New genetic evidence for the origins of castes indicates that the upper castes are more European than Asian.

R. RAMACHANDRAN

THE caste-based social hierarchy is deeply entrenched in Indian society even today, but the origins of the system as sociologists and historians now understand, remain an enigma. It certainly goes as far back as the second millennium B.C. when the Aryans, the migrating Indo-Iranian or Indo-European people, entered the country from the northwest and drove southward the proto-Asian and Dravidic speaking populations inhabiting the north. Literary evidence for the stratification of the society, at least in terms of references to the duties of the highest caste, namely the Brahmin, exists in the oldest text of the land, namely the Rig Veda (1500-1200 B.C.). The emergence of the caste system is thus associated with the arrival of the Aryans.

However, many sociologists believe that some kind of a hierarchical social order, in terms of an individual's occupation and duties, was in place perhaps ahead of the arrival of the Aryans. Its evolution into the caste or the varna system as we know today - with the four distinct castes of Brahmin, Kshatriya, Vaisya and Sudra in the order of social standing - probably occurred with the settling of the Aryans who sanctified and legitimised the social order in their own terms which had a distinct religious underpinning. Some sociologists hold that the societal stratification in terms of rights and duties of the individual was a creation of the Aryans in their bid to exercise power over the indigenous proto-Asian populations of North India.

An anthropologically pertinent question, therefore, is what really are the origins of the caste Hindu populations of today who make up nearly 80 per cent of India's one billion population. In recent times, with the rise of strident nationalism in the form of "Hindutva" ideology, which rejects the premise that Aryans were outsiders and views them as part of the continuum from the Indus valley civilisation, an unequivocal answer to this may have political implications. While material evidence of ancient history has not been able to resolve this issue, modern population genetics, based on analyses of the variations in the DNA in population sets, has tools to provide a more authoritative answer. Certain inherited genes carry the imprint of this information through the ages.

An international study led by Michale J. Bamshad of the Eccles Institute of Human Genetics of the University of Utah of caste origins has found (the findings have been reported in a recent issue of the journal Genome Research) that members of the upper castes are genetically more similar to Europeans, Western Eurasians to be specific, whereas the lower castes are more similar to Asians. This finding is in tune with the expectations based on historical reasoning and the prevalent views of many social historians. In exercising their superiority over native proto-Asian populations, the Aryans would have appointed themselves to higher rank castes. The 18-member research team includes scientists from the United States, the United Kingdom, India and Estonia. The collaborating Indian scientists were anthropologists Bhaskar Rao, J. Mastan Naidu and B. V. Ravi Prasad from Andhra University, Visakhapatnam, and P. Govinda Reddy from the University of Madras.

There have been genetic studies in the past that tried to answer this question but their results have been equivocal, in the sense that some have found European origins and some Asian origins. According to Partha P. Majumder, a population geneticist with the Indian Statistical Institute (ISI), Kolkata, who has written a commentary on the work in the same issue of the journal, the primary reason for this was the lack of data on a large uniform set of genetic markers from populations of India and central/west Asia. This study, where the researchers have used a battery of genomic markers and DNA sequences spanning three genomic regions, is a landmark, says Majumder. "The study provides an incisive genomic view of castes and their origins," he has written.

"It is conceivable that the Aryan contact should have been progressively lower as one descended the varna ladder. The genetic expectation, therefore, is that the proportions of those genes (or genomic features) that 'characterised' the Aryan speakers should progressively decline from the highest varna to the lowest and a reverse trend should be observed with respect to those genes that 'characterised' the indigenous Indians," Majumder says.

The three different genomic regions the study has looked at include two gender-specific genes and one biparentally inherited gene. Mitochondrial DNA (mtDNA), the DNA contained in mitochondria which are tiny organelles in each cell that generates the energy required by the cell, is exclusively derived from the mother. Similarly, the Y-chromosome, which defines the male gender in mammals, is passed on exclusively by the father.

Interestingly, an analysis of the genetic variations in the markers associated with the maternally inherited mtDNA and paternally inherited Y-chromosome show strikingly different trends. Maternally inherited DNA was overall found to be more similar to Asians than to Europeans, though the similarity to Europeans increases as we go up the caste ladder. Paternally inherited DNA, on the other hand, was overall more similar to Europeans than to Asians but, unlike in the case of maternal inheritance, with no significant variation in affinity across the castes. This is intriguing, but there is a plausible explanation. Migrating Eurasian populations are likely to have been mostly males who integrated into the upper castes and took native women. Inter-caste marriage practices, while generally taboo, are occasionally allowed, in which women can marry into an upper caste and move up in the social hierarchy. However, such upward mobility is not permissible for men. The caste labels of men are thus permanent, while women, by means of their limited mobility, cause a gene flow across caste barriers. This is the reason, according to the researchers, for the differing affinities of gender-specific genes among castes to continental populations.

In fact, in a study carried out in 1997, the results of which were published in 1998 in Nature, the same research group had mapped this female gene flow among caste groups in Andhra Pradesh. Analogously, in 1999 Majumder and colleagues examined the genetic impact of this social custom preventing upward mobility of males in the caste hierarchy. They looked at six genetic markers for the male inherited Y-chromosome and found that there was little sharing between castes of the features pertaining to the markers. This phenomenon has been described by Bamshad and company as "modulation of evolutionary forces by social processes" instead of through the normal, purely natural, processes of genetic drift and mutation.

