Yes. In the immune system, confining somatic mutations to a narrow part of the genome that is relevant to antibodies (the VDJ region) dramatically increases the likelihood of a beneficial mutation relative to mutating in an untargeted way across the whole genome. Would you like to work out the math?
I donât think math regarding a specific location relative to the entire genome would be relevant. What would be relevant would be the ratio of adaptive to nonadaptive mutations in those locations.
It is well known that many types of mutations are not at all random with respect to location.
1 Like
That is the point. We are not looking within those locations alone, but across the whole genome. A whole lot of unhelpful mutations are avoided by targeting mutations in particular area.
Exactly, and in this case it biases things towards beneficial mutations.
2 Likes
Thank to you and @mercer for this helpful and informative exchange. I have a question: is there is a causal relation between the external stimulus and the nature of the mutation?
-
Does the location of the mutation vary in a way which depends on the nature of the stimulus.
-
Does the nature of the mutation within a location vary with the nature of the external stimulus?
-
Does the rate of mutation vary with the nature of the stimulus â that is, do some type of stimuli increase the rate more than others in a predictable way?
I see 3 â the rate of change impact â as qualitatively different from 1 and 2. For 1 and 2 are claims about each individual mutation, whereas 3 is a claim about the rate of mutation and not the nature of each individual mutation.
I do recognize that if all three are possible, then there is the further complication of modeling how and if they occur together.
ETA: #1 for clarity,.
But apparently, some very unhelpful mutations are caused by imperfect targeting, contributing to lymphomagenesis.
I donât know how one would count the actual numbers of mutations, since AFAIK we only see them after filtering by selectionâin one direction that is good for the organism (reactive B cells) and in one that is bad (lymphomas).
1 Like
Perhaps this is what @Perry_Marshall is referring to when he says evolution is not random. He seems to believe that evolution is directed, so there is a reduction in uncertainty.
1 Like
Yes this is exactly what he means, and this is an abuse of terminology.
2 Likes
@BruceS, your suppositions are not entirely precise enough. Doing my best to fill in the gaps, I think the answers areâŚ
Yes.
Yes.
Yes.
1 Like
Working out the math would make sense of this. T
here is an approx. million-fold increase in ratio of beneficial to non-beneficial mutations. Iâm gonna have to look up how large the region of somatic hypermutation is, but letâs suppose for a moment that it is 1,000 bp. The whole human genome is 3 billion bp. Let us say that there is x beneficial mutations in this region (we donât know how many).
So, the chances of getting a beneficial change to the antibody if we are mutating across genome is proportional x/3billion, in the hyper-variable region, the chances are x/1,000. The ratio of these two chances is 3 billion / 1,00, or 3 million. We increase the chance of a beneficial mutation by about 3 million fold by limiting mutations to the hyper variable region.
And yes, these are still ârandomâ mutations.
1 Like
Somewhere far above I mentioned censoring, and this sort of filtering is exactly what I meant.
1 Like
But youâre ignoring that the hypermutation is selective, not completely limited to this location. And youâre only considering harmful mutations to antibody secretion, not ones that cause lymphomas.
And youâre still ignoring the effects of selection.
1 Like
I think you are just computing a different entity than I am. Iâm computing the pre-selection likelihood of a beneficial mutation per mutation.
I donât think your estimation is valid. Where did you find the numbers for beneficial vs. harmful mutations PRE-selection? Where could anyone?
I must say that I am surprised that you are ignoring the clinical data given your medical training. How did you factor in in the fact that lymphoma is likely far more harmful than additional variance in antibodies is beneficial?
And that somatic hypermutation also yields an increased frequency of self-reactive antibodies?
I just donât see that you (or anyone else) have the data that would allow you to calculate anything definitive.
Once again we are just computing different numbers. Iâm not sure precisely your point. The fact that some types of negative mutations are more likely does not change that the positive mitigationâs are more likely. Im making a restricted point.
My educated guess is that the observed rates are what is estimated. These will be slightly biased against negative mutations due to selection.
1 Like
No, Iâm not computing anything.
My point is that there are no numbers on which either of us can base any computation.
One thing that we do know is that the hypermutation is merely selective as to location in the genome, not simply specific to antibody genes.
Arenât you claiming to be computing the ratio of positive to negative?
1 Like
Or extremely biased. We donât know.
From the view of a molecular biologist, the main point is that the cellular mechanisms producing the mutations donât âknowâ which mutations are helpful and which are neutral or harmful. Rather, bases are ârandomlyâ changed in a specific genetic locus, and the beneficial mutations are selected for through other independent mechanisms. For example, when you are given a tetanus vaccine the processes mutating the tetanus toxin specific antibodies have no idea which amino acid changes will improve avidity and specificity. To use another example, itâs a bit like randomly dropping depth charges in an area where you suspect a submarine may be.
2 Likes
I agree with everything you wrote but this. The data indicate that hypermutation is selective, not specific.
1 Like
What do you mean by that?
1 Like