Page 11 of 11 FirstFirst ... 91011
Results 101 to 109 of 109

Thread: What is the relationship between South Indian Dravidians and Austro-Asiatic Tribals?

  1. #101
    Registered Users
    Posts
    6,091

    Quote Originally Posted by bmoney View Post

    ...

    R1a1 - older in SA than we thought
    ...

    The Indian R1a subclade seems to be separated by a long time and deep in India hence the high diversity - this is why i asked about tribal R1a
    ...
    These variances and TMRCAs from STRs are useless IMO.

    When the pioneering Sengupta paper came out it made some astounding claims. Since the authors included Zhivotovsky and the top R1a1 researcher - Underhill - it was considered authoritative at that time:

    "Associated microsatellite analyses of the high-frequency R1a1 haplogroup chromosomes indicate independent recent histories of the Indus Valley and the peninsular Indian region ... In HGs R1a1 and R2, the associated mean microsatellite variance is highest in tribes (table 12), not castes. This is a clear contradiction of what would be expected from an explanation involving a model of recent occasional admixture ... it is entirely plausible that they harbor as-yet-undiscovered subsequent haplogroup diversification that approximates the phylogeographic patterns revealed for HG L ...
    Further, the relative position of the Indian tribals (fig. 6), the high microsatellite variance among them (table 12), the estimated age (14 KYA) of microsatellite variation within R1a1 (table 11), and the variance peak in western Eurasia (fig. 4) are entirely inconsistent with a model of recent gene flow from castes to tribes and a large genetic impact of the Indo-Europeans on the autochthonous gene pool of India. Instead, our overall inference is that an early Holocene expansion in northwestern India (including the Indus Valley) contributed R1a1-M17 chromosomes both to the Central Asian and South Asian tribes prior to the arrival of the Indo-Europeans."



    This was my interpretation of their (& Sahoo's) data in 2007:

    "Sengupta paper which says – “Indian tribals, (fig 6), the high microsatellite variance among them (table 12).” I was trying to distinguish between ‘within’ and ‘among.’ The other papers I had in mind were: 1. Cordaux, et al. (Independent Origins … 2004, Table 3) which showed molecular variance lowest within tribal populations but highest among them; 2. Corduax, et al (The North-east Indian … 2004) “low Y-chromosome diversity (both at haplogroup and STR levels);” 3. Gayden, et al (the Himalayas … May 2007) on Nepal which found low R1a1 age of 1k years within the indigenous Newar (Table 2).

    As far as the Sahoo paper is concerned they state – “the STR haplotype diversity on the background of R1a in Central Asia (and also in Eastern Europe) has already been shown to be lower than that in India” relying on a Kivisild et al table which showed variance of R1a at 0.32 for India 0.37 for Pakistan, 0.38 for Iran, 0.19 for Estonians, and 0.14 for Czechs. This does not take into account the high R1a1 age in the Balkans or even a paper on Iceland which showed the variance there as 0.36 and an age of 7467 years - it is known that Iceland was occupied in about 870AD – indicating that coalescent time and historical age are not the same. This same paper (Helgason et at 2000) gave HG3 variances as 0.50 in Norwegians and 0.60 in Scandinavians.

    Looking at the data in the Sengupta paper we see the following on R1a1 in tribals – Chakma TB 0/4, Jamatia TB 2/30 – both same, Mog TB 1/5, Mizo TB 0/27, Tripuri TB 1/21, Ho AA 0/30, Lodha AA 0/20, Santhal AA 0/14, Irula DR 0/30, Konda DR 0/30, Kota DR 2/16, Koya DR 2/27 - separation of one at two locations, Kurumba DR 0/19, Toda DR 1/8, Halba IE 4/21 – separation one total, Muria DR 0/21.

    A total of 11 R1a1 chromosomes spread over 16 tribes with significant isolation and geographic separation. With so may singletons or close pairs and one close quartet, I am not comfortable with making a statement on age or any statement at all except that R1a1 close to missing or very low in AA and TB tribals. Sahoo also confirms the near absence of R1a1 in AA & TB tribals in Table 3 – 1/85 AA, 1/51 TB. No determination of age should be made based on these singletons.

    In such a low data set one sample can skew results significantly eg the one Mog in Sengupta had DYS389AB at 20. Sengupta et al due to the lack of sufficient samples, I believe, do not make any age determination for in Table 11 for IE, TB, or AA tribes. Nevertheless one can glean from their diversity map in Figure 4, the R1a1 diversity in India on the Indus (0.40), decreasing eastward, southward, and northward, falling to low levels where DR tribals live (0.23), even lower where AA tribals live (0.17) and vanishing in areas where TB tribals live (0.06).

