Racial Composition and History of India

December 10, 2012

DNA evidence confirms what historians, linguists and anthropologists have long known but nationalists have denied: that Indians are mainly a mix of indigenous Australoids and intrusive Caucasoids. They're composed of two genetic components, one related to Andaman Islanders and the other to Western Eurasians, which is higher in upper castes. The estimated dates of admixture between the two are consistent with the introduction of Indo-Aryan languages from the northwest and probably also earlier events related to the spread of Dravidian languages and even agriculture.

India has been underrepresented in genome-wide surveys of human variation. We analyse 25 diverse groups in India to provide strong evidence for two ancient populations, genetically divergent, that are ancestral to most Indians today. One, the "Ancestral North Indians" (ANI), is genetically close to Middle Easterners, Central Asians, and Europeans, whereas the other, the "Ancestral South Indians" (ASI), is as distinct from ANI and East Asians as they are from each other. By introducing methods that can estimate ancestry without accurate ancestral populations, we show that ANI ancestry ranges from 39-71% in most Indian groups, and is higher in traditionally upper caste and Indo-European speakers. Groups with only ASI ancestry may no longer exist in mainland India. However, the Andamanese are an ASI-related group without ANI ancestry, showing that the peopling of the islands must have occurred before ANI-ASI gene flow on the mainland. Allele frequency differences between groups in India are larger than in Europe, reflecting strong founder effects whose signatures have been maintained for thousands of years owing to endogamy. We therefore predict that there will be an excess of recessive diseases in India, which should be possible to screen and map genetically.



Reich et al. "Reconstructing Indian Population History". Nature, 2009.


Metspalu et al. "Shared and Unique Components of Human Population Structure and Genome-Wide Signals of Positive Selection in South Asia". Am J Hum Genet, 2011.

Linguistic and genetic studies have shown that most Indian groups have ancestry from two genetically divergent populations, Ancestral North Indians (ANI) and Ancestral South Indians (ASI). However, the date of mixture still remains unknown. We analyze genome-wide data from about 60 South Asian groups using a newly developed method that utilizes information related to admixture linkage disequilibrium to estimate mixture dates. Our analyses suggest that major ANI-ASI mixture occurred in the ancestors of both northern and southern Indians 1,200-3,500 years ago, overlapping the time when Indo-European languages first began to be spoken in the subcontinent. These results suggest that this formative period of Indian history was accompanied by mixtures between two highly diverged populations, although our results do not rule out other, older ANI-ASI admixture events. A cultural shift subsequently led to widespread endogamy, which decreased the rate of additional population mixtures.

Moorjani et al. "Estimating a date of mixture of ancestral South Asian populations", Evolutionary and Population Genetics, 2012.

The paper provides an overview of the spatial and temporal aspects of human morphological variation in India. Four morphological types — Australoids, Negritos, Mongoloids and Caucasoids — have been discerned in the contemporary Indian population. The Australoids appear to be the oldest and have evolved in India. The Caucasoids are physically heterogeneous and suggests incorporation of more than one physical type involving more than one migration. The within-type variance compared to between-type variance for characters studied is smaller. The paper further discusses the observed variability in terms of Indian social organization as well as in terms of endogamy, small numerical strength of the groups and varying ecological conditions prevalent in India.

K.C. Malhotra. "Morphological Composition of the People of India". J Hum Evol, 1978.


Indian Male Composite
Indian Female Composite


Assumed parental groups:


Andamanese Australoid
Iranian Caucasoid


Degrees of admixture:


Austroasiatic speaker
(Juang)
Austroasiatic speaker
(Santhal)


Dravidian speaker
(Paniya)
Dravidian speaker
(Hallaki)


Indo-European speaker
(Meghawal)
Indo-European speaker
(Kashmiri Pandit)

26 comments

Anonymous said...

You're not serious about the pictures choice do you ?

Racial Reality said...

Yeah, what's wrong with them?

Anonymous said...

Agree about the picture choices. ANI and ASI are composites so they can't be represented by individual modern phenotypes. Also, no one has ever mentioned that Iranians are representative of ANI. Moorjnai actually found that Georgians have the lowest z-score and share genetic drift with the ANI. As for ASI, they have indicated no modern population is representative of it since the populations with the highest amounts of ASI have minor non-ASI admixture. However, the Andamanese Islanders are a "distantly" related population to the ASI which separated from them about 30,000 years ago. Using them as representative of the phenotype of ASI is hardly scientific considering how distantly related they are.

Racial Reality said...

This blog post was made before that Moorjani paper came out. But Iranians are a good enough proxy for ANI, just like Andaman Islanders are a good enough proxy for ASI. Look at the genetic plot from Metspalu et al. The South Asian populations cluster right between Iranians and Andaman Islanders. And Moorjani's plot looks the same, with Iranians being very close to Georgians.

Anonymous said...

The science has proved the main Indian population is an admixture of various racial groups(mainly Australoid and Caucasoid with some mixing of Negrito and Mongaloid races) and only differing in percentage in different regions and ethnicities. It has been always the case in India that political and other hidden agendas prevent any unbiased scientific truths being disclosed. Therefore I am mystified by the choice of your photos accentuating the stereotypes and myths of Indo Aryan and Dravidian speakers. You have selected well known Former PM as an example of an Indo European speaker but selected unknown tribals as Dravidian speakers, surely for an unbiased portrayal, you could have posted well known actress Ashwarya Rai as a Dravidian speaker.

Onur said...

Aishwarya Rai is a high-caste person (the Kshatriya caste, i.e., the warrior caste), so she can in no way represent Dravidian speakers in general, the vast majority of whom are members of the low castes.

Anonymous said...

Photos are representing the linguistic groups, not the castes .Caste is common to both major linguistic groups in India.
Saying that Indian people are a mixture of many races whether they are South Indians or North Indians(including the different castes) and with the same breath saying the Dravidian speakers all look Australoid and Negreto and all Indo European speakers look cucasoid shows India has a long way to go. Even the scientists are motivated by different agendas. It was a shame the introduction of caste system put a halt to the mixing of different groups. The practice of endogamy has made the Indian population more insular and less progressive.

Anonymous said...

This is interesting, as I am most commonly mistaken for a MENA person in the US, by Arabs, Persians, and other Middle Easterners. I would say if I grow my beard out I look 100% Persian. I've had Arabic people talk to me in their language, Iranians talk to me in Farsi, etc. In India all my relatives look similar to me, except with (in some cases) skin that is lighter by about three shades, different eye shapes (as is expected in any population) and noses that are sharp but suited to individual faces. In other words, my entire family could pass in the entire MENA region. As far as our caste, we are Indo-European speakers that are settled in North India, originally from Rajasthan, now near the border of Nepal. We are Rajputs/Kshatriyas and have intermarried among members of the Kshatriya caste for centuries. I was wondering why there are blatant similarities between North Indian High Castes and good looking Middle Eastern people- the theory postulated here could account for some of it- but looking at the base populations its hard to imagine such blatant changes over time, resulting in obvious phenotype changes in Indo European/Dravidian speakers and Higher/lower castes. I guess thousands of years is enough time to see such drastic changes. As far as Aishwarya representing the Dravidian community- I suppose that is fair, but only if you keep in mind that Aishwarya is an upper caste Dravidian- and thus has susbstantial ANI ancestry- and it is known that many higher castes from North India settled in the South, which gave rise to the creation of a fourth caste in that region. Therefore, if we look at the native Dravidian population that is not composed of migrants from the North- it is fair to say that Aishwarya would be a poor example. And she is not particularly good looking either- I've seen much better looking females among higher castes in North India.

Anonymous said...

Also, is it fair to say that all Indo European speakers are 100% Caucasian? I mean their physical characteristics are all caucasoid, down to their hair and hirsutism. In addition, the children I have seen of even lower caste Indo European speakers and Europeans come out looking 100% European- and vice versa. Therefore, it is safe to assume that a large proportion of the population in India is Caucasoid- and this is mostly found among the upper and lower caste Indo European Speakers, along with a minority of Dravidian speakers that are upper caste/not lower caste. As I said before, the similarities among MENAs and NI Upper caste individuals like myself are mind boggling. There is no way to distinguish between the two in many instances.

Anonymous said...

It doesn't matter whether the blog post was made before Moorjani's paper came out. Neither Reich or Metspalu mentioned that Iranians are representative for ANI in any of their work. In fact, Reich originally used the NEU (Utahn White Americans of North Euro ancestry) as a proxy for ANI in his original ANI-ASI paper. Moorjani's group improved on that aspect as Georgians were a much closer match. I assume you used Iranians because they share a linguistic connection with South Asia and the affinity they have on PCA plots toward South Asia in comparison to other West Asians. That is simply because they share some ANI ancestry with very minor ASI ancestry. Other than that, they aren't actually a "very good" proxy for ANI since it was much more "Caucasus" like than the admixture of Iranians who have affinities toward Southwest Asia and the Mediterranean too. At least on admixture runs. No to mention that ANI is a composite so even if Georgians a proxy for population genetics, a population representative for ANI in terms of phenotype is simply not possible. As for the the Andaman Islanders, they are a much worse proxy for ASI (especially in terms of phenotype) than Iranians are for ANI. Firstly, they are 30,000 years apart from the ASI and secondly, ASI is a composite. It's highly probably that ASI is made up of diverse indigenous populations in South Eurasia that could have been fairly divergent among themselves. Reich mentioned in his original paper that the Andamanese Islanders were only used as proxy because a purely ASI population no longer existed and the Andamanese Islanders were an ASI-related (albeit distantly) population with no ANI admixture.

As for Metspalu's PCA plot, that plot is only meant to be an intra-South Asian plot. Otherwise, people would actually think that Tajiks and even Uzbeks, Hazara and Uyghurs were more West Eurasian than the HGDP Pashtuns, Baloch, Sindhis, etc. which is not the case at all. In fact, on 23andMe's Global Similarity plot and McDonald's 2 dimension world PCA plot, these Pakistani populations are much closer to Euros and West Asians than those Central Asian populations other than Tajiks who are roughly similar in distance. In addition, the Andamanese Islanders would also be much further on 23andMe's global similarity plot or McDonald's plot.

It's not even about the genetic distance so much but portraying the Andmanese Islanders as a modern day phenotypical representation of the ASI is not exactly very scientific.

Racial Reality said...

You don't know what you're talking about. Metspalu's plot is not "intra-South Asian". It's comparing South Asians to West Eurasians, East Asians and Austro-Melanesians. The Central Asians you mentioned are clearly intermediate between West and East Asians, while South Asians (except for the Tibeto-Burman speakers) are intermediate between West Eurasians and Austro-Melanesians.

And I told you, Iranians are very close to Georgians and a perfectly good proxy for ANI. And Andaman Islanders are also a good proxy for ASI. Do you understand what "proxy" means?

Stop spamming my blog with your nonsense and making me repeat myself endlessly.

Anonymous said...

I never directly said it was an intra-South Asian plot but implied it should be interpreted as such as it is not the most informative World PCA (excluding Africa). It has a North-South Axis with the South Asian populations with more South Asian ancestry further South. Although, West Asians and East Asians are on the plot, the PCA plot doesn’t actually show how West Eurasian each South Asian population is like 23andMe’s Global Similarity or McDonald’s world PCA plot which are built more on a West-East axis for West Eurasian vs. East Eurasian levels of ancestry or North and South for West Eurasian vs. SSA levels of ancestry.
Also, not all Central Asians are clearly intermediate between West and East Asians. Look at where Tajiks and Turkmen cluster versus Baloch, Brahui and Pashtun. On Metsplau’s PCA, Turkmen and Tajiks seem much closer to Iranians and West Eurasia than the Baloch, Brahui and Pashtun which is not the case at all. Turkmen are roughly similar in terms of West Eurasian ancestry as Baloch, Brahui and Makrani (although more Northern shifted) but Tajiks are less West Eurasian than Baloch, Brahui and Pashtun which is what most autosomal admixture studies indicate.
This McDonald PCA is much more informative than that Metsplau’s plot:

http://oi39.tinypic.com/2cxvyfa.jpg
So, what if Iranians are very close to Georgians in terms of distance? They aren’t phenotypic substitutes for one another since there is phenotypical differences between the two on average. They also aren’t the most accurate proxy for ANI which means your photos are misleading. Especially, considering that ANI is a composite like ASI. You can’t use a modern human population as a phenotypical proxy for an ancestral population like ANI or ASI which consist of various differentiated populations.
Also, how are Andaman Islanders a good proxy for ASI? Reich and Moorjani never indicated that. That’s your interpretation of their research. Reich clearly indicated they are the only proxy since they lack non-ASI admixture. 30,000 years apart from a sister population hardly makes a “good” proxy. It makes you a proxy. Nothing more. I understand what proxy means. You don’t seem to understand what “good proxy” means though.
With nonsense? Half of your blog is pseudoscience nonsense. The other half is regurgitating genetic studies with “your” interpretation of them. Any credible geneticist or anthropologist would take it for a bad joke. Although, of course to you, they are just a bunch of politically correct race-deniers.

Anonymous said...

