Very interesting project! I can imagine usefulness even beyond the intended use case, namely in joining the information from multiple large genomics studies. The sharing and access of such data becomes increasingly difficult, and by this approach one could increase the effective sample size while keeping the data private.