    Sahoo also reports a 3/20 (15%) R1a1 in the Chenchu. Cf 5/5 (100%) for UP Thakur, 6/7 (86%) for UP Khatri, 9/12 (75%) for Bihar Rajput, 13/18 (72%) for Bihar Brahmin, 12/20 for Bhumihar (60%). Please note the latter three live in very close proximity with Bhumij, Birhor, Ho, Kharia, Munda, and Santhal who collectively show 0/56 R1a1. How could have two groups –IE since the Holocene, and AA (who supposedly are earlier) and still have such a difference – almost being mutually exclusive?

    IMO, Dienekes is correct that we should look for recent historical explanations rather than prehistoric ones. For example, in the case of a small isolated island of 300000 residents like Iceland, where the R1a1 age is calculated as 7467 years, instead of assuming a Neolithic occupation of Iceland we should rather consider the significant genetic age possessed by the immigrants in 870AD and/or that the immigration was multiple sourced (eg from Sheltand – though this particular paper states that they saw little if any Gaelic influence.)."




    That is why to me Sharma's R1a1 paper that followed was so startling and then upsetting. That was the first paper that looked to have made a genuine discovery of old R1a1 STR and SNP branches in South Asia. But all for naught - they could not justify their results when challenged.
    Last edited by parasar; 03-07-2018 at 09:05 PM. Reason: typo

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

     bmoney (03-08-2018),  redifflal (03-07-2018)

  3. #102
    Banned
    Posts
    3,063
    Sex
    Y-DNA
    L1a1
    mtDNA
    M30

    These variances and TMRCAs from STRs are useless IMO. This is not an argument

    Ok first discount Newar or NE Indians - they seem to be a west moving TB pop with recent admixture dates, we both know this

    "Associated microsatellite analyses of the high-frequency R1a1 haplogroup chromosomes indicate independent recent histories of the Indus Valley and the peninsular Indian region" This is true - Dravidian and Central Indian tribal R1a is older than North Indian R1a (on average due to North Indians have newer clades in addition to older ones) is what im inferring from the studies and is consistent with everything we've read so far

    I'm not saying UP Brahmin R1a has the same age as tribal R1a. SA has had multiple migrations from the steppe which probably brought further clades of R1a J and L1a2.

    Also R1a is still more frequent in tribals than L or J

    Another conclusion which is contrary to yours

    Similarly, the coalescent time for M82-H1a haplogroup in Bhil and M17-R1a and M95-O2a in Sahariya was estimated to be 13.18±3.24 kya, 10.97±1.86 kya and 16.48±3.06 kya, respectively (Table 1). However, the expansion times of different lineages can be considered as the upper boundary of the migration rather than referring to the time of origin of these tribal groups. The above results suggests that the frequent Y-chromosomal haplogroups in this region are highly unlikely a newcomer or recently migrated, they are present there at least since pre-Neolithic time.

    http://journals.plos.org/plosone/art...l.pone.0032546

    Nevertheless one can glean from their diversity map in Figure 4, the R1a1 diversity in India on the Indus (0.40), decreasing eastward, southward, and northward, falling to low levels where DR tribals live (0.23), even lower where AA tribals live (0.17) and vanishing in areas where TB tribals live (0.06).

    This is the only study to suggest diversity is lower in DR tribals. This is still consistent with multiple R1a clade entry into northern SA such as the Sri Lankan R1a I posted from Yfull which seems to be significantly older, and the strong possibility that DR tribal R1a is much older than NW R1a on average- so low diversity due to less migratory events, but older (old TMRCA). This also is not inconsistent with the other studies showing that R1a is still as old as L in the south based on diversity

    Northern Indus or Southern Indus? - if southern Indus its a whole new ball game and I personally find it hard to believe as Yamna R1a should follow a Gilgit entry point based on deduction and distribution

    Anyway what do you think about R1a coming along with L and J (aside from some old J) along with the ANI component south mixing in that Moorjani timeframe.

    I say this because any tribe/caste that has non-negligible L or J in South India always has more R1a - my caste northern elite Nairs are 50% R1a based on 23andme relatives

    This would then suggest the Kushan mix dated by Moorjani brought further levels R1a into the north but that R1a was already there in the North since the Neolithic at least

    Look at the pattern here, South Indians 12% L1 and 11% J overall vs 27% R1a, R1a is the biggest lineage for a pop that had its last admixture event @ 4500 ybp:

    https://www.ncbi.nlm.nih.gov/pmc/art...1241/table/T1/

    All of what I cited is consistent with this Sengupta statement from your post:

    an early Holocene expansion in northwestern India (including the Indus Valley) contributed R1a1-M17 chromosomes both to the Central Asian and South Asian tribes prior to the arrival of the Indo-Europeans."

