Not exactly a computational approach but with the continuing fall of the costs of sequencing a human genome there are plans appearing for sequencing the genome of more individuals. The cost is still too high to do it on any wide scale (approx $20million for a complete human genome) but still drastically less than the several billion the human genome project.

It is likely with the current trend in falling prices and the appearance of technology such as lab on a chip devices that we could reach a cost of a few thousand per genome within 10 years. At this point it would be feasible to sequence complete genomes of a large number of people from diverse backgrounds, ethnic groups and so on. If you then link this with detailed information on medical conditions and family history you can start looking for genetic similarities between groups on a genome wide scale. This could lead to important insights into genetic predispositions to illness and so forth.

Not really what you were talking about but has some similarities and sounds like it could be a promising approach to me.

There are of course downsides to this widespread sequencing including some major privacy issues to assess, but also some other positives such as the ability to tailor medical treatments to individuals based on their genetic information.