Daniel Tabin @DanTabin Twitter profile

Last Seen Profiles

@HulalaJmb

@aboutsjk

@cxqmaggie

@stwmaniax

@Artistatsydney

@sutasususu

@stw_pdg

@KibangouMi1985

@ErenKor53671853

@hiroshitsujino

@Vickimcc101

@CotyDavila3

@Taka32233

@ptrancyber

@trephacardcanon

@PositiveMaren

@sarichan_ai

@dark_angelik

@FabiAn83995036

@FParianen

@3Rap7K2psmxqlwy

@Deniizkzi

@lalaluvlalisa

@CrossFitMayhem

@die_BGE

@keik0vr

@NielsDVV

@muscle_god_

@ReerTargaryen

@freddyzinho

@kisakisa1162947

@noader7

@chiiiikawachan

@platon751

@Dewie_Depok

@ZTheril

Daniel Tabin

@DanTabin

21 days

If you are interested in the origins of Turkic languages and are at EMBL, please come find my poster (115) and me at poster session 1 on Wednesday. I'd love to hear your thoughts, comments, and critique! And if you aren't at EMBL, please feel free to email me or reply here!

31

91

544

Daniel Tabin

@DanTabin

11 months

I will be presenting a poster about my work on the origins of Turkic peoples this Saturday from 2:15 to 4:15 at #ASHG2023 (poster # PB3093). It is unpublished, preliminary, and subject to change! If you're interested in chatting, find me there, reach out here, or email me!

15

30

141

Daniel Tabin

@DanTabin

21 days

It feels worth mentioning that this is a more developed version of work that I posted on here earlier, but it remains unpublished, somewhat preliminary, and subject to change

Daniel Tabin

@DanTabin

11 months

I will be presenting a poster about my work on the origins of Turkic peoples this Saturday from 2:15 to 4:15 at #ASHG2023 (poster # PB3093). It is unpublished, preliminary, and subject to change! If you're interested in chatting, find me there, reach out here, or email me!

15

30

141

6

3

56

Daniel Tabin

@DanTabin

3 months

Interested in FST and or Information Theory? I have a poster connecting both at #SMBE2024 ! Find me at poster session 1 on Monday (or here) if you're interested and want to chat! This is unpublished, preliminary, and subject to change. I'd love thoughts, critique, and feedback!

0

3

23

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva This is almost surely not real. G25 (and its downstream methods) are based on a PCA which does not capture the variation found within Ranis / Zlaty Kun. These super early Eurasians seem to be outgroups to every Eurasian group that came after (including modern groups) 1/n

2

17

Daniel Tabin

@DanTabin

21 days

@nrken19 The samples I am using come from Kazakhstan. I don't think I can say any more details without first consulting the archaeologists

1

0

15

Daniel Tabin

@DanTabin

12 days

@nrken19 @rampadus61899 @Sulkalmakh Dzudzuana isn’t “CHG” as in the type of ancestry found in Satsurblia / mesolithic Georgia. It seems to be an earlier type of ancestry with more affinity to Anatolia Neolithic groups. The Paleolithic Georgian from a recent Willerslev group paper also seems Dzudzuana like

1

16

Daniel Tabin

@DanTabin

22 days

Lots of selection happened in our past! Much more than previously known or expected! Check out this amazing new work from @aliakbari23

Pervasive findings of directional selection realize the promise of ancient DNA to elucidate human...

We present a method for detecting evidence of natural selection in ancient DNA time-series data that leverages an opportunity not utilized in previous scans: testing for a consistent trend in allele...

www.biorxiv.org

1

3

9

Daniel Tabin

@DanTabin

21 days

@jack_meditates2 Biggest changes are 1) the emphasis on Bulan Koby being Shaz Turk specific, and 2) regional modeling which fixes things like Uralic ancestry in some Tatars appearing to be Bulan Koby ancestry. The idea of Bulan Koby being important and ancestral to modern Turkic peoples remains

1

0

9

Daniel Tabin

@DanTabin

20 days

@steppestones Nothing beyond the poster right now. I have not yet put out any preprint nor submitted it for publication, but I hope to do both at some point!

