Check out the grant hosted at:

www.perfectgenome.com

We built this in a few days. Basically, a query engine for whole-genome interpretation. You can submit your own python script to interpret the genomes on our backend - tied to phenotypic/medical information. We have a scheme to keep at patient information private and display measures of predictive power of the script/establish FDA standards/optimize runtime. If enough people submitted scripts for specific disorders, we could build a library for a plug and play system.