New paper in PLOS Genetics, Low Levels of Genetic Divergence across Geographically and Linguistically Diverse Populations from India. Here's the conclusion:
Populations from India, and groups from South Asia more generally, form a genetic cluster, so that individuals placed within this cluster are more genetically similar to each other than to individuals outside the cluster. However, the amount of genetic differentiation among Indian populations is relatively small. The authors conclude that genetic variation in India is distinctive with respect to the rest of the world, but that the level of genetic divergence is smaller in Indians than might be expected for such a geographically and linguistically diverse group.
It's in PLOS, so you can read the whole thing. This figure is pretty illustrative. a) Brown people form a distinct genetic cluster. South Asians that is. This shouldn't surprise. b) South Asians are more related to other South Asians than non-South Asians. Punjabis (Northwestern India) might resemble Iranians and Arabs more than other South Asians, but they are still more like other South Asians than Iranians or Arabs. c) This study showed very little internal population substructure within South Asia. I think the caveats are important, the study looks at American South Asians. This isn't going to be as rich a sample space as all South Asians, there are caste, regional and socioeconomic biases. Within the next 5 years you'll see a paper on South Asia just like this: European Population Substructure: Clustering of Northern and Southern Populations. d) Please be cautious about taking comments like this literally:
The correlations are increased by using a linear combination of allele frequencies with ∼2/3 contribution from Europe/Middle East and ∼1/3 contribution from East Asia. At the same time, however, the separate cluster for India in population structure analysis indicates that allele frequencies in India are distinctive, so that predictions obtained based on European and East Asian groups cannot fully explain allele frequencies in Indian populations.
The "take home" message some get is that this means South Asians are 2/3 group A and 1/3 group B. That's probably not what's going on. It wasn't the case that 10,000 years ago a Ur-European race and an Ur-Asian race got together in India and mated. Rather, South Asia is a crossroads in Eurasia, and it makes sense that the flow of genes would reflect influences from both the west and the east. You notice that populations in Eastern India show the biggest influence from East Asia, and populations from Western India show the biggest influence from West Asia. Geography matters!