Page 17 of 282 FirstFirst ... 715161718192767117 ... LastLast
Results 161 to 170 of 2818

Thread: New Samples from Migration Era and Early Medieval Moravia

  1. #161
    Gold Class Member
    Posts
    2,675
    Sex
    Location
    United States
    Ethnicity
    LebaGermish
    Y-DNA (P)
    P312>DF19>DF88
    mtDNA (M)
    J2a1a1e
    Y-DNA (M)
    J2a1 Z6065>Y7708>M47
    mtDNA (P)
    H1j

    United States of America Lebanon Germany United Kingdom Belgium Switzerland
    Quote Originally Posted by Bygdedweller View Post
    There’s the IA-sample RISE174 from Scania, which sure enough is from a later timeframe, but has the necessary genetic profile to be ancestral to modern North Germanics and potentially other Germanic groups too. A Baltic/Uralic-shift, assuming that’s what it is, may have been present in small amounts to give it that profile, as it looks more eastern-like than what’s been found in some contemporary Germanic samples from the continent and elsewhere. There’s something to be said for differing ancestry-proportions from BA-populations too, but that’s another debate.
    Similarly, it looks to me as though both Íland Iron Age-samples (VK579 and VK522) have some type of Baltic/Uralic shift, although perhaps to a slightly higher degree than RISE174. VK579 is dated to 200-400 CE, that should be close chronologically (assuming these Wielbark-samples are indeed the same ones from Stolarek’s study), although we shouldn’t take for granted that the genetic landscape went unchanged from the time of the two preceding centuries. The Jutland Iron Age-sample from Alken Engen (VK582) is from the 1th century CE and rather low quality, but looks comparably western as shown in the PCA.

    Beyond that, I agree that there are many confounding factors. There’s the question of whether the samples represent anything like a homogenous population, the current low resolution of the samples and the possibilities of local admixture-events. That makes it hard to say whether the samples have Swedish-specific drift or not, but I’m finding the seeming overlap with IA-Swedes a bit too coincidental. It makes sense that such a genetic profile would have been brought over from the other side of the Baltic.
    Ahh, I was wondering what distinguished Íland Iron Age-samples.
    VK333 was buried there, but isotopes pointed to being raised somewhere else. However, when I ran him against all the IA sample averages available, it seems about half his ancestry looked like IA Íland:

    Target: VK2020_SWE_Oland_VA:VK333
    Distance: 1.8174% / 0.01817381
    46.0 VK2020_SWE_Oland_IA
    29.0 FRA_IA_north
    23.0 BGR_IA
    1.0 CHN_Yellow_River_LBIA
    0.8 KEN_Pastoral_IA
    0.2 VK2020_DNK_Sealand_IA

    Or, more specifically:
    Target: VK2020_SWE_Oland_VA:VK333
    Distance: 1.8163% / 0.01816272
    45.2 VK2020_SWE_Oland_IA:VK522
    25.0 FRA_IA_north
    23.2 BGR_IA
    4.8 VK2020_SWE_Oland_IA:VK579
    1.0 CHN_Yellow_River_LBIA
    0.8 KEN_Pastoral_IA

    ^^He was buried in a style typical for Íland, despite showing as "non-local," had perimortem sharp-force traumas (swords?) and had carved incisors.
    R1b>M269>L23>L51>L11>P312>DF19>DF88>FGC11833 >S4281>S4268>Z17112>BY44243

    Ancestors: Francis Cooke (M223/I2a2a) b1583; Hester Mahieu (Cooke) (J1c2 mtDNA) b.1584; Richard Warren (E-M35) b1578; Elizabeth Walker (Warren) (H1j mtDNA) b1583;
    John Mead (I2a1/P37.2) b1634; Rev. Joseph Hull (I1, L1301+ L1302-) b1595; Benjamin Harrington (M223/I2a2a-Y5729) b1618; Joshua Griffith (L21>DF13) b1593;
    John Wing (U106) b1584; Thomas Gunn (DF19) b1605; Hermann Wilhelm (DF19) b1635

  2. The Following User Says Thank You to Dewsloth For This Useful Post:

     Bygdedweller (04-19-2021)

  3. #162
    Registered Users
    Posts
    5,262
    Sex
    Location
    Zagreb
    Ethnicity
    Kaikavian Croatian
    Nationality
    Croatian
    Y-DNA (P)
    I2a1a2b-A815
    mtDNA (M)
    HV0-T195C!

