Page 3 of 18 FirstFirst 1234513 ... LastLast
Results 21 to 30 of 171

Thread: New Simulated AASI G25 Coordinates (+ Updates)

  1. #21
    Registered Users
    Posts
    920
    Ethnicity
    Brahmin (mixed)
    Nationality
    Indian
    mtDNA (M)
    M-30
    Y-DNA (P)
    R-1A (Z-93)

    Quote Originally Posted by DMXX View Post
    @All,

    You're very welcome Poi, and thank you for uploading the files so soon.

    Here's a reasoned run with 262.., Censored, Kulin and Reza (individuals such as these gents will be the ones who benefit from the expansion of AASI_S).

    Running first without my simulations or a living proxy for AASI (let's called this the "GenePlaza-worthy" model). All runs cyc=1k, bat=500, pen=def:

    Code:
    	Sample	Details					Fit	Map		IRN Shahr I Sokhta BA1	LAO Hoabinhian	NPL Chokhopani 2700BP	RUS Sintashta MLBA	TKM Gonur1 BA	TKM Namazga Tepe En
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;		7.694 	Open Map	13.4			61.4		0			4.2			18.8		2.2
    2	Bengali_Bangladesh:Reza_AGUser	Modern;		8.9554 	Open Map	15.8			64.4		0			1.4			17.2		1.2
    3	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser	9.3652 	Open Map	47.6			41.2		0			0.8			10.4		0
    4	Custom:AGUser_censored				10.1407 Open Map	52.4			47.6		0			0			0		0
    Reza and Kulin, being ~60% ancient SE Asian with a fit of 10? Definitely GenePlaza-worthy.

    Now, without my simulations (replaced with the Onge, which we would've had to have resorted to a year ago).

    Code:
    Sample	Details								Fit	Map		IRN Shahr I Sokhta BA1	LAO Hoabinhian	NPL Chokhopani 2700BP	Onge	RUS Sintashta MLBA	TKM Gonur1 BA	TKM Namazga Tepe En
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;				7.8941 	Open Map	5.8			37.8		0			24.2	9.6			19.4		3.2
    2	Bengali_Bangladesh:Reza_AGUser	Modern;				9.3278 	Open Map	7			47.8		0			18	7.4			17.4		2.4
    3	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser			9.2133 	Open Map	50.4			1.8		0			39	2.2			6.6		0
    4	Custom:AGUser_censored						9.9598 	Open Map	54			5.6		0			40	0.4			0		0
    The fit either doesn't change or improves slightly, though the results don't make much sense, due to the inappropriate use of the highly-drifted Onge as an AASI proxy. Humourously, Lao_BA looks as good as the Onge based on this.

    Replacing Onge with my AASI_S:

    Code:
    	Sample	Details					Fit	Map		IRN Shahr I Sokhta BA1	LAO Hoabinhian	NPL Chokhopani 2700BP	RUS Sintashta MLBA	Simulated AASI South By DMXX	TKM Gonur1 BA	TKM Namazga Tepe En
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;		3.0433 	Open Map	7.6			13.2		4.8			12.2			36.2				23.6		2.4
    2	Bengali_Bangladesh:Reza_AGUser	Modern;		4.0586 	Open Map	9			10.6		7.2			9.8			38.8				21		3.6
    3	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser	4.0334 	Open Map	30.4			5.2		1			8.4			32.8				22.2		0
    4	Custom:AGUser_censored				3.4833 	Open Map	29.2			0		0			4.6			43.6				21.4		1.2
    Drastic improvements in the fits. Can we do better?

    Adding in AASI_NW:

    Code:
    	Sample	Details						Fit	Map		IRN Shahr I Sokhta BA1	LAO Hoabinhian	NPL Chokhopani 2700BP	RUS Sintashta MLBA	Simulated AASI NW By DMXX	Simulated AASI South By DMXX	TKM Gonur1 BA	TKM Namazga Tepe En
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;			2.6927 	Open Map	5.2			19.8		5			13.6			8.8				22.6				21.2		3.8
    2	Bengali_Bangladesh:Reza_AGUser	Modern;			3.5705 	Open Map	6.4			24.6		2.8			10.8			11.6				20.8				20.2		2.8
    3	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser		2.6733 	Open Map	27			5.2		2.6			11			11.8				20				22.4		0
    4	Custom:AGUser_censored					2.2382 	Open Map	32.2			3.6		1.2			10.2			13.4				24.6				14		0.8
    The fits have once more improved.

    You guys can probably find a more robust individualised fit for each individual, but that's a good start. There's probably a need for something WSHG-ANE-rich to account for the MLBA steppe inflation IMO, which could push things down to the excellent 1-2 fit range.
    Like I said before, we love you DMXX. This is pretty amazing stuff.
    No more puns tho.
    The onge coin one was too much

  2. The Following 4 Users Say Thank You to 26284729292 For This Useful Post:

     bmoney (07-15-2019),  DMXX (07-14-2019),  FrostAssassin0701 (07-14-2019),  MonkeyDLuffy (07-15-2019)

  3. #22
    Administrator
    Posts
    3,688
    Sex
    mtDNA (M)
    D4j5*
    Y-DNA (P)
    R2a*-M124 (L295-)

    England
    Update to the above model - Adding in MA1 seems to have improved things significantly and pushed us closer to the dream 1-2 fit range:

    Code:
    	Sample	Details							Fit	Map		IRN Shahr I Sokhta BA1	LAO LN BA	NPL Chokhopani 2700BP	RUS MA1	RUS Sintashta MLBA	Simulated AASI NW By DMXX	Simulated AASI South By DMXX	TKM Gonur1 BA
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;				2.4469 	Open Map	7.6			4.4		3.8			3.2	12.8			4.2				40				24
    2	Bengali_Bangladesh:Reza_AGUser	Modern;				2.995 	Open Map	8			5		4.8			2.6	11			9.2				35.4				24
    3	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser			2.6272 	Open Map	27.6			1		2.4			1.4	10.8			12.4				21				23.4
    4	Custom:AGUser_censored						2.0829 	Open Map	26.2			1.2		0.8			2.2	8.6			15.6				24.4				21
    I've also replaced Lao_Hao with Lao_BA (as we all know, excessively old samples yield worse fits generally speaking).

    [Edit]: Just realised something intriguing, now that we have a robust pair of AASI simulations...

    Observe the SiSBA1 and Gonur1_BA (BMAC) proportions here... Reza and Kulin are Gonur1_BA > SiSBA1, whereas it's the opposite for Censored and 26... - Could we finally be seeing evidence of a Dravidian:true BMAC split with respect to Iran-Turan agriculturalist admix?

    Quote Originally Posted by 26284729292 View Post
    Like I said before, we love you DMXX. This is pretty amazing stuff.
    No more puns tho.
    The onge coin one was too much
    AG be the exclusive love-fest;
    Pure platonic proceedings while we attain the best tests;
    The rest of academia and the 'net feel collective stress;
    As we G25-dive towards AASI,
    Supported by Kale's 'Adm sets;
    You didn't rule out rhymes,
    So I'll lay it public sans fixed fees,
    But a gentle reminder that;
    I accept free Andaman eCheese:

    dn84gZS.png

     

    The feeling is mutual - You gals and guys are exactly who I'd wanted to see in this community back when we launched in '12.

  4. The Following 6 Users Say Thank You to DMXX For This Useful Post:

     26284729292 (07-15-2019),  Censored (07-14-2019),  Kulin (07-14-2019),  MonkeyDLuffy (07-15-2019),  Reza (07-14-2019),  traject (07-14-2019)

  5. #23
    Moderator
    Posts
    774
    Sex
    Location
    Sarai-Batu
    Ethnicity
    Bengali
    Nationality
    Canadian
    mtDNA (M)
    A11-a*
    Y-DNA (P)
    J2b2a

    Canada Bangladesh Jammu and Kashmir India Pakistan Sri Lanka
    Quote Originally Posted by DMXX View Post
    Update to the above model - Adding in MA1 seems to have improved things significantly and pushed us closer to the dream 1-2 fit range:

    Code:
    	Sample	Details							Fit	Map		IRN Shahr I Sokhta BA1	LAO LN BA	NPL Chokhopani 2700BP	RUS MA1	RUS Sintashta MLBA	Simulated AASI NW By DMXX	Simulated AASI South By DMXX	TKM Gonur1 BA
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;				2.4469 	Open Map	7.6			4.4		3.8			3.2	12.8			4.2				40				24
    2	Bengali_Bangladesh:Reza_AGUser	Modern;				2.995 	Open Map	8			5		4.8			2.6	11			9.2				35.4				24
    3	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser			2.6272 	Open Map	27.6			1		2.4			1.4	10.8			12.4				21				23.4
    4	Custom:AGUser_censored						2.0829 	Open Map	26.2			1.2		0.8			2.2	8.6			15.6				24.4				21
    I've also replaced Lao_Hao with Lao_BA (as we all know, excessively old samples yield worse fits generally speaking).