3

0

8

Daniel Tabin

@DanTabin

21 days

@jack_meditates2 from a direct ancestral relationship. I also should caveat that other saka groups share some IBD with later Turks. But the amount of IBD connecting Turks and Bulan Koby is very large and very unique. I also find the qpAdm and ADMIXTURE results convincing 2/3

1

0

9

Daniel Tabin

@DanTabin

19 days

@xiang_zhua93543 Via Bulan Koby, yes

0

7

Daniel Tabin

@DanTabin

20 days

@benevetsenhayir Not until they are published unfortunately

1

0

8

Daniel Tabin

@DanTabin

6 months

@Beyoku @iosif_lazaridis It is common for ancient samples (you can see this in a number of the oldest Americans) to appear admixed. This is because they have had less time to "drift" from the ancestral population and thus are closer to the original allele frequencies 1/n

2

1

7

Daniel Tabin

@DanTabin

21 days

@jack_meditates2 It did and I personally believe they are! The strongest evidence of this is the IBD which shows direct relationships between Bulan Koby and Turks. I should caveat and mention that IBD between samples could be from a shared ancestor (ie samples are cousins) rather than 1/n

1

0

7

Daniel Tabin

@DanTabin

10 months

@CPlusPlusOpenG1 @iosif_lazaridis Hello, this is an important question. I have a few points. A) This model is not well optimized for Tajiks (p=0.02, and I designed it for Turks and Mongols). B) Kura Araxes and Sogdians are genetically quite similar, as both have most of their ancestry coming from (1/2)

2

0

6

Daniel Tabin

@DanTabin

21 days

@jack_meditates2 Yes it is as both are very "Siberian" (Mongolia Neolithic and Paleosiberian) Which group are you referring to? Azelino or Bulan Koby? I'm not sure I understand the second part of your quesiton

1

0

5

Daniel Tabin

@DanTabin

21 days

@based_dardanian Based on the evidence that Lir Turks and Shaz Turks share Xiongnu ancestry, I'd say that proto-Turkic was spoken by the Xiongnu (or at least within the Xiongnu). I am not (yet) convinced by any external relationship (such as the Altaic hypothesis).

1

0

4

Daniel Tabin

@DanTabin

10 months

@CPlusPlusOpenG1 @iosif_lazaridis CHG / Iran_N like sources; the main difference between them is Sogdians have Indo-European and some Siberian ancestry, while Kura Araxes have more Near-Eastern type ancestry. C) Kura Araxes + Bulan Koby is somewhat similar to Sogdian + Middle Eastern + Siberian / Bulan Koby.

1

0

3

Daniel Tabin

@DanTabin

11 months

@igp_rostamd @iosif_lazaridis No, these are unpublished. There is another separate study by @yc_ozdemir also looking at other Bulan Koby samples, but neither of our works are published yet

1

0

4

Daniel Tabin

@DanTabin

19 days

@JLingPystynen It’s a Cisbaikal rich source

1

0

4

Daniel Tabin

@DanTabin

21 days

@jack_meditates2 Take a look at my poster and tell me what you think!

1

0

4

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva Because Ranis doesn't share any drift with specific later Eurasians, but does share drift with Eurasians as a whole (specifically the drift all Eurasians share with one another), it will be placed roughly in the middle of the Eurasian part of the PCA. 2/n

2

0

3

Daniel Tabin

@DanTabin

11 days

@nrken19 Yes, its from this paper:

1

0

3

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva Because you would be comparing f3s, you also could compute f4(Ranis, Yoruba; Eurasian1, Eurasian2), which would give you whether Eurasian1 or Eurasian2 is closer to Ranis (positive->Eurasian1 and negative->Eurasian2) I hope this is useful! 12/12

1

0

3

Daniel Tabin

@DanTabin

22 days

Ali also developed this awesome website: Use it to look at any SNP you'd like and see how selection affected it in West Eurasia over the past 10,000 years!

0

1

3

Daniel Tabin

@DanTabin

1 year

If anybody is interested in working with David Reich (or another Harvard HEB faculty member) as a PhD student, and wants more information, you should attend this:

2024 HEB Graduate Program Open House | Department of Human Evolutionary Biology

heb.fas.harvard.edu

0

3

4

Daniel Tabin

@DanTabin

11 months

@321Hpc @Orhan57323150 1) In an allele frequency based model, BK are similar to other Saka, so Tajik BK ancestry is likely somewhat explained by that. 2) Tajiks can be modeled without Xiongnu ancestry, another group involved in the formation of Turks 3) Tajiks likely have some Turkic ancestry.