If you run the Admixtools software on Iranians using the same formula that Moorjani et al used to calculate ANI, you end up with a value of approximately 85%. So you can't say Iranians are ancestral to Indians. If you run the same program on Armenians you end up with almost 100%.
Anyways, how can you say that the Onge types are the ASI ancestors. That's like saying that humans evolved from apes.
By the way, the Georgians as B in the F4 formula does not give correct ANI number either because of the fact that Brahmins as well as NW Indians have higher Northern European admixture than Georgians. Not to mention that it won't fully capture the Mal'ta like ancestry.

Racial Reality said...

Tajiks and Turkmens look phenotypically exactly how they plot genetically -- like Caucasus/Iranian people with slight Mongoloid admixture. And according to Reich et al., Pakistanis (Pathans and Sindhis) are still only ~75% ANI, which is why those samples cluster farther away from West Eurasians.

That McDonald plot looks just like the Metspalu and Moorjani plots. South Asians are intermediate between Iranian Caucasoids and Australoids (Papuans and Melanesians this time since there are no Andaman samples). That makes those two groups good proxies for ANI and ASI. Period.

You're just another one of these self-hating South Asian trolls who denies being descended from Aryan invaders and Negrito-like natives because of your weird nationalist/racist fantasies. I've dealt with your kind of nonsense before.

Don't post here again.

Anonymous said...

You don’t know the first thing about Tajiks and Turkmen do you? They can range from looking pseudo-North Euro (especially isolated mountain Tajiks) to very East Eurasian influenced. Autosomally, there is a large range as well with Afghan Tajiks being even more diverse than Yunusbayev Tajikistan Tajiks. The Afghan Tajik samples also seem to be more East Eurasian and South Asian influenced. You would know that if you bothered to follow any of the new studies.

Also, using Moorjani’s formula to calculate ANI or West Eurasian ancestry for Tajiks and Turkmen, you would get 66% and 70% respectively while the HGDP Pashtuns get 70% and Sindhis get 64%. So, while those numbers are merely rough estimates, it proves that the HGDP Tajiks and Turkmen are not really any more West Eurasian than the HGDP Sindhis and Pashtuns and are likely less West Eurasian than the HGDP Baloch, Brahui and Makrani.

The McDonald plot doesn’t look just like Metspalu and Moorjani’s plots. On their plots, there were certain South Indian and tribal groups clustering almost immediately next to the Andamanese. However, Dr. McDonald’s South Indians (mostly tribal and low caste according to him) are quite distant from Melanesians and Papuans and aren’t even on a proper cline with them. When you get to the HGDP Burusho and other HGDP Pakistani populations, the cline drops even more off and if anything, seems to be a cline with East Eurasians rather than any Negrito populations. This makes sense since substituting Dai for Onge gets ANI-ASI estimates very similar to those of Reich and Moorjani.

Also, you seemed to ignore the Anonymous post below mine. Iranians don’t even show up as 100% ANI using Moorjani’s formula but closer to 85% so how could they possibly make a good proxy for ANI? They also have an affinity or pull toward South Asia that is lacking in other West Asians. As for, Papuans and Melanesians making “good” proxies for ASI, that’s even more absurd since they are likely even much more distant from ASI than the Onge (30,000 years). How one could reasonably infer that they make a “good” proxy for ASI because they are simply in between South Asians and Iranians is very presumptuous and not a credible argument. In addition, lumping ANI and ASI into modern day single populations is not only disingenuous but unscientific. Moorjani and Reich were careful in labeling Georgians as representative for ANI as they mentioned that ANI is multi-layered from diverse populations. They only used them as a “genetic” proxy and never alluded to using them as some type of phenotypic proxy that you have taken the liberty to do with various populations on your blog.

As for a self-hating troll, that’s a rich comment coming from someone who thinks his blog has anything credible to offer toward genetics and anthropology. Not to mention, someone who is obsessed with protecting the purity of Southern Italians. By your same reasoning, denying studies that show the affinity of Southern Italians toward the Near East, North Africa and even SSA (albeit minor) makes you a self-hating Southern Italian. Also, don’t forget your efforts to prove that the phenotype differences between North and South Italians are negligible. The fact is they aren’t and Italians form a cline with more Northern Italians having a closer proximity to other Euros (especially Northern and Central Euros) while Southern Italians have closer proximity to Near Easterners and Jews.

Anonymous said...

I should clarify I meant affinity regarding than overall proximity regarding North and South Italians. Regardless, this affinity can be clearly seen on McDonald's PCA plots that include Sicilians, Tuscans or or other Italians.

Racial Reality said...