    Croatia European Union
    Quote Originally Posted by Huijbregts View Post
    Are you confident that your method to decompose POH44 into a Slavic and a Germanic component is more reliable than an UMAP analysis?

    IMO 31% Germanic is quite a lot.
    In my UMAP analysis in #149 I find that POH44 plots in the center of the Balto-Slavic cluster or even more to the East.
    My list of nearest neighbors is in #158. For Migration.1 I find the next 14 nearest neighbors:

    Baltic_EST_MA:s19_IIf_1
    DEU_Krakauer_Berg_MA:KRA001
    VK2020_SWE_Gotland_VA:VK60
    Migration: POH41
    VK2020_POL_Bodzia_VA:VK156
    Scythian_UKR:scy009
    VK2020_POL_Bodzia_VA:VK154
    VK2020_UKR_Lutsk_MA:VK541
    HUN_Avar_Szolad:Av1
    HUN_Avar_Szolad:Av2
    DEU_Krakauer_Berg_MA:KRA011
    Migration: POH11
    DEU_Krakauer_Berg_MA:KRA003
    VK2020_SWE_Gotland_VA:VK439

    I don't doubt that this list is not exact, but it is hard to avoid the conclusion that these are very eastern samples.
    The only "German" samples in the list are the three DEU_Krakauer_Berg_MA samples.
    But Krakauer_Berg_MA is very eastern itself.
    My experience with UMAP is that it neglects minor parts of ancestry.

    If his Slavic part looks like on my map then his Germanic part also looks like on my map and ancestry breakdown gives such percents.
    Of course, if his Slavic part was more western shifted, then his Germanic part would look more eastern shifted. And percents of Germanic could be smaller if his Germanic is at the same time more western shifted.
    It is the play with possible components which I have choosen from the available samples from 1 CE to 850 CE.
     
    All simple calculations, maps and plots I make for free, but for more complicated maps and calculations I ask for a donation via Hidden Content PayPalHidden Content account

  4. The Following 3 Users Say Thank You to ph2ter For This Useful Post:

     Brent.B (04-21-2021),  leonardo (04-19-2021),  Michał (04-19-2021)

  5. #163
    Registered Users
    Posts
    5,262
    Sex
    Location
    Zagreb
    Ethnicity
    Kaikavian Croatian
    Nationality
    Croatian
    Y-DNA (P)
    I2a1a2b-A815
    mtDNA (M)
    HV0-T195C!

    Croatia European Union
    Quote Originally Posted by Brent.B View Post
    So PO44 is from the migration era, or early medieval?

    I was under the impression LIB samples are from migration era, and POH from early medieval?
    POH44 is from the second part of the 9th century.
     
    All simple calculations, maps and plots I make for free, but for more complicated maps and calculations I ask for a donation via Hidden Content PayPalHidden Content account

  6. The Following 2 Users Say Thank You to ph2ter For This Useful Post:

     Brent.B (04-19-2021),  Coldmountains (04-19-2021)

  7. #164
    Registered Users
    Posts
    276
    Sex
    Location
    Monts du Lyonnais
    Ethnicity
    European
    Nationality
    French
    Y-DNA (P)
    E-V22
    mtDNA (M)
    H4a1a

    France Lyon Switzerland Chile France Bourgogne
    Hi, which of these samples could be considered to be goth?
    G25
    Barcin 48.2; Yamnaya 41.2; WHG 9.4; LapaDoSanto 1.2

    K36 nMonte3
    1 Rheinland-Pfalz 1.17
    2 Baden-Wurttemberg 1.22
    3 Swiss_German 1.22
    4 Saarland 1.23
    5 Tirol 1.24
    6 Ile-de-France 1.26
    7 FR_Swiss 1.27
    8 Alsace 1.31

    Oracles
    FR_Savoie+FR_Savoie+Ukraine_Bukovina+South_Dutch 5,600388
    FR_Savoie+FR_Savoie+Ukraine_Bukovina+Norway_south-east 5,625698

    67% DE_Rheinland-Pfalz + 33% FR_Savoie
    62.7% FR_Normandy + 37.3% IT_Lombardia
    73% DE_Saarland + 27% IT_Lombardia

  8. #165
    Registered Users
    Posts
    2,065
    Sex

    Quote Originally Posted by ph2ter View Post
    My experience with UMAP is that it neglects minor parts of ancestry.

    If his Slavic part looks like on my map then his Germanic part also looks like on my map and ancestry breakdown gives such percents.
    Of course, if his Slavic part was more western shifted, then his Germanic part would look more eastern shifted. And percents of Germanic could be smaller if his Germanic is at the same time more western shifted.
    It is the play with possible components which I have choosen from the available samples from 1 CE to 850 CE.
    The only thing which might be missed is that there were not just Germanics, Slavs, Iranians and Balts around, but also Eastern Celts, as well as possible other, potentially closer to Illyrian-Pannonians and Dacians.
    It should be kept in mind that Germanics were newcomers and Slavs proper probably still not around or being just one among other people of the wider region.
    We really need to know more about the local, mostly Eastern Urnfield derived people, with the Lusatians being a central element.
    Like similar mixtures and positions on the PCA can be explained in very different ways then, just like in us moderns now.
    This is even more true in such a complex and variable sample, LIB11 in particular.

  9. The Following 6 Users Say Thank You to Riverman For This Useful Post:

     Alain (04-19-2021),  Brent.B (04-19-2021),  etrusco (04-19-2021),  leonardo (04-20-2021),  Megalophias (04-19-2021),  Michał (04-19-2021)

  10. #166
    Registered Users
    Posts
    1,052
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA (P)
    I2a2a1b2-CTS1977
    mtDNA (M)
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by ph2ter View Post
    My experience with UMAP is that it neglects minor parts of ancestry.

    If his Slavic part looks like on my map then his Germanic part also looks like on my map and ancestry breakdown gives such percents.
    Of course, if his Slavic part was more western shifted, then his Germanic part would look more eastern shifted. And percents of Germanic could be smaller if his Germanic is at the same time more western shifted.
    It is the play with possible components which I have choosen from the available samples from 1 CE to 850 CE.
    I don't understand your actual computation, but anyway thanks for your answer.
    IMO the problem that we face here, is that you try to split a not measurable tree (the ancestry of POH44) into two even less defined trees.
    Even in the best circumstances, that would demand a huge number of samples and a huge variance in the data.
    Your target is a single sample, and the variance of the comparison samples is severely restricted, by narrowly restricting them in space and time.
    In these circumstances all the doors are opened to overfitting.
    I don't see what criteria can be used to validate the result.

    I agree with you that UMAP has it own problems. The first problem is that UMAP is not a distance-based model but a connection based model.
    Relatively slight changes in included samples can cause unexpected changes in topology. So it can be hard hard to choose the optimum size of the model.
    Also UMAP needs a relatively large number of samples.
    But at least the algorithm is mathematically well defined.
    Although I cannot prove that an UMAP model is correct (after all it is mainly used as an exploration tool), there are often encouraging signs.
    For instance, in the case of POH44, its location is nearly exactly in the center of the Balto-Slavic cluster.
    Also the nearest neighbors are encouraging. For instance POH44 can choose 14 out of 647 samples as its neighbors; it would be very rare if it would choose all three of the DEU_Krakauer_Berg_MA samples.
    Also the number of POH samples as a near neighbor is above chance level.

