There are 4 nucleotides (ATGC) that can be mutated into the remaining 3, so there are 12 possible mutations. If they have an equal change of occurring, you expect to see each mutation 8.33% of the time. But C->T occurs 63% of the time in silent mutations in the variants.
1
2
9
What is going on here? Mutagenesis occurs through 2 different mechanisms: 1. When the RNA genome is copied by RNA-dependent RNA polymerase (RdRp), mistakes are made resulting in mutations. 2. "RNA editing" by the mammalian host cell can change the sequence directly.
3
3
12
So, apparently more than half of the silent mutations in SARS2 variants are produced by an RNA editing enzyme that changes a C to a T. The nice thing about silent mutations is that they are *usually* not selected for or against. So they inform unbiased about mechanisms at work
2
2
11
Because silent mutations aren't biased by selection pressure, they accurately reflect the mechanisms that produce them. Those same mechanisms should then be responsible for functional mutations. Are they?
2
3
2
12
This is where it gets interesting. In Omicron, the functional/defining mutations have too few C->T mutations. And the difference if shocking. They are not produced by APOBEC RNA editing. See picture below. Just look of the density of C>T mutations highlighted in green.
3
11
1
24
Don’t we also need to examine all the NS mutations in the other variants, and see what the C->T ratio is for NS mutations in those? Perhaps those also have a low prevalence of C->T which could simply mean that a C->T error/edit is less likely to produce a NS change.
2
Yes. And I will look at that. But what could possibly the mechanism? Low fidelity replication and RNA editing are both blind to to codon assignments. Neither knows whether they are mutating a silent or functional residue.
1
Also, as you mentioned, selection pressure applies with N-S mutations, but not with S. Could be there are many N-S C->Ts that don't survive? Presumably most of the surviving S C->Ts are in wobble position? It's interesting in itself the attraction away from C e.g. wrt CGG-CGG.
1
1
Is there anything interesting about the distribution of the other possible mutations?
1
Third base C->T is always non-synonymous. That would account for some of the higher survival rate, but why not also T->C?
3
I made a C>G substitution matrix. It shows all the NS mutations you can get with C>G. It is very sparse. Mutations to F and L are highly favored.

Jan 12, 2022 · 10:14 PM UTC