2h-YP5585
    Roopkund pilgrims 770–887 AD Roopkund_A I6942/R57R1a1a1b2a1a-L657

    I am sorry that ISOGG 2019 and yfull are not fully aligned in this respect. I wanted to add the specification in letters, but ISOGG tree and yfull give different paths.

  14. The Following 6 Users Say Thank You to Kristiina For This Useful Post:

     Alain (06-01-2020),  artemv (06-01-2020),  Kulin (06-19-2020),  misnomer (06-01-2020),  parasar (06-01-2020),  Pribislav (06-01-2020)

  15. #8
    Gold Class Member
    Posts
    7,536

    Quote Originally Posted by Coldmountains View Post
    well there is a big problem with this theory here. R1a was absent in Yamnaya, Catacomb and any ancient Armenian/South Caucasian samples. Modern day R1a among Armenians belongs to Iranic clades and is of rather recent origin. His R1a definetly arrived from Central Asia there is no point in denying it. He has complex ancestry and it is not easy to model him, but even if he had the best fit with only Catacomb and Yamnaya this would say nothing about the origin of his R1a. Many of this models have vastly different results when a bit different source pops are used. It is not even relevant here if he is Z93+ or L657+. R1a in this region not existed prior to Indo-Iranians. There is untill yet not any evidence for a Pre-Indo-Iranian existence of R1a-M417 in West Asia. So even if he turns out to be Z93- it is still much more likely that he belonged to a rare R1a lineage among Indo-Iranians like the R1a-Z283 in Srubnaya than to some R1a clade from Yamnaya, Ancient Armenians or Hittites.
    I had heard that R1a was found in Yamna, but yes no published paper shows that.

  16. #9
    Moderator
    Posts
    1,399
    Sex
    Y-DNA (P)
    R1a-Z93
    mtDNA (M)
    H28/H1b5 (Wife)

    Quote Originally Posted by Kristiina View Post
    To my knowledge, Sintashta has only yielded R1a1a1b2a2a-Z2125 and upstream. Z93 has many branches with a TMRCA of 4000-3000 ybp. In my opinion it is simplistic to claim that all Z93 must have passed through Sintashta if Sintashta only yields a limited set of branches, possibly only one branch, Z2125 and upstream.

    These branches seem not have been detected in ancient samples to date: R1a1a1b2a3-Y40 (TMRCA 4300), R1a1a1b2c-YP1451 (TMRCA 3800), R1a1a1b2g-Y34351 (TMRCA 3600).

    R1a1a1b2h-YP5586/YP5585 (TMRCA 4600) has been detected in Roopkund and Srubna:
    Srubno-alakulskaya Kazburun Bashkorstan kzb002 1875-1665 BC R1a1a1b2h-YP5585
    Roopkund pilgrims 770–887 AD Roopkund_A I6942/R57R1a1a1b2a1a-L657

    I am sorry that ISOGG 2019 and yfull are not fully aligned in this respect. I wanted to add the specification in letters, but ISOGG tree and yfull give different paths.
    We can for quite sure say that most branches of modern day R1a-Z94+ are not from Sintashta. For example most Sintashta samples were R-S23592, which later became dominant in North Eurasia among Karasuk, Saka and Huns. But for example Indo-Aryan Y3+ (ancestral to and most other Iranic clades like Z2123, YP413 (dominant among Pashtuns) and F1345 were not found there. But this is not important because Sintashta is not represeting all Proto-Indo-Iranians and many of this clades were found in Sintashta-like cultures before and after. Sintashta is just one early Indo-Iranian culture and many tends to exaggerate it importance. Abashevo and eastern Fatyanovo will show a bigger diversity of Z94+ and here probably the major Indo-Iranian lines will be found.

    For example basal Y3+ was found in Sredny Stog (I6561), which looks like a PIE culture ancestral to Abashevo>Sintashta. Z2123 was found in Srubnaya (I0430) and YP431 in Tazabagyab culture. With better sampling we will find other basal clades related to other Indo-Iranian lines. Untill yet no Z2124+ or Y3+ clade was found in a Neolithic or Bronze Age cultures without an Indo-Iranian context/substrate at least i am not aware of such samples. There is a chance for basal Z93+ moving deep into the balkan already in the Sredny Stog stage but this Megiddo outliners don't really seem to have ancestry from the balkan and instead show BMAC-like admix so i very much agree with Pegasus and consider it much more likely that his R1a line is ultimately of Indo-Iranian orgin. The clear Indo-Aryan adstrate of local cultures in the region is also very much pointing to this.
    Last edited by Coldmountains; 06-01-2020 at 04:33 PM.

  17. The Following 2 Users Say Thank You to Coldmountains For This Useful Post:

     btmacdowell (06-03-2020),  Kulin (06-19-2020)

  18. #10
    Moderator
    Posts
    1,399
    Sex
    Y-DNA (P)
    R1a-Z93
    mtDNA (M)
    H28/H1b5 (Wife)

    Quote Originally Posted by parasar View Post
    I had heard that R1a was found in Yamna, but yes no published paper shows that.
    well this are rumours. Yamnaya is now one of the best samples cultures in the region and still has zero R1a. If R1a will be found than rather in the forest zone bordering CWC related groups but this is still far from the Caucasus.
    Last edited by Coldmountains; 06-01-2020 at 04:02 PM.

  19. The Following User Says Thank You to Coldmountains For This Useful Post:

     artemv (06-01-2020)

Page 1 of 5 123 ... LastLast

Similar Threads

  1. V88 in Ancient Israel?
    By Rb1-V88 in forum R1b Early Subclades
    Replies: 14
    Last Post: 06-29-2020, 06:35 PM
  2. Outliers and "Noise"
    By Ski in forum 23andMe
    Replies: 10
    Last Post: 10-16-2016, 02:38 PM
  3. Replies: 2
    Last Post: 09-18-2016, 08:59 PM
  4. mystery structure off Israel
    By MikeWhalen in forum Archaeology (Prehistory)
    Replies: 1
    Last Post: 05-23-2013, 03:24 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •