3 ms·

I keep re-reading what you wrote, and I think it's less far fetched than you imply, for the following reasons. Huge datasets labelled as to ancestry exist, cha

I keep re-reading what you wrote, and I think it's less far fetched than you imply, for the following reasons. Huge datasets labelled as to ancestry exist, characterizing distributions of alleles that imply "race", which is how Ancestry and 23andme, for American company examples, do some of their analyses. At least two Nobel prizes were recently awarded, one to Doudna and Charpentier related to Crispr-CAS9 for context-sensitive genome edits, and another for mRNA (which was in development for a decade or so before the entire Covid repurposing arose). So, I think the part missing, to your point, is the "vector" that would deliver a sequence of patch files, perhaps via CRISPR-cas9 etc, which would perhaps bind to the constellation of race-implication genome sites to collectively build some proteins of whatever sort of interest, whether for fitness repair or fitness reduction in the host.

dekhn 3y ago
I said what I said above based on 30+ years of experience in this field. Be aware that the entire field of gene therapy was paused for 20 years because a single patient died in a single study in the 90s. Race is a proxy variable for genetic history and we have enough genetic history to identify what "race" people belong to, in a computational way, with a great deal of human judgement. Implementing that "logic" in the form of a gene therapy delivered as a virus is much harder. Molecular logic, these days, is fairly simple and doesn't allow us the level of accuracy you would need. If you don't require 100% sensitivity and specificity the job is "easier". For example, what should your theoretical virus do to people who are children of two parents with very different genetic history? Which "race" do they belong to? And if you go down that path, you're basically going to have to convince your funders (because this would be an extremely expensive research project) that your weapon of mass destruction will likely kill some fraction of allies. Those are just a couple things off the top of my head; I could write a whole book on how hard delivering gene therapy is.
- jjtheblunt 3y ago
  yep, totally understand and agree it's a super vague notion, with super imprecise possible consequences.

dekhn 3y ago

I said what I said above based on 30+ years of experience in this field. Be aware that the entire field of gene therapy was paused for 20 years because a single patient died in a single study in the 90s. Race is a proxy variable for genetic history and we have enough genetic history to identify what "race" people belong to, in a computational way, with a great deal of human judgement. Implementing that "logic" in the form of a gene therapy delivered as a virus is much harder. Molecular logic, these days, is fairly simple and doesn't allow us the level of accuracy you would need. If you don't require 100% sensitivity and specificity the job is "easier". For example, what should your theoretical virus do to people who are children of two parents with very different genetic history? Which "race" do they belong to? And if you go down that path, you're basically going to have to convince your funders (because this would be an extremely expensive research project) that your weapon of mass destruction will likely kill some fraction of allies. Those are just a couple things off the top of my head; I could write a whole book on how hard delivering gene therapy is.

jjtheblunt 3y ago
yep, totally understand and agree it's a super vague notion, with super imprecise possible consequences.