    AG be the exclusive love-fest;
    Pure platonic proceedings while we attain the best tests;
    The rest of academia and the 'net feel collective stress;
    As we G25-dive towards AASI,
    Supported by Kale's 'Adm sets;
    You didn't rule out rhymes,
    So I'll lay it public sans fixed fees,
    But a gentle reminder that;
    I accept free Andaman eCheese:

    dn84gZS.png

     

    The feeling is mutual - You gals and guys are exactly who I'd wanted to see in this community back when we launched in '12.
    Awesome runs, DMXX! Btw, does this model indicate that ancestors of ethnic Bengalis largely bypassed the IVC, and that most of their Iranian farmer ancestry is through the BMAC? I seem to score the lowest SISBA1 as well as NW AASI in your model.

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

     DMXX (07-14-2019)

  7. #24
    Administrator
    Posts
    3,688
    Sex
    mtDNA (M)
    D4j5*
    Y-DNA (P)
    R2a*-M124 (L295-)

    England
    Quote Originally Posted by Kulin View Post
    Awesome runs, DMXX! Btw, does this model indicate that ancestors of ethnic Bengalis largely bypassed the IVC, and that most of their Iranian farmer ancestry is through the BMAC? I seem to score the lowest SISBA1 as well as NW AASI in your model.
    Per the edit in the post you quoted, it provisionally looks like SiSBA1 might be a catch-all component for the Iran_N rich ancestry that probably predominated in both the IVC and among the (assumed) Dravidian speakers across the Subcontinent.

    The Sintashta + AASI_NW (partial) + Gonur1_BA combination might well be what the pre-IVC Indo-Aryans carried, based on your and Reza's proportions (f.ex. ~60-65% BMAC, ~30-35% Sintashta, ~5% AASI_NW). Basically, a more BMAC and AASI-shifted version of modern Pamiris (and highly similar to the Kalash).

    That 8% SiSBA1, per the above framework, could've been picked up anywhere between the IVC through to historical Bangladesh. We don't have the necessary intermediaries to speculate where the bulk of it came from.

    If the above is true, then that'd put Bengalis at somewhere around 40-45% post-IVC Indo-Aryan. Don't know how well that compares with the Ror-Kalash substitute runs that others have attempted.
    Last edited by DMXX; 07-14-2019 at 08:19 PM. Reason: final comm

  8. The Following 4 Users Say Thank You to DMXX For This Useful Post:

     bmoney (07-15-2019),  Kulin (07-14-2019),  kush (07-15-2019),  Reza (07-14-2019)

  9. #25
    Bronze Class Member
    Posts
    2,210
    Sex
    Location
    Krasnoyarsk

    Siberian Tatars Tajikistan
    Guys, beware that SIS1 has about 5% AASI itself. It should not be used alongside the simulated AASI in my opinion(if the sole purpose is to determine AASI levels, otherwise its perfectly fine). As for the split between SIS1 and BMAC it's something that should certainly be investigated further.
    Last edited by Censored; 07-14-2019 at 08:36 PM.

  10. The Following User Says Thank You to Censored For This Useful Post:

     bmoney (07-15-2019)

  11. #26
    Administrator
    Posts
    3,688
    Sex
    mtDNA (M)
    D4j5*
    Y-DNA (P)
    R2a*-M124 (L295-)

    England
    You could swap it with Tepe Hissar_ChL just fine - Though it wouldn't markedly affect the outputs. Let's find out.

  12. #27
    Global Moderator
    Posts
    645
    Sex
    Location
    UK
    Ethnicity
    Bengali
    mtDNA (M)
    M4b1
    Y-DNA (P)
    R1a-L657

    Really interesting - thanks for modelling DMXX.

    Just re: the AASI simulations - in the absence of an Austroasiatic component in the run - I imagine there would be a preference for AASI based on S Indian tribals (plus whatever goes into Lao_LN_BA).

    Why does the addition of MA1 reduce fits? For the pre-historic ANE admixture? Would Sarazm be a better component than Gonur1?