    But IMO later Indo-Iranian R1a made it to SA bringing the language shift

    @BMG any thoughts?
    Last edited by bmoney; 03-08-2018 at 01:15 AM.

  4. The Following 2 Users Say Thank You to bmoney For This Useful Post:

     redifflal (03-08-2018),  Sapporo (03-09-2018)

  5. #103
    Registered Users
    Posts
    6,091

    Quote Originally Posted by bmoney View Post
    These variances and TMRCAs from STRs are useless IMO. This is not an argument

    ...

    Another conclusion which is contrary to yours

    Similarly, the coalescent time for M82-H1a haplogroup in Bhil and M17-R1a and M95-O2a in Sahariya was estimated to be 13.18±3.24 kya, 10.97±1.86 kya and 16.48±3.06 kya, respectively (Table 1). However, the expansion times of different lineages can be considered as the upper boundary of the migration rather than referring to the time ...
    "M17-R1a was estimated from microsatellite variation within the haplogroup using the method described by Zhivotovsky et al. [19] and updated in Sengupta et al. [10]"

    I doubt even Zhivotovsky uses Zhivotovsky any more. Roy King has noted (Zhivotovsky's co-author on the Senguta update mentioned above) that the authors have realized it is problematic and have moved on.

    That is why I had mentioned it as useless.

    But it unfortunately is still cropping up.
    See also:
    Quote Originally Posted by lgmayka View Post
    ...

    Zhivotovsky's fudge factor was utterly illogical the very day it was published, as Ken Nordtvedt--the world's foremost expert in Y-STRs--made clear. For years, we all laughed at research papers which continued to apply Zhivotovsky's folly to haplogroup ages, with ridiculous results. Thankfully, those days are now mostly behind us.
    Quote Originally Posted by lgmayka View Post
    Quite hilarious! A 2015 paper that uses 12 Y-STRs and the infamous Zhivotovsky fudge factor. I didn't realize that the Paris Institute of Molecular Anthopology has an endowed chair for comedy.
    Last edited by parasar; 03-08-2018 at 06:37 PM.

  6. The Following User Says Thank You to parasar For This Useful Post:

     bmoney (03-08-2018)

  7. #104
    Banned
    Posts
    3,063
    Sex
    Y-DNA
    L1a1
    mtDNA
    M30

    Quote Originally Posted by parasar View Post
    "M17-R1a was estimated from microsatellite variation within the haplogroup using the method described by Zhivotovsky et al. [19] and updated in Sengupta et al. [10]"

    I doubt even Zhivotovsky uses Zhivotovsky any more. Roy King has noted (Zhivotovsky's co-author on the Senguta update mentioned above) that the authors have realized it is problematic and have moved on.

    That is why I had mentioned it as useless.

    But it unfortunately is still cropping up.
    See also:
    Could you break this down in layman terms? I'm not understanding why Zhivotovskys method is unreliable

    Whats the next best approximation we can use apart from this method?

  8. #105
    Registered Users
    Posts
    1,518
    Sex
    Ethnicity
    Assamese
    Nationality
    Indian
    Y-DNA
    R1a1
    mtDNA
    M13'46'61

    India China Tajikistan Mongolia Russian Federation Papua New Guinea
    Quote Originally Posted by bmoney View Post
    Could you break this down in layman terms? I'm not understanding why Zhivotovskys method is unreliable

    Whats the next best approximation we can use apart from this method?
    Please note that NE Indians are not just TB, my MTDNA haplogroup is more present in lahu and dai and i also score dai ancestry in Dna.land and geneplaza, ahom are more admixed, whereas the 5 other tai communities in NE India are primarily of Dai ancestry, Tai Phake, Tai Turung, Tai Aiton, Tai Khamti and Tai Khamyang as they only came to Assam on the invitation of the Ahom king in the last 200-250 years from Northern Myanmar or Northern Thailand. Ahom are more TB admixed but still have Dai traits, the ones admixed with Dai are deficient in Haemoglobin E as well as other genetic traits

  9. The Following 2 Users Say Thank You to jortita For This Useful Post:

     khanabadoshi (03-10-2018),  redifflal (03-09-2018)

  10. #106
    Registered Users
    Posts
    2,122
    Location
    Gonur Tepe

    Afghanistan Jammu and Kashmir United States of America Canada
    Quote Originally Posted by bmoney View Post
    Haplogroup age estimates among sampled South Indian pops

    Oldest L1a1 - 4,483 ybp
    Oldest R1a - 4,878 ybp

    http://journals.plos.org/plosone/art...l.pone.0050269
    I disagree with R1a preceding L1a , its strong corelation with proto Dravidian IMO makes that impossible, also the R1a in South India was largely brought by demic diffusion of Central Indians which occurs much later.

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

     bmoney (03-10-2018),  redifflal (03-11-2018)

  12. #107
    Banned
    Posts
    3,063
    Sex
    Y-DNA
    L1a1
    mtDNA
    M30

    Quote Originally Posted by pegasus View Post
    I disagree with R1a preceding L1a , its strong corelation with proto Dravidian IMO makes that impossible, also the R1a in South India was largely brought by demic diffusion of Central Indians which occurs much later.
    I agree too, but my argument is L1a in South India being the same age as R1a, which is what the studies are indicating - L1a was ancient only in the NW IMO.

    This probably means that L and certain J clades came as part of the Central Indian R1a expansion as every south Indian caste which has L or J has higher levels of R1a

    Parasar was suggesting that Tamil L1a1/L1a2 was some ancient pocket and older than any other in L clade in SA, and one of the oldest haplogroups in the south. The studies addressed this question using Tamil low castes and found that its as old as the R1a in them, which means the oldest R1a and L1a is not in the South.

    Also that proposal is completely at odds with Yfull; the oldest L1a1 is in Kuwait and Saudi (8700ybp TMRCA), and the L1a1 clades to which all the SAs belong (L-Z5924 and L-Z5927) have a TMRCA 4100 and 3700 ypb which makes it of a similar age as the oldest z94 in SA, R1a-Y40 with a TMRCA of 3900

    To me it suggests a coastal diffusion of L1a1 both west and east of southern Iran, with the west migration being earlier than the SA eastern migration. L1a2 would have a completely different story IMO, from northern Iran/Caucasus instead of Southern Iran
    Last edited by bmoney; 03-10-2018 at 03:33 AM.

  13. The Following User Says Thank You to bmoney For This Useful Post:

     Sapporo (03-10-2018)

  14. #108
    Global Moderator
    Posts
    3,244
    Sex
    Location
    Chicago
    Ethnicity
    Baloch Kashmiri Uzbek Kho
    Nationality
    USA
    Y-DNA
    R-Y17491 > R-YP4858
    mtDNA
    A8a

    Pakistan United States of America
    Quote Originally Posted by bmoney View Post
    Could you break this down in layman terms? I'm not understanding why Zhivotovskys method is unreliable

    Whats the next best approximation we can use apart from this method?
    Zhivotovsky (2004) published a paper in which he
    attempted to get around these difficulties by calculating
    an “effective mutation rate” that is empirically derived
    from a set of descendants of an ancestor who lived at a
    known time in the past. All of the unknown factors
    such as the genealogy or the population dynamics, are
    simply averaged out in calculating the effective mutation
    rate, assuming 25 years per generation (which may be
    too small). This can work well if there are a number of
    such case studies that can be analyzed and the resulting
    average rates can be averaged (Zhivotovsky averaged
    the rates from three population groups), and if the cases
    that are included are representative of the situation to
    which the derived rate is to be applied. In practice, it is
    not so easy to guess whether the case studies have the
    necessary characteristics to be appropriate.

    Zhivotovsky’s “effective” mutation rate is averaged over
    just a few traditionally measured markers. However,
    Chandler’s relative rates for all 37 markers can also be
    calibrated to Zhivotovsky’s effective rates, resulting in a
    complete set of individual effective rates. Or, any subset
    of the 37 can be averaged to obtain the corresponding
    effective average rate for any panel of markers. These
    would then be appropriate for application to a dataset
    of haplotypes where the genealogy is unknown and the
    time depth is similar to that used by Zhivotovsky for his
    determination of the effective rates.
    However, in using three different datasets and averaging
    the result from each, Zhivotovsky seems to have
    introduced a small problem: the markers used in the
    different datasets were not exactly the same, especially
    for the third dataset, so he was averaging rates over
    different markers. Even with unlimited sample size, the
    rates from the three groups should not be the same.
    Zhivotovsky averaged them anyway.

