Background

Human genetic diversity observed in Indian subcontinent is second only
to that of Africa. This implies an early settlement and demographic
growth soon after the first 'Out-of-Africa' dispersal of anatomically
modern humans in Late Pleistocene. In contrast to this perspective,
linguistic diversity in India has been thought to derive from more
recent population movements and episodes of contact. With the
exception of Dravidian, which origin and relatedness to other language
phyla is obscure, all the language families in India can be linked to
language families spoken in different regions of Eurasia.
Mitochondrial DNA and Y chromosome evidence has supported largely
local evolution of the genetic lineages of the majority of Dravidian
and Indo-European speaking populations, but there is no consensus yet
on the question of whether the Munda (Austro-Asiatic) speaking
populations originated in India or derive from a relatively recent
migration from further East.
Results

Here, we report the analysis of 35 novel complete mtDNA sequences from
India which refine the structure of Indian-specific varieties of
haplogroup R. Detailed analysis of haplogroup R7, coupled with a
survey of ~12,000 mtDNAs from caste and tribal groups over the entire
Indian subcontinent, reveals that one of its more recently derived
branches (R7a1), is particularly frequent among Munda-speaking tribal
groups. This branch is nested within diverse R7 lineages found among
Dravidian and Indo-European speakers of India. We have inferred from
this that a subset of Munda-speaking groups have acquired R7
relatively recently. Furthermore, we find that the distribution of
R7a1 within the Munda-speakers is largely restricted to one of the sub-
branches (Kherwari) of northern Munda languages. This evidence does
not support the hypothesis that the Austro-Asiatic speakers are the
primary source of the R7 variation. Statistical analyses suggest a
significant correlation between genetic variation and geography,
rather than between genes and languages.
Conclusion

Our high-resolution phylogeographic study, involving diverse
linguistic groups in India, suggests that the high frequency of mtDNA
haplogroup R7 among Munda speaking populations of India can be
explained best by gene flow from linguistically different populations
of Indian subcontinent. The conclusion is based on the observation
that among Indo-Europeans, and particularly in Dravidians, the
haplogroup is, despite its lower frequency, phylogenetically more
divergent, while among the Munda speakers only one sub-clade of R7,
i.e. R7a1, can be observed. It is noteworthy that though R7 is
autochthonous to India, and arises from the root of hg R, its
distribution and phylogeography in India is not uniform. This suggests
the more ancient establishment of an autochthonous matrilineal genetic
structure, and that isolation in the Pleistocene, lineage loss through
drift, and endogamy of prehistoric and historic groups have greatly
inhibited genetic homogenization and geographical uniformity.

http://www.biomedcentral.com/1471-2148/8/227/abstract