2

0

4

Daniel Tabin

@DanTabin

20 days

@oglu_ural49842 Please email me or reach out!

1

0

3

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva ANE shares drift with both as it is ancestral to both! Many ANE samples are also very old. Thus these groups (which are LESS admixed than Native Americans or Europeans) show up between them, as would a population with the correct mixture of Native American and European ancestry

1

0

2

Daniel Tabin

@DanTabin

10 months

@bibliophiliste I modeled them so that a single model could be applied to all of my targets. I used the same rotating sources for them as my other populations. I don't think my model of Xiongnu disagrees with the one you posted. Also, I agree Xiongnu nearly the same genetically as Xianbei

1

0

2

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva Perhaps a a stronger example @hhhhhhbbhhbbbb : ANE groups appear between Native Americans and Europeans / West Eurasians. PCA implies this is because they are mixed West Eurasian and Nativa American but in fact the opposite is true!

1

0

2

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva In G25, it appears that the closest groups to where this arbitrary center lies are South Asians, but you can see that a few populations who have East and West Eurasian Admixture (Maori) and groups with some African Admixture (Iranian Bandari Zanji) also end up somewhat nearby 3/n

1

0

2

Daniel Tabin

@DanTabin

6 months

@Beyoku @iosif_lazaridis Which are roughly equal to an average of all of the descendant groups (this is because at each SNP, the drift is independent and random in each group, and has mean zero with some variance). 2/n

1

0

3

Daniel Tabin

@DanTabin

19 days

@Metal_Terium Archeological uncertainty basically. But Archaeologists might be able to resolve some of the individuals

0

3

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva A final note: Yana is known to (despite being more East Asian as Native Americans are) be less ancestral to Native Americans than Yana is. The ANE ancestral to Native Americans is closer to MA1. Nicely we see MA1 being closer to the Native American pole than Yana.

1

0

2

Daniel Tabin

@DanTabin

6 months

@Beyoku @iosif_lazaridis Thus, for ancient populations ADMIXTURE and similar methods often show them as mixtures of many groups. You will notice other ancient groups in that plot show the same patterns of being highly mixed including with the "European" "East Asian" and "African" components 3/n

2

0

3

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva Typo correction: Yana is less ancestral to Native Americans than MA1 is

0

2

Daniel Tabin

@DanTabin

8 months

@nrken19 Slightly slowed down as I am teaching this semester, but otherwise good :) I'm working on a more formal write up that I will probably pre-print. Hopefully it will be ready somewhat soon, but no promises on a date.

1

0

3

Daniel Tabin

@DanTabin

6 months

@Beyoku @iosif_lazaridis This is not a phenomenon unique to this paper, and it also shows up in PCA as well as ADMIXTURE (older samples often appear closer to the center of the plot, even when unadmixed). I hope this is helpful :)

0

3

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva In either case though, ADMIXTURE and PCA based methods (like G25) are unlikely to give you meaningful results for Ranis / Zlaty Kun. G25 is a great tool which can give interesting information, but it simply isn't the tool for the job when it comes to Ranis. 10/n

1

0

2

Daniel Tabin

@DanTabin

13 days

@PhilistiaForeva I’m glad you found it useful :)

0

2

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva In methodologies like PCA (and also ADMIXTURE) highly admixed populations often look similar to extremely ancient populations (ie UP and IUP smaples). This is why in ADMIXTURE many ancient groups have small amounts of ancestry from many different components. 4/n

1

0

1

Daniel Tabin

@DanTabin

11 months

@dogangurpinar2 Also, I absolutely agree that the East Eurasian ancestry of Turkic and Mongolic speaking groups differs. Not only do Mongolic groups have more Yellow-River ancestry, but Bulan Koby themselves have East-Eurasian ancestry (which is related to Uralic ancestry as both are APS-rich)

1

0

2

Daniel Tabin

@DanTabin

6 months

@MollyWop9 @Beyoku @iosif_lazaridis Look at USR1. It is also Native American and shares the main bottle neck but also has non-Native American components. These are early Native Americans. I think the "African" in the Brazilian sample is Population Y affinity, which I believe this sample is known to have

