Indian arranged marriages and genes

The debate was triggered when, talking about genes, Ankit mentioned that people with same surnames (implying same lineage) do not usually get married in arranged Indian marriages. He explained the advantages of this practice as the benefits of greater genetic differences between the bride and the groom.

The disadvantage of marrying someone with similar gene pool was explained to me like this: Say you are of gene pool AB, your spouse is also of gene pool AB. Now your child would carry some genes from both parents and discard some. In this case, however, his gene set could be AA, AB, or BB. Compare this with a couple who have AB and CD gene sets. Then their child may have AB, AC, BC, BD, or AD - a larger range of gene pool ensuring the progeny has chances of a diverse set of genes to choose from in the long run. In evolution, this capability would ensure that nature selects the genes that would ensure adaptation to changing environment. In case of the earlier case, where both the parents had AB genes, say the progeny got AA. Then if he marries someone who is also AA, their progeny would miss out on the B set of genes altogether, thereby making them less competitive in the evolution process - weaker candidates in the survival of the fittest.

The above may be very simplistic and even incorrect understanding. However, my concern is about the truth of this hypothesis - Indian arranged marriages allow for better gene diversity and thus greater evolutionary adaptive ability. If you test it by taking two samples, one belonging from India where this tradition of arranging marriages to an individual of different surname is followed to another sample from any other part of the world, is there (or can there) any perceptible difference in the adaptation of human beings to a changed environment? Can we say people in India are better disposed to evolutionary changes than people where marriages are not arranged on the basis of lineage?

A parallel discussion that also went forth was the benefit of marrying between different gene pools is chances of riddance from existing defects (genetic diseases) in a gene pool. If a diabetic marries another diabetic, chances are their children would carry the diabetic genes. However, test this against the example above: In the gene pool AB, say B contributes to diabetes. Marry to another AB, and there is a two in three chance that the child may get a B diabetes carrying gene (AB and BB, the third possibility would be AA). Marrying in the same gene pool seems to have poorer chance of getting rid of the defective B genes (clearly, marrying AB with CD implies that the chances of carrying B genes are two in four assuming that the child picks up at least one gene from both the parents - AC AD BC BD).

(I know somebody even slightly literate in genetics may be laughing all the way over my simplistic understanding of this whole issue!)

However, I wonder if this hypothesis is really true. Surnames are based on a patriarchal system but clearly, genes are not. Genes are contributed from both the mother and the father. So, by arranging marriages based on patriarchal system, the genes from the mother’s side are a free interactor and may dilute the whole theory to a point of random redundancy.