    Of course you are right in asserting that UMAP neglects minor parts of the ancestry. If it didn't it wouldn't be a good model.
    I will appreciate concrete contra-examples. That is how we can learn.

  11. The Following User Says Thank You to Huijbregts For This Useful Post:

     ph2ter (04-19-2021)

  12. #167
    Registered Users
    Posts
    5,262
    Sex
    Location
    Zagreb
    Ethnicity
    Kaikavian Croatian
    Nationality
    Croatian
    Y-DNA (P)
    I2a1a2b-A815
    mtDNA (M)
    HV0-T195C!

    Croatia European Union
    Quote Originally Posted by Riverman View Post
    The only thing which might be missed is that there were not just Germanics, Slavs, Iranians and Balts around, but also Eastern Celts, as well as possible other, potentially closer to Illyrian-Pannonians and Dacians.
    It should be kept in mind that Germanics were newcomers and Slavs proper probably still not around or being just one among other people of the wider region.
    We really need to know more about the local, mostly Eastern Urnfield derived people, with the Lusatians being a central element.
    Like similar mixtures and positions on the PCA can be explained in very different ways then, just like in us moderns now.
    This is even more true in such a complex and variable sample, LIB11 in particular.
    In case of POH44 it is about 875 in Medieval Moravian dukedom. There were Slavic people and Frankish empire. Slavic-Germanic mix is perfectly plausible especially when we know that this man was a warrior of higher rank.
     
    All simple calculations, maps and plots I make for free, but for more complicated maps and calculations I ask for a donation via Hidden Content PayPalHidden Content account

  13. The Following 7 Users Say Thank You to ph2ter For This Useful Post:

     Alain (04-19-2021),  Brent.B (04-19-2021),  Coldmountains (04-19-2021),  leonardo (04-20-2021),  Michał (04-19-2021),  parastais (04-19-2021),  Riverman (04-19-2021)

  14. #168
    Registered Users
    Posts
    2,065
    Sex

    Quote Originally Posted by ph2ter View Post
    In case of POH44 it is about 875 in Medieval Moravian dukedom. There were Slavic people and Frankish empire. Slavic-Germanic mix is perfectly plausible especially when we know that this man was a warrior of higher rank.
    I agree for that sample, more generally speaking, but even in this case, I doubt that the Moravians at that time or in later times had to be exclusively Germanic and Slavic. The real question is what represents the pre-Slavic and pre-Germanic local population the best, and is there just one, or even more. Because the Celtic tribes might have been different from the unknown Lusatian-related, Pannonian, Sarmatian and so on. We would need probably a dozens of samples to assert that properly and the mix of samples we have from Moravia is very interesting, but not necessarily representative.

  15. The Following 6 Users Say Thank You to Riverman For This Useful Post:

     Alain (04-19-2021),  Brent.B (04-19-2021),  Megalophias (04-19-2021),  Michał (04-19-2021),  ph2ter (04-19-2021),  Ryukendo (06-02-2021)

  16. #169
    Registered Users
    Posts
    5,262
    Sex
    Location
    Zagreb
    Ethnicity
    Kaikavian Croatian
    Nationality
    Croatian
    Y-DNA (P)
    I2a1a2b-A815
    mtDNA (M)
    HV0-T195C!

    Croatia European Union
    Quote Originally Posted by Huijbregts View Post
    I don't understand your actual computation, but anyway thanks for your answer.
    IMO the problem that we face here, is that you try to split a not measurable tree (the ancestry of POH44) into two even less defined trees.
    Even in the best circumstances, that would demand a huge number of samples and a huge variance in the data.
    Your target is a single sample, and the variance of the comparison samples is severely restricted, by narrowly restricting them in space and time.
    In these circumstances all the doors are opened to overfitting.
    I don't see what criteria can be used to validate the result.