1

0

2

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva I'll note here (for the Hoabinhian+Iranian+Mota model) that Mota is, while non Eurasian, closer to Eurasians than to other African groups. Because of this (and the fact that backward-in-time-drift is also martingale) Mota also works as a source that can average out to "p*" 8/n

2

0

1

Daniel Tabin

@DanTabin

10 months

@321Hpc @Orhan57323150 Sorry for the late reply, I haven't been on twitter for a while. The Bulan Koby samples are on the upper Ob River in the Altais. We have no evidence of their language, but they are culturally and genetically similar to Saka, so they likely spoke a Saka dialect at some point

0

1

Daniel Tabin

@DanTabin

6 months

@MollyWop9 @Beyoku @iosif_lazaridis BTW, I mentioned this elsewhere but felt worth mentioning here: this sample is published and I haven’t seen any other analyses showing a real African affinity. Also, there isn’t a clear population that would absorb the population Y type drift.

0

2

Daniel Tabin

@DanTabin

20 days

@BulgarChanyu Interesting idea! What makes you say this?

1

0

2

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva Another intuition for this: Each population has allele frequencies drifting away from the ancestral one (which I will call p*). Each neutral SNP should move randomly up and down in frequency (due to drift), but given that moving up OR down is equally likely, the mean is p* 5/n

1

0

1

Daniel Tabin

@DanTabin

11 months

@chronicfutility I have looked at this. It is a fantastic study, and the lead author (TC), who is extremely talented, is a collaborator of mine. Thanks again for your input!

0

1

Daniel Tabin

@DanTabin

11 months

@dogangurpinar2 I also never claimed the burial traditions were totally continuous, but there is clear connection and continuity. It is neither identical, nor fully discontinuous

1

0

2

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva A better test would be computing f3(Ranis, Eurasian; Yoruba), which calculates how much drift Ranis and the Eurasian shared with one another compared to Yoruba. The Eurasian that maximizes that value would actually be the closest living population to Ranis. 11/n

1

0

1

Daniel Tabin

@DanTabin

11 months

@dogangurpinar2 I am not claiming that Bulan Koby is proto-Turkic, but rather that it is an important source in later Turkic groups. I fully agree that there is other ancestry. Turks clearly have near-ubiquitous Xiongnu ancestry in my model. Bulan Koby simply differentiates them from others

1

0

1

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva Its true that there’s an unresolved relationship between Goyet 116 and Tianyuan, but it seems unique to Goyet 116 and there’s no clear relationship between Tianyuan and other Aurignacians. What are your thoughts on my ANE example? I hope it will be very elucidating for you

1

0

1

Daniel Tabin

@DanTabin

11 months

@TheScythman I never suggested Bulan Koby were the "original" Turkic speakers. I instead am suggesting they were an important population in the formation and spread of Turkic languages. These are different things. I appreciate the critique and I should be more clear.

0

2

Daniel Tabin

@DanTabin

11 months

@dogangurpinar2 Finally, I actually feel that there is a lot of continuity between Bulan Koby and later Turkic burials just archeologically. This is what first drew us to these samples. See

0

1

Daniel Tabin

@DanTabin

11 days

@nrken19

Genomic analyses of hair from Ludwig van Beethoven

Begg et al. perform genomic analyses of eight locks of hair attributed to composer Ludwig van Beethoven. They sequence a high-coverage genome, finding strong genetic risk for liver disease that may...

www.cell.com

0

1

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva I want to state that it IS POSSIBLE that Ranis and Zlaty Kun share more ancestry with present day South / Southeast Asians than other groups (although this would be a novel finding). Bacho Kiro clearly shares more with later East Asians than later Europeans. 9/n

1

0

1

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva Thus for each descendant population 1, 2, 3, ... the allele frequency at the SNP in question (dubbed p1, p2, p3) can roughly be drawn from a normal distribution with mean p* and some variance term (representing drift). If frequency reaches 0 or 1 it becomes fixed. 6/n

1

0

1

Daniel Tabin

@DanTabin

11 months

@Iranic_Genomes I don't think I can give details on the sites on Twitter without first consulting with the archeologists. My apologies. In theory they will be released with a mature version of the above work. I don't have a date as to when that will be.

