2. An inactivated vaccine will have to be given again and again. A live attenuated vaccine can drive the dangerous strains to extinction. Also attenuated vaccines protect those not vaccinated.

3. In theory you could engineer a virus with deletions, but nobody knows what deletions will work. By using natural deletions you know are mild you get around this problem. It also solves the regulatory issue as you are not going to be able to quickly test a lab made strain.

are there already databases of sequenced strains?

and how would the regulatory aspect work? if the harmless individual is no longer infectious, how would people be infected

2. Yes there are, but we need to look for a quite rare mutant.

3. We would be genome sequencing clinical samples - these would be kept in storage until the data comes back.