    However, we can illustrate the approach to recalibrating
    Chandler’s mutation rates to the effective rate by just
    using the results Zhivotovsky obtained on seven
    markers that were tested in the dataset of Maori and
    Cook Islanders, where he obtained a mutation rate
    averaged over the seven markers of .000705. The seven
    markers were: DYS19, DYS389I, DYS389II, DYS390,
    DYS391, DYS392, and DYS393. If we average
    Chandler’s father-son mutation rates over the same
    seven markers, we obtain an average value of .00183.
    Therefore, we can calculate the ratio of Zhivotovsky’s
    effective average rate to Chandler’s father-son average
    rate on those seven markers, as .000705/.00183 = .385.
    Zhivotovsky’s effective rate is just 38.5% of the father-
    son rate. Armed with that conversion factor, we can
    convert all 37 of Chandler’s individual father-son
    mutation rates to effective rates a la Zhivotovsky by
    simply multiplying each by 0.385. With this complete
    set of rates, we can then average them over any subset
    of markers if desired for a particular application.
    [http://www.jogg.info/pages/32/editorial.pdf]

    Basically, there is an inherent "fudge factor", it has to be re-calibrated.

    In order to understand microsatellites mutation mechanisms, reliable rate estimates for STRs, in general, and Y-STR, specifically, have long been considered of scientific interest. Different approaches have been utilized for the estimation of Y-STR mutation rates, some of which use direct counting such as counting mutation events in deep routed pedigrees with known history (Heyer et al., 1997), or in father–son pairs (Ge et al., 2009; Kayser et al., 2000), and also in sperm (Holtkemper et al., 2001). Other methods try to estimate the mutation rates indirectly; for example, Zhivotovsky et al., 2004, use the diversity of Y-STRs in modern-day population samples with documented population founding or population splitting events in the last 1000 years (Gypsies and Pacific Islanders). They use TD, an estimator for a population divergence time based on inter- and intra-population variance of STR repeat number (Zhivotovsky, 2001), to estimate the average STR mutation rates. For the eight Y-STRs they consider they get an estimate of 6.9 × 10−4±5.9 × 10−4 mutations per generation.

    All of the different approaches observed significant variation in the mutation rates of different STRs. See for example the comparable mean and standard error for mutation rates as estimated by Zhivotovsky et al. (2004). However, there is a roughly 3-fold gap between the rates estimated from geneologies and those estimated from historical or phylogenetic data (Zhivotovsky et al., 2004, 2006). This has generated extensive attention in the literature, with some explanations offered by Zhivotovsky et al. (2006), but in our view it largely remains unresolved.
    [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2935444/]

    tl;dr -- Use SNPs not STRs.

    Parasar has an older post where he explained it all very nicely, if I remember.
    Last edited by khanabadoshi; 03-10-2018 at 08:02 PM.
    “Chahar chez est tohfay Multan, Gard-o- Garma, Gada-o- Goristan”.

    Four things are the gift of Multan: Dusty winds, hot seasons, beggars and graveyards.




  15. The Following 3 Users Say Thank You to khanabadoshi For This Useful Post:

     bmoney (03-11-2018),  redifflal (03-11-2018),  Sapporo (03-12-2018)

  16. #109
    Registered Users
    Posts
    2,122
    Location
    Gonur Tepe

    Afghanistan Jammu and Kashmir United States of America Canada
    Quote Originally Posted by pegasus View Post
    Thats not possible ,for very simple reasons. NO large urbanized civilization with large populations is going to be a monolith, its going to be diverse, its not Mesolithic Europe, with scant populations. Also the Pania like ancestry found in Southern Iranians is not recent, some mtdna clades of M are archaic and have been found to specific to the Iranian plateau. In the same way Barcin farmers absorbed extra WHG once in Europe in the Neolithic, the same can easily be inferred for those early Neolithic Iranians in Balochistan. Also those admixture dates are not a good marker . Those Horner like samples from Tanzania pretty much proved that ie they found Horner like woman all the way down in Tanzania, and she was older than 3Kya date given for admixing. So you have Luxmunda types and Mota types existing in the same period.
    Well I predicted this lmao , its good to see my prediction was right esp with the 14 samples coming up.

Page 11 of 11 FirstFirst ... 91011

Similar Threads

  1. Replies: 5
    Last Post: 08-10-2017, 05:06 AM
  2. Replies: 18
    Last Post: 05-12-2016, 10:02 AM
  3. New: South Indian Karnataka Result (23andme)
    By pava@grahana.net in forum 23andMe
    Replies: 55
    Last Post: 04-18-2016, 05:12 AM
  4. Replies: 28
    Last Post: 08-30-2013, 09:38 PM
  5. South Indian Y-Chromosomes
    By J Man in forum Other
    Replies: 7
    Last Post: 12-22-2012, 10:06 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
  •