    I ran all the Bengali samples we have using your modelling:

    Code:
    Sample Fit IRN Shahr I Sokhta BA1 LAO LN BA NPL Chokhopani 2700BP RUS MA1 RUS Sintashta MLBA Simulated AASI NW By DMXX Simulated AASI South By DMXX TKM Gonur1 BA
    Bengali_Bangladesh:Average 1.6443 7.5 4.17 5.83 3.33 8.33 9.17 33.33 28.33
    Bengali_Bangladesh:Kulin_AGUser 1.9046 9.17 6.67 4.17 6.67 12.5 8.33 31.67 20.83
    Bengali_Bangladesh:Reza_AGUser 2.3676 12.5 5 7.5 5 10 14.17 25.83 20
    Bengali_Bangladesh:Reza_BIL_AGUser 2.7928 6.67 4.17 6.67 2.5 11.67 10 31.67 26.67
    Bengali_Bangladesh:Reza_Cousin_AGUser 2.4359 11.67 5.83 5.83 6.67 7.5 11.67 25.83 25
    Bengali_Bangladesh:Reza_Father_AGUser 1.9711 10.83 5.83 5 5 8.33 13.33 29.17 22.5
    Bengali_Bangladesh:Reza_Feni_AGUser 2.6978 7.5 5.83 3.33 8.33 13.33 13.33 30 18.33
    Bengali_Bangladesh:Reza_Mother_AGUser 2.5121 11.67 3.33 5.83 5.83 8.33 6.67 31.67 26.67
    Bengali_Bangladesh:Reza_Wife_AGUser 2.7946 12.5 7.5 2.5 8.33 5.83 9.17 32.5 21.67
    Bengali_Bangladesh:Tanzil_AGUser 2.8716 13.33 8.33 4.17 5 8.33 12.5 28.33 20
    Brahmin_West_Bengal:Average 1.409 20.83 3.33 1.67 5 16.67 15 17.5 20
    Brahmin_West_Bengal:frostassassin_AGUser 2.4388 23.33 0 1.67 5 20 11.67 24.17 14.17
    Last edited by Reza; 07-14-2019 at 08:41 PM.
    Paternal; Y-DNA: R1a-L657> Y6> Y11> Y920*
    Paternal; mtDNA: M5b'c
    Maternal; Y-DNA: R1a-L657> Y6
    Maternal; mtDNA: M4b1

  13. The Following 8 Users Say Thank You to Reza For This Useful Post:

     bmoney (07-15-2019),  DMXX (07-14-2019),  FrostAssassin0701 (07-14-2019),  Kulin (07-14-2019),  kush (07-15-2019),  laltota (07-15-2019),  MonkeyDLuffy (07-15-2019),  poi (07-15-2019)

  14. #28
    Administrator
    Posts
    3,688
    Sex
    mtDNA (M)
    D4j5*
    Y-DNA (P)
    R2a*-M124 (L295-)

    England
    Quote Originally Posted by Reza View Post
    Why does the addition of MA1 reduce fits? For the pre-historic ANE admixture? Would Sarazm be a better component than Gonur1?
    Correct.

    Sarazm and Namazga don't seem to do as good a job as WSHG or MA1 alone. I ran a bunch of models looking into this several weeks ago. Surprisingly, MA1 was favoured over even WSHG. That's despite WSHG being only 10% E. Eurasian in formal stats (the rest was on the EHG-AG3 cline).

  15. The Following 4 Users Say Thank You to DMXX For This Useful Post:

     bmoney (07-15-2019),  Kulin (07-14-2019),  Reza (07-14-2019),  traject (07-14-2019)

  16. #29
    Administrator
    Posts
    3,688
    Sex
    mtDNA (M)
    D4j5*
    Y-DNA (P)
    R2a*-M124 (L295-)