    I agree with you that UMAP has it own problems. The first problem is that UMAP is not a distance-based model but a connection based model.
    Relatively slight changes in included samples can cause unexpected changes in topology. So it can be hard hard to choose the optimum size of the model.
    Also UMAP needs a relatively large number of samples.
    But at least the algorithm is mathematically well defined.
    Although I cannot prove that an UMAP model is correct (after all it is mainly used as an exploration tool), there are often encouraging signs.
    For instance, in the case of POH44, its location is nearly exactly in the center of the Balto-Slavic cluster.
    Also the nearest neighbors are encouraging. For instance POH44 can choose 14 out of 647 samples as its neighbors; it would be very rare if it would choose all three of the DEU_Krakauer_Berg_MA samples.
    Also the number of POH samples as a near neighbor is above chance level.

    Of course you are right in asserting that UMAP neglects minor parts of the ancestry. If it didn't it wouldn't be a good model.
    I will appreciate concrete contra-examples. That is how we can learn.
    You are right. But the nature of G25 is such that it is not possible to know the exact ancestry of anyone. You can get the same result with different sets of kits.
    But my result is plausible and your umap only tells you that this person is in majority of the Slavic ancestry.
    When we try to compose his g25 values from available genetic profiles it is obvious that we can't get his values without Germanic components. Your umap can't tell you anything about this fact.
    Last edited by ph2ter; 04-19-2021 at 08:43 PM.
     
    All simple calculations, maps and plots I make for free, but for more complicated maps and calculations I ask for a donation via Hidden Content PayPalHidden Content account

  17. The Following 2 Users Say Thank You to ph2ter For This Useful Post:

     Brent.B (04-21-2021),  Michał (04-19-2021)

  18. #170
    Gold Class Member
    Posts
    452
    Sex
    Y-DNA (P)
    R-Y33
    mtDNA (M)
    J1c2
    Y-DNA (M)
    E-Y6938
    mtDNA (P)
    G2a

    Russian Federation Russia Tatarstan Arms Russia Mordovia Timurid
    Been trying to assign a deeper subclade for the R-L151 individual - POH27.

    It looks like he's R-Z49*, more specifically he is derived for several SNPs that define the R-BY55682 subclade and at ancestral state for others.


    Here's a lit of derived SNPs for POH27 that aren't on YFull but included on the FTDNA tree for R-BY55682

    BY128477 (2 reads)
    BY96884 (2 reads)
    BY199460 (3 reads)
    YDNA (P): R-Y33
    YDNA (P, maternal line): R-Y20756
    YDNA(M): E-Y6938

  19. The Following 9 Users Say Thank You to altvred For This Useful Post:

     Brent.B (04-19-2021),  Bygdedweller (04-19-2021),  Dewsloth (04-19-2021),  Megalophias (04-19-2021),  Michał (04-19-2021),  Principe (04-20-2021),  Riverman (04-19-2021),  Strider99 (04-20-2021),  Waldemar (04-20-2021)

Page 17 of 282 FirstFirst ... 715161718192767117 ... LastLast

Similar Threads

  1. Medieval Pashtun Samples
    By J Man in forum Central
    Replies: 0
    Last Post: 12-28-2020, 05:37 AM
  2. G25 Distance Maps to selected Early Medieval Samples
    By ph2ter in forum Autosomal (auDNA)
    Replies: 80
    Last Post: 05-09-2020, 02:33 PM
  3. Early Medieval Czech DNA (years 600-900 AD)
    By Tomenable in forum Central
    Replies: 55
    Last Post: 04-19-2018, 01:07 PM
  4. Early Medieval Germanic barrow burials
    By JonikW in forum History (Medieval)
    Replies: 1
    Last Post: 03-31-2018, 12:34 PM
  5. Replies: 0
    Last Post: 05-27-2016, 06:50 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
  •