Don2 (Don1 Revised)
Contributor
I was interested enough to click the link, and then click to the research article they link to. I was especially interested in how they derived the "one-in-three-trillion chance."Uh-oh;
Fresh suspicion that Covid may have been tinkered with in a lab emerged today after scientists found genetic material owned by Moderna in the virus's spike protein. They identified a tiny snippet of code that is identical to part of a gene patented by the vaccine maker three years before the pandemic.
It was discovered in SARS-CoV-2's unique furin cleavage site, the part that makes it so good at infecting people and separates it from other coronaviruses. The structure has been one of the focal points of debate about the virus's origin, with some scientists claiming it could not have been acquired naturally. The international team of researchers suggest the virus may have mutated to have a furin cleavage site during experiments on human cells in a lab. They claim there is a one-in-three-trillion chance Moderna's sequence randomly appeared through natural evolution. But there is some debate about whether the match is as rare as the study claims, with other experts describing it as a 'quirky' coincidence rather than a 'smoking gun'.
Daily Mail
I noticed right away that the Daily Mail had mutated a "three in one trillion" chance to the 9 times rarer number, but this was the least of problems in their calculation.
I was afraid that the calculation would involve a lot of genetics particulars I'd be unable to understand or check, but in fact it was a trivial combinatorial calculation. I've attached their Figure 2. Here's a rough transcription in case that image breaks:
P1 = Prob(19-nt seq appears in 30,000-nt genome)
= (30000-18)*1/4^19
= 1.09 * 10^-7
P2 = Prob(19-nt seq appears in 3300-nt seq)
= (3300-18)*1/4^19
- 1.19 * 10^-8
P2c = Prob(19-nt seq appears in one of 24712 seq of 3300 each)
(24712*P2*(1-P2)^24711 = 0.00029 [the reddened part is irrelevant, but particularly silly]
P3 = Prob (id seq appears in each
= P1 * P2c = 3.21 * 10^-11
We have others adept at simple probability calculations. Check my work please! I think they committed the fallacy
"What is the probability that Lady Lucille Morningstar is in BOTH databases A and B?" without noting that Morningstar's membership in database A was the only reason she got involved in the question!When I do the calculation my way, instead of 3-in-a-trillion I get a chance much closer to unity!
I feel like conspiracy type things are usually missing a kind of meta-thinking. It's hard to describe, but it's more like, "hey I noticed this real big coincidence. Now, let me go find a p-value for it. Wow, it's significant!" The meta-thinking should look more at how you always find a bunch of coincidences in a far greater context, like let's say life. It's hard to apply sometimes, but let me try: what kind of coincidence would have made someone jump up and down and say, "see look, this proves it!"--and what are the chances due to randomness you find such a coincidence? What is the probability of the thing you wanted to know in this much greater context? So, this 19nt sequence is some kind of oncology sequence from 3300 nt, but there's a ton more sequences from other things patented by Moderna and also Pfizer, J&J, and even govt entities and others that also would have made a conspiracy person jump up and down. Any of those things could have matched by chance and this one is weird because it's from an oncology gene, not something you'd expect to be used for some kind of synthetic whatever in virus editing, especially just 19nt. 19nt is more like the length of a primer or probe design. And wouldn't you expect an insert to be a multiple of 3nt or it'd change the open reading frame? Anyway, 19nt is also the match but it's also somewhat arbitrary because some of the nucleotides surround the area of interest and if you look at 30nt, it fails equivalency, or if you look at some other thing you might find that 21nt matches some other thing some other company patented but more of the nucleotides are on one side or you might find 16nt (less) match some other thing. In this latter case, I think anything 12nt or greater matching the 12nt section and any nt on the sides would have been cause for someone to jump up and down.
ETA: I am still looking at the nucleotide sequences as I need to clear something up there...will post later as well.
Last edited: