My, admittedly limited, understanding is that the immuno response to a natural infection is "broader" and targets more of the viral proteins ("we show that multiple neutralizing epitopes, within and outside the receptor binding domain (RBD), are variably targeted by human polyclonal antibodies") while the vaccine only introduces a single or small number of proteins for the immune system to respond to.

If that is the case then we would want vaccines that include more mRNA instructions for creating more unique viral proteins to expand the number our immune system has been exposed to. I have no idea how challenging that is.

> I have no idea how challenging that is.

Technically it's easy, Moderna already has multi target vaccines (for other diseases), one with 7 different targets. It can be as easy as literally creating 7 different vaccine liquids and then mixing them all in a single vial.

The problem is deciding what to put in, what ratio, possible side effects, etc... Lots of combinations which can mean lots of expensive and slow clinical trials.

That's roughly my understanding of this as well, pulling from whatever I remember from old immunology classes.

The human adaptive immune response is massively parallel, and you can liken it to multiple progressive selection rounds for antibody candidates. As such, it is intuitive that this process would create a set of antibodies that targets not just a single antigen in multiple ways, but multiple antigens in multiple ways, resulting in more robust response (see polyclonal b-cell response)

As for the latter, the issue, as already stated, is in understanding the side effects and efficacy profiles of said vaccine. Clinical trials are long and expensive processes.

> should prove exceptionally difficult for the virus to overcome through evolutionary pressure

Which is why we need gain of function experiments to accelerate this process and get ahead of the virus so that we are prepared for the eventual successful mutation.

Gain of function experiments are not all the same. If you use an artificial selection approach (wherein selective pressure is applied to a huge library of random variations), then maybe.

If you have labs picking-and-choosing what mutations to make (what usually happens), then no. That's just hubris. People stumble around in the dark, occasionally find a truffle, and pat themselves on the back for having such a great truffle-finding method.

Even for good artificial selection systems, mutational space is so gigantic that it's hard to cover properly: not just the point mutations (i.e. converting one amino acid to another) but also insertions, deletions and transpositions. There's no artificial selection system I'm aware of that can recapitulate mutational space.

Which is why we need "intelligent design" to "help" the virus jump over this combinatorial problem.

For example we can try spikes from different related viruses, or pick and choose various other tricks, like furin cleavage sites, for example like in this rejected 2018 EcoHealth coronavirus research DARPA grant proposal:

https://twitter.com/JamieMetzl/status/1439989291858513929

I’m not an expert for sure though my read is the most expedient way out of this situation is plasma donation / transplantation.

The issue with plasma donation is that the plasma contains antibodies which are temporary (6ish months). What we want is to elicit a response which produces memory B cells so the vaccinated individual can create their own antibodies. That can only happen through vaccination or infection.