0

2

Daniel Tabin

@DanTabin

14 days

@PhilistiaForeva From here, it is simple to see that p* can be re-estimated by averaging p1, p2, p3, ..., with a higher number of descendant populations giving a better estimate of p*. Thus ends the other intuition and explains why okay-ish fits can be obtained mixing different sources 7/n

1

0

1

Daniel Tabin

@DanTabin

10 months

@CPlusPlusOpenG1 @iosif_lazaridis Thus, it is more likely that the "true" model for Tajiks has a source of Near Eastern ancestry, and due to my model not including this, it prefers Kura Araxes to Sogdians for the Iran Neolithic like ancestry. This also applies to Uzbeks and Turkmen

0

1

Daniel Tabin

@DanTabin

11 months

@TheScythman There are actually differences between Pazyryk and Bulan Koby, which I can seen in multiple of my analyses, but I agree that there is also a connection between the two groups. I mention it in my poster I believe.

0

1

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva The same applies for the essentially fully West Eurasian Sunghir and Kostenki which also are shifted towards the center of the plot on this image

2

0

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva Salkhit is both mixed and ancient. Look at Tianyuan, who is generally thought to be fully on the East Asian “branch” with essentially no West Eurasian DNA. They too are closer to the middle, as while they share some drift with later East Asians, they also don’t yet have all of it

2

0

Daniel Tabin

@DanTabin

1 year

Postglacial genomes from foragers across Northern Eurasia reveal prehistoric mobility associated...

The North Eurasian forest and forest-steppe zones have sustained millennia of sociocultural connections among northern peoples. We present genome-wide ancient DNA data for 181 individuals from this...

www.biorxiv.org

0

1

Daniel Tabin

@DanTabin

14 days

@hhhhhhbbhhbbbb @PhilistiaForeva As do ancient groups. Which is exactly why they look similar to one another in these methods

0

1

Daniel Tabin

@DanTabin

10 months

@bibliophiliste Thank you for your interest!

0

1

Daniel Tabin

@DanTabin

10 months

@LucreSnooker I can't speak on it yet, but we have already looked into something similar to this idea. It is an interesting one to say the least

0

1

Daniel Tabin

@DanTabin

10 months

@kackinturko There are Mongolic and Turkic speaking Yugurs. The Category is a simplification for the poster. It is good to point out though. I meant it as "living in China with Majority Han-like ancestry"

0

1

Daniel Tabin

@DanTabin

11 months

@TheScythman I am familiar with this paper, and have myself linked it elsewhere in this thread. This paper aligns with my conclusions. Bulan Koby are influential in later Turkic burial practices

0

1

Daniel Tabin

@DanTabin

11 months

@Nafis_theBeast The Sogdian samples are unpublished

0

1

Daniel Tabin

@DanTabin

14 days

@AntiAlles6 No Sakas were most likely spoke Iranic languages. Turks have some Saka ancestry though, namely via Bulan Koby

2

0

1

Daniel Tabin

@DanTabin

11 months

@dogangurpinar2 Again, thank you for this interesting discussion. I will think deeply about this, and look into the Karajakupovo and Krasnoyarsk_BA samples

0

1

Daniel Tabin

@DanTabin

10 months

@CPlusPlusOpenG1 @iosif_lazaridis No, I don't disagree with her opinions there. Thanks for asking!

0

1

Daniel Tabin

@DanTabin

11 months

@konowryyyy_ @iosif_lazaridis I have noticed this as well. I don't yet know the answers. as I mentioned this is a work in progress. It is an interesting question!

0

1

Daniel Tabin

@DanTabin

11 months

@Newlearn35 Sogdians are most similar to Kanju. I haven't calculated a direct measure of genetic difference like an f2 or Fst, but I will probably do this at some point.

1

Daniel Tabin

@DanTabin

11 months

@TheScythman What makes you say this? It should be noted that qpAdm is showing the simplest passing model, not necessarily the truth.

0

1

Daniel Tabin

@DanTabin

6 months

@Beyoku @iosif_lazaridis I see your point, but also don't think this is real African Ancestry. This sample is not new, and other studies haven't seen any such affinity. A more likely explanation is a "population Y" signal, which I think this sample has been shown to have

0

1

Daniel Tabin

@DanTabin

11 months

@chronicfutility Are you basing this on PCA? I will look into this more. Thank you for the feedback

1

0

1