Bamshad and associates examined 40 additional bi-parentally inherited genes as well, which also confirmed the results obtained from mtDNA and Y-chromosome markers that Hindu upper castes are genetically closer to Europeans. They thus conclude that Indian caste Hindus "are more likely to be of proto-Asian origin with West Eurasian admixture resulting in rank related and sex-specific differences in their genetic affinities to Asians and Europeans."

Basically the study carried out three sets of comparisons of genetic variations respectively in the mtDNA, the Y-chromosome and the 40 specific autosomal (of chromosomes other than the sex chromosomes X and Y) gene sequences in a sample of 265 males, belonging to eight Telugu speaking castes, from Visakhapatnam district. Comparisons were made within this sample and to 400 individuals from tribal and Hindi-speaking populations within the country and 350 Africans, Asians and Europeans.

The eight castes chosen were Niyogi and Vydiki Brahmin, Kshatriya, Vysya, Telega and Turpu Kapu, Yadava, Relli, Madiga and Mala. Significantly, the castes were ranked as 'upper', 'middle' and 'lower' instead of the four-level hierarchy of the traditional varna classification. Such a classification has in recent times apparently become more popular among anthropologists. Brahmins, Kshatriyas and Vysyas were grouped as 'upper' caste, Kapu and Yadavas as 'middle' caste and the remaining three as 'lower'. "In studies pertaining to origins of castes, one is liable to draw incorrect inferences by including castes belonging to different varnas in the same ranked cluster," points out Majumder.

For the extraction of DNA from the sampled population, after obtaining informed consent, about 8 ml of whole blood or five plucked scalp hairs were collected from each participant. The DNA extraction and its amplification by the Polymerase Chain Reaction (PCR) technique was carried out at Andhra University by Indian scientists. To conform to the ethical guidelines of research with the human genome, approvals and clearances were obtained from Andhra University and the Government of India, according to the authors of the paper. (These DNA samples are being maintained by Andhra University where a laboratory has been set up to carry out such analyses.) Analysis of "genetic distances" - a measure of genetic similarity and affinity - of markers of mtDNA, the maternally inherited DNA, between caste populations and continental populations shows that, irrespective of caste rank, each caste group is most closely related to Asians and is most dissimilar from Africans. And as one moves from the lower castes to the upper castes, the genetic distances to Asians increases, suggesting that Indian populations are predominantly proto-Asian but with affinities to West Eurasian genes. The West Eurasian admixture is, however, proportional to the ranking among castes. Analysis of a special set of mtDNA markers (called haplotypes), whose loci in the genome are closely linked and which tend to get inherited together, also showed that the West Eurasian admixture amounted to 20-30 per cent of mtDNA haplotypes.

Similar "genetic distance" analysis using the paternally inherited Y-chromosome presented, as indicated earlier, a distinctly different pattern of population relationships among castes and among castes and continental populations. In contrast to the mtDNA distances, Y-chromosome data do not suggest a closer affinity to Asians. The upper castes are more similar to Europeans than to Asians, the middle castes are equidistant from the two groups and the lower castes are most similar to Asians. The genetic distances between caste populations and Africans increase as one moves from lower to upper caste groups.

Looking at the variations in a particular special set of Y-chromosome markers, the study disaggregates the European population into Northern, Southern and Eastern Europeans. The analysis of genetic distances shows that each caste is most closely related to Eastern Europeans. Moreover, the genetic distance between Eastern Europeans and upper castes is half the distance between the middle or lower castes and the Eastern Europeans. The authors interpret this as the Indian Y chromosomes, particularly upper caste Y-chromosomes, being more similar to European than to Asian Y-chromosomes. One limitation of the study is the restricted geographical region, namely a single district of Andhra Pradesh, from which the sample of caste Hindu populations have been obtained. The likely reason is that of the logistics of achieving rapport with local populations and getting their consent for genetic analysis.

But according to the researchers this also helped in "minimising the confounding effect of geographical differences between populations." Moreover, the sample size of 265 is too small for drawing conclusions about a Hindu caste population of about 800 million. For example, the number of Kshatriyas in one comparison set is as small as 10. The authors do recognise this limitation in their paper and emphasise the need for carrying out similar analysis in other regions of the country. They, however, remark that because of the ubiquity of the caste system, it is reasonable to predict similar patterns in caste populations in other areas. But according to Majumder, replicating the study in other areas is, in fact, imperative before general conclusions about origins of Indian caste populations can be drawn.

"It is not generally realised that the caste society in a sense was a very elastic society and a caste bearing the same name may have very different origins in different geographical regions," he points out. According to him there are examples when a tribe dispersed over a large geographical region took up different occupations in different sub-regions and fitted itself into the caste hierarchy on different rungs. Different Brahmin castes of Maharashtra, for example, probably had different origins, he says. "Thus, the origin of caste populations may not be uniform over the entire country," adds Majumder. It is also reasonable to assume that northern societies are more likely to reflect more truly the real origins of caste than societies down south where Dravidic features are likely to be reflected in the genetics of the populations. Also, several social forces may have interfered to result in the stratification as is evident today.