You're the one who doesn't know anything about Tajiks and Turkmens. Figure 3 of Di Cristofaro et al. (2013) shows them clustering exactly like they do on Metspalu's plot. They're predominantly Caucasoid. Period.

And as for South Asians clining toward Papuans and Melanesians, they do the same thing even more clearly in Figure S1 of Auton et al. (2009). They're a mix of Caucasoids and Australoids/Negritos. Deal with it.

Like South Asia, Europe has two distinct ancestral components varying from north to south -- "Ancient Northern Eurasian" and "Ancient Western Eurasian" (Lipson et al. 2012) -- which, unlike you, I don't pathetically try to deny. The cline in Italy comes from that, and it is indeed genetically and phenotypically negligible.

Your next endless post of denial with no links to any peer-reviewed studies will be deleted. Final warning.

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...
This comment has been removed by a blog administrator.
Racial Reality said...

An endless five-post reply spamming the same nonsense I've already refuted, and still not a single link/quote from a peer-reviewed study. SMH

And btw, we do know what ASI looked like from anthropologists who've studied South Asian crania, including Mesolithic Balangoda Man. All show Australoid/Negrito affinites (follow the links on this page).

Anonymous said...

Spamming the same nonsense? Your blog is nothing but nonsense. SMH…

“For example, the true ancestral populations of India were probably not homogeneous as we assume in our model but instead were likely to have been formed by clusters of related groups that mixed at different times. “

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2842210/

This quote is from Reich’s original ANI-ASI paper. This quote alone makes your individual phenotype proxy assumptions an utter joke. You can’t use an individual phenotype as a proxy for an ancestral population which is not homogeneous.

“However, the Andamanese are an ASI-related group without ANI ancestry.”

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2842210/

The Andamanese are ASI related. They aren’t pure ASI like you’re assuming.

“Genetic evidence indicates that most of the ethno-linguistic groups in India descend from a mixture of two divergent ancestral populations: Ancestral North Indians (ANI) related to West Eurasians (people of Central Asia, the Middle East, the Caucasus and Europe) and Ancestral South Indians (ASI) related (distantly) to indigenous Andaman Islanders.”

http://genetics.med.harvard.edu/reich/Reich_Lab/Welcome_files/2013_AJHG_Priya_India_Date.pdf

When will you get it through your thick skull that that ASI (an ancestral population) is not a homogenous population that Reich’s model assumed? It cannot be represented through the phenotype of a single modern day population. Not to mention that the Onge are only “distantly” related to ASI. Humans are also distantly related to modern Chimps and Gorillas but that doesn’t mean our ancestors looked like Chimps and Gorillas. We simply shared an ancestor just like ASI shared an ancestor with the Onge but only distantly. Hence, the distant relationship.

Dienekes’ work on putting ASI in context simply puts to light what Reich and Moorjani already stated. That no modern population is closely related to ASI.

http://dodecad.blogspot.com/2011/06/ancestral-south-indian-asi-in-context.html

Also, we don’t know what ASI looked like. Quoting outdated anthropologists and studies from 30+ years ago doesn’t prove anything. I can’t believe I’m wasting my time arguing with someone foolish enough to think Coon was actually a credible anthropologist. Modern anthropologists and geneticists laugh or cringe at the mention of his name. Oh wait, they are politically correct race-deniers….

Racial Reality said...

Moron, when you need to represent "not homogenous" populations that don't exist anymore, you have to use a proxy, which is a substitute that will never be 100% perfect. But of all living people, Andaman Islanders are the closest to ASI and therefore the best possible proxy. That's why Reich et al. used them, and that's why I used them too. But really, any Australoid would do because that's what all the evidence says ASI was, and that's what you can't accept. No realistic proxy would be good enough for you.

Your "outdated anthropologists" line just got you permanently banned from here. It proves you're just a denier of evidence you don't like, which comes from being a self-hating racist troll who's butthurt about having Negrito ancestry. Coon is still respected by modern anthropologists, but the sources I cited were mostly Indian anthropologists anyway, and they all agreed about the racial make-up of ancient and modern South Asians, which in turn agrees with genetics.

This discussion is over.

Indian said...

while Indians are a hybrid population of ANI and ASI, ANI are not derived from an outside populations who arrived a couple of thousand years ago according to an unproven linguistics theory. ANI have been in North India for 45000 years. So stop pushing the euro-trash feel good meme, Indians, ANI and ASI are ancestral populations, meaning only the Africans came before us, nobody else, that includes all the special little snowflakes.