Assuming that one located a naturally mild strain of virus (which I personally remain bearish on), for it to produce useful (in the regulatory sense) information a challenge trial, you would need to establish that the response of individuals to the mild strain (which by definition differs from the more lethal strain) is representative of the response to the more lethal strain of virus.
Yes, but that is where you use epidemiology. If you track down each person in the local area infected with the mutant strain and follow their clinical course you can work this out. You can also do serology to make sure the immune response is similar, etc.
There will always be the possibility that the attenuated strain varies is some major but subtle way from the non-attenuated strains, but this is not something ever seen before with other viruses. If you demand to know everything and the answers to all possible risks up front medical science doesn’t advance.
I demand to know enough information to be able to give proper informed consent to study volunteers.
I think your ideas are possible, but the likelihood of success is quite low and the effort required to do the proper studies is far greater than you estimate. Our resources are better spent on much more developed programs with better defined odds of success.
My helpful closing suggestion is to speak with some epidemiologists and clinical researchers to see if you can get some budgetary efforts for what you propose, or failing that, try to locate a review article or two. Good luck.
A bit off subject, but I'm curious if you've ever run across any information that indicates certain length genomic sequences (say 8 nucleotides in a row or so) follow some sort of "order" in their numbering in the entire sequence?
