To provide additional protection (especially since given that timing it almost certainly wasn't the delta variant that you got). Maybe you are protected by your previous infection and maybe you aren't.
It was my understanding that you basically just punch these mRNA sequences into a computer and you could have a new formulation ready for manufacture over the weekend. If it's a regulatory issue, maybe the regulations should be adjusted. Why have a booster of the original stuff that's not working as well?
It's really not that simple. I'm not current, but I studied immunology as part of medical school a long time ago.
The Delta variant replicates really quickly, that is its evolutionary advantage over other strains, it's not that your immune system is weaker to it. What happens is that at first it ramps up a lot faster than your immune response, and during this period, it produce a lot more virions, and your immune system has a bigger fight on its hands.
There are many immunologists working on identifying the nature of the immune response to sars-cov-2, and so far, something on the order of 80 unique antibodies have been identified which are induced by vaccination, and about 200 unique antibodies which are induced by actual infection. The convalescent antibodies target the nucleocapsid (the ball) more than the spike protein.
It appears that the delta variant escapes approximately 1/8 of the antibodies in both cases, so you still have a very strong immune response, just one which doesn't ramp up as fast as the virus.
Now, as for the mRNA sequences, we can create them in a DNA printer (then transcribe to RNA), however, this triggers new safety and efficacy trials at the FDA.
Are you willing to bet that the tweaked Moderna vaccine (the trials of it was announced, right?) shows worse efficacy than the original? Odds 51:49 in your favor, null hypothesis of the same efficacy is a tie. That's generous I think, if nobody knows.
Most proteins are safe, of course, and the odds that any particular new mRNA target is safe is pretty good. But it's not trivial to predict with computational models, and so it requires testing.