    England
    Code:
    	Sample	Details	Fit	Map	IRN Tepe Hissar C	LAO LN BA	NPL Chokhopani 2700BP	RUS MA1	RUS Sintashta MLBA	Simulated AASI NW By DMXX	Simulated AASI South By DMXX	TKM Gonur1 BA
    1	Bengali_Bangladesh:Kulin_AGUser	Modern;	2.5893 	Open Map	10	3.8	4	4.2	11.6	4.6	40	21.8
    2	Bengali_Bangladesh:Reza_AGUser	Modern;	3.0095 	Open Map	11.6	4.2	5.4	3.4	10.2	10.2	34.8	20.2
    3	Brahmin_Khas_Nepal:poi1_AGUser	Modern;	3.536 	Open Map	0	1.8	2.8	5.6	15.2	9.4	13.4	51.8
    4	Brahmin_Tamil_Nadu_Mixed:26284729292_v3_AGUser		2.838 	Open Map	2.6	1.4	2.8	3	6.4	15.2	17.6	51
    5	Custom:AGUser_censored		2.1938 	Open Map	5.8	0.2	1.6	3	7	14.4	26.6	41.4
    6	Punjab_Ramgharia:MonkeyDLuffy_AGUser	Modern;	3.9247 	Open Map	0	2.8	1.6	4.8	14.4	7	11.8	57.6
    7	Punjabi_Jat:Sapporo_AGUser	Modern;	3.338 	Open Map	0	2	0.8	4	15.4	7.6	8	62.2
    No real change to AASI, as predicted based on the above values.

    Also worth noting that the fits have worsened. SiSBA1 is favoured over TH.

  17. The Following 5 Users Say Thank You to DMXX For This Useful Post:

     Censored (07-14-2019),  Kulin (07-14-2019),  MonkeyDLuffy (07-15-2019),  Reza (07-14-2019),  scobar (07-16-2019)

  18. #30
    Bronze Class Member
    Posts
    2,210
    Sex
    Location
    Krasnoyarsk

    Siberian Tatars Tajikistan
    "sample": "Irula:Average",
    "fit": 1.3417,
    "Simulated_AASI_South_by_DMXX": 70,
    "TJK_Sarazm_En": 27.5,
    "Simulated_AASI_NW_by_DMXX": 2.5,

    "sample": "Paniya:Average",
    "fit": 1.4231,
    "Simulated_AASI_South_by_DMXX": 78.33,
    "TJK_Sarazm_En": 20,
    "Simulated_AASI_NW_by_DMXX": 1.67,

    Good work, seems the non-AASI/Iran_N values for these two tribal groups are now in the generally accepted range, whereas in the past they were deflated. Not to mention they now score only noise level "NW AASI" whereas in the past they would often get a large chunk of it.

    While Im at it might as well run some more tribals, not using simulated AA as the updated version has not been released yet:

    "sample": "Birhor:Average",
    "fit": 1.2315,
    "Simulated_AASI_South_by_DMXX": 64.17,
    "LAO_LN_BA": 18.33,
    "TJK_Sarazm_En": 12.5,
    "Simulated_AASI_NW_by_DMXX": 5,

    "sample": "Chenchu:Average",
    "fit": 2.0976,
    "Simulated_AASI_South_by_DMXX": 50.83,
    "TJK_Sarazm_En": 40,
    "Simulated_AASI_NW_by_DMXX": 6.67,
    "LAO_LN_BA": 2.5,

    "sample": "Gond:Average",
    "fit": 1.4146,
    "Simulated_AASI_South_by_DMXX": 62.5,
    "TJK_Sarazm_En": 21.67,
    "LAO_LN_BA": 10,
    "Simulated_AASI_NW_by_DMXX": 5.83,

    "sample": "Sakilli:Average",
    "fit": 2.6104,
    "Simulated_AASI_South_by_DMXX": 54.17,
    "TJK_Sarazm_En": 38.33,
    "Simulated_AASI_NW_by_DMXX": 7.5,
    "LAO_LN_BA": 0,

    "sample": "Santhal:Average",
    "fit": 1.3468,
    "Simulated_AASI_South_by_DMXX": 63.33,
    "LAO_LN_BA": 16.67,
    "TJK_Sarazm_En": 15,
    "Simulated_AASI_NW_by_DMXX": 5,

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

     bmoney (07-15-2019),  DMXX (07-14-2019),  FrostAssassin0701 (07-14-2019),  kush (07-15-2019),  laltota (07-15-2019),  MonkeyDLuffy (07-15-2019),  poi (07-15-2019),  Reza (07-14-2019),  Tsakhur (07-16-2019)

Page 3 of 18 FirstFirst 1234513 ... LastLast

Similar Threads

  1. Replies: 8
    Last Post: 08-01-2019, 05:13 AM
  2. I got my g25 coordinates
    By Magnetic in forum Autosomal (auDNA)
    Replies: 86
    Last Post: 07-02-2019, 07:54 AM
  3. How to determine my AASI amount
    By Censored in forum Southern
    Replies: 8
    Last Post: 08-07-2018, 03:54 AM
  4. Replies: 5
    Last Post: 07-25-2018, 09:39 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
  •