I’m not the most knowledgeable person on this, even on LW, but I can sketch out my understanding of the topic at a high level.
The basic idea of a new-strain-robust peptide vaccine is pretty simple. When designing a peptide vaccine, we choose peptides to match against particular segments of COVID proteins. A single-base-pair mutation in COVID’s DNA will produce (at most) a single-amino-acid change in the corresponding protein sequence. Key point: these mutations are local; the rest of the protein remains unchanged, so if our peptide matches part of the protein which did not mutate, then it should usually keep working just fine. (Nonlocal mutations are also a thing, but they tend to change things more radically—most likely they just make the virus not work, and even if it does work it will be a very different virus.)
Now combine that with conserved regions: some regions of the protein play an important functional role, so most changes to their sequence will result in the virus failing to reproduce very much and dying out. Ideally we want to target those regions. (Conversely, other regions play little functional role or a more flexible functional role and are often targeted by the immune system, so mutations in those regions are very likely to be positively selected.)
Note that the specificity of peptide vaccines plays an important role here. Other kinds of vaccines just show the immune system the whole protein, so they don’t control exactly which subsequence the system learns to recognize. Radvac was designed with this in mind.
Zooming out a level… two relevant meta-considerations:
new strains are likely to be less deadly. Partly this is just reversion to the mean, partly it’s evolutionary trade-offs from the virus’ perspective.
the more people get infected, the more new strains will pop up—I expect new strains at a rate roughly proportional to infection count. India is very bad news from that perspective.
Yeah. Bad news at the international level has been causing me to wonder how to set up a fallback plan where “if I see X then I should check Y about the RadVac design and if adequate I should order some peptides”.
My current second best candidate for “X” is to watch Israel (where vaccination is very high) for a spike that is explained by new strains. A better canary would be a place with decent vaccination rates but otherwise a catastrophically poorly run public health service (like in terms of clinics and testing systems and coordinated travel policies and so on)… and basically I think maybe the US is that canary? :-(
Trying to imagine what “Y” would be is harder, like… I just searched for [peptide epitope search tool] and found the immune epitope database which could be relevantly helpful.
> I’m not the most knowledgeable person on this, even on LW...
Coming from you, this seems like extremely high praise for the forum. Care to name names? Also, I wonder if we could improve the attempt to summon them by promising to be thankful for their contributions? ;-)
I’m not the most knowledgeable person on this, even on LW, but I can sketch out my understanding of the topic at a high level.
The basic idea of a new-strain-robust peptide vaccine is pretty simple. When designing a peptide vaccine, we choose peptides to match against particular segments of COVID proteins. A single-base-pair mutation in COVID’s DNA will produce (at most) a single-amino-acid change in the corresponding protein sequence. Key point: these mutations are local; the rest of the protein remains unchanged, so if our peptide matches part of the protein which did not mutate, then it should usually keep working just fine. (Nonlocal mutations are also a thing, but they tend to change things more radically—most likely they just make the virus not work, and even if it does work it will be a very different virus.)
Now combine that with conserved regions: some regions of the protein play an important functional role, so most changes to their sequence will result in the virus failing to reproduce very much and dying out. Ideally we want to target those regions. (Conversely, other regions play little functional role or a more flexible functional role and are often targeted by the immune system, so mutations in those regions are very likely to be positively selected.)
Note that the specificity of peptide vaccines plays an important role here. Other kinds of vaccines just show the immune system the whole protein, so they don’t control exactly which subsequence the system learns to recognize. Radvac was designed with this in mind.
Zooming out a level… two relevant meta-considerations:
new strains are likely to be less deadly. Partly this is just reversion to the mean, partly it’s evolutionary trade-offs from the virus’ perspective.
the more people get infected, the more new strains will pop up—I expect new strains at a rate roughly proportional to infection count. India is very bad news from that perspective.
Yeah. Bad news at the international level has been causing me to wonder how to set up a fallback plan where “if I see X then I should check Y about the RadVac design and if adequate I should order some peptides”.
My current second best candidate for “X” is to watch Israel (where vaccination is very high) for a spike that is explained by new strains. A better canary would be a place with decent vaccination rates but otherwise a catastrophically poorly run public health service (like in terms of clinics and testing systems and coordinated travel policies and so on)… and basically I think maybe the US is that canary? :-(
Trying to imagine what “Y” would be is harder, like… I just searched for [peptide epitope search tool] and found the immune epitope database which could be relevantly helpful.
> I’m not the most knowledgeable person on this, even on LW...
Coming from you, this seems like extremely high praise for the forum. Care to name names? Also, I wonder if we could improve the attempt to summon them by promising to be thankful for their contributions? ;-)