I don’t think it’s stated explicitly, in which case you’re asking me to quote what they don’t say.
I see you failed to find where they use a different value when varying the functional proportion of the genome.
Here is what you first said : « They seem to assume that the number of deleterious mutations is independent of the fraction of the genome that is functional. This is not possible. »
Now, you admit that they didn’t explicitly state such a thing. So I guess that you’ve found somewhere in the paper at least one passage implicitly conveying such a thing. Can you point the relevant passage? If you cannot, the best thing to do would be to retract your initial statement.
PS: note that if the authors had even implicitly suggested that the number of deleterious mutations is independent of the fraction of the genome that is functional, they would never have passed the peer reviewed barrier.
I would have a hard time finding such a thing since nowhere in the paper do the authors use, manipulate or compute the parameter « number of deleterious mutations » (let’s denote d this parameter).
I am wandering whether you are not confusing d, the number of deleterious mutations, with u, the probability of a deleterious mutation at a functional site. Whereas d obviously depends on the proportion of the genome that is functional, this is not the case of u.
I intended to delete this when I noticed pre-moderation applies. Whether mutations are deleterious, neutral or beneficial is an extrinsic property. Context is all. That’s all I have to say
2 Likes
… Lesecque et al. consider the case nu = 10 … In humans, nu is on the order of 1–10. … nu (the number of de novo deleterious mutations per individual) …
I can find nowhere that says nu would increase with the functional proportion of the genome.
Codswallop. Peer review is not perfect.
My emphasis:
As we have shown, the properties of this distribution depend not just on nu (the number of de novo deleterious mutations per individual) but also on s , the selection coefficient against deleterious mutations.
You didn’t look hard enough.
Testing? I don’t understand. What do you mean?
Unfortunate, since what you did say seems of questionable relevant to @Giltil 's post. You probably need to explain.
I’m pretty sure it matters to both, if all else is equal.
If you distribute 100 mutations randomly in two genomes of equal size, one of which is 10% functional and the other being 90% functional, then the total number of deleterious mutations d, and the probability of a deleterious mutation at a functional site u, will both be higher for the case where a genome is 90% functional.
Since there’s more functional genome, the total number of deleterious mutations possible will be higher. And in the same way, since there’s more functional genome available to mutate, the probability of a deleterious mutation at a functional site will also be higher.
What am I missing?
I suspect the difference between ‘a deleterious (mutation at a functional site)’ vs ’ (a deleterious mutation) at a functional site’
1 Like
Fair enough. But the passage you quote precisely shows that d, the number of deleterious mutations per individual depends on n, the number of functional sites, which, obviously, incorporates f, the functional fraction of the genome [n=(1/2)fg, with g being the number of diploid sites in the human genome]. So you can see that the authors leave no room to the idea that the number of deleterious mutations is independent of f.
Yes, you are missing something. Here is the relevant passage that should help you see where you are wrong:
The value of u is very small and following common practice we ignore terms of order u 2 in our development, as well as the exceedingly small effect of back mutations from M to A . Empirical estimates of human mutation rates include all mutations that occur. However, not all mutations, even at a functional site, are necessarily deleterious. We use v to denote the empirically estimated rate of mutation per base pair in the human genome and p to denote the probability that a mutation is deleterious. Then, u = vp is the probability of a deleterious mutation at a functional site. We adopt the value v=1.2×10−8 in agreement with recent studies (Lesecque et al. 2012; Besenbacher et al. 2015; Milholland et al. 2017). Graur (2017)provides evidence that 0.4 is a reasonable value for p , the probability that a mutation at any functional site is deleterious, and we also adopt this value.
So, as you can see, u=vp and so is independent of the proportion of the functional genome. Given the values chosen for v and p, u is a constant that is equal to 0,48x10-8.
Thanks. I see what they mean. The problem seems to be that your statement was not sufficiently clear in describing that u is the probability of a deleterious mutation at a functional site(apparently meaning the probability that a mutation that happens at one specific locus will be deleterious) as opposed to the probability of a deleterious mutation at any functional site across the entire genome.
2 Likes
The sections I quoted do not mention the number of functional sites. d is a variable you just made up, which isn’t in the paper at all. The authors actually state the opposite, that f depends on nu and incorporates n. You have it backwards.
Since you haven’t quoted anything written by the authors, have misunderstood what they wrote, couldn’t find a phrase that you were explicitly searching for and are inventing variables they didn’t use, I see no such thing.
Perhaps if you quoted a relevant section from their text I might see that. Until then, sayonara.
1 Like
That’s the probability of a deleterious mutation happening at a specific functional site. It is not the same as the probability that a mutation that happens is at a functional site and deleterious. The former is not relevant to genetic load; the latter is.
3 Likes
Conditional probability strikes again! 
1 Like
No, this is not the meaning of u. U is the probability of occurrence of a deleterious mutation at any given functional site. According to your interpretation of u in your quote above, u would depend on f, which is not the case, since u=vp.
Exactly.
Correct
Your are wrong here. U being the per-base deleterious mutation rate, it is of course relevant to mutation load. All things being equal, the higher u, the higher the load.
You’re wrong here. Below is the section you quoted:
As we have shown, the properties of this distribution depend not just on nu (the number of de novo deleterious mutations per individual) but also on s , the selection coefficient against deleterious mutations.
As you can see, the section mentions n, which is precisely the number of functional sites!
The quote you refer to mentions the number of deleterious mutations. So I didn’t just made this variable up. For practical reason, I only gave this variable a name, namely d.
Wrong again. Please read the abstract.
Wrong again. I quote myself: But the passage you quote precisely shows that d, the number of deleterious mutations per individual depends on n, the number of functional sites, which, obviously, incorporates f, the functional fraction of the genome [n=(1/2)fg, with g being the number of diploid sites in the human genome]. So you can see that the authors leave no room to the idea that the number of deleterious mutations is independent of the functional fraction of the genome.
As every one can see, it’s you who have misunderstood what they wrote, starting with your initial claim that the authors seem to say that the number of deleterious mutations is independent of f, the functional fraction of the genome. Will you have the elegance to recognize that this initial claim of yours was false ?
- You aren’t one of the authors, so quoting yourself is pointless.
- The short quote included in that post was added in an edit made after I said you hadn’t quoted anything written by the authors. So I wasn’t wrong at all.
If you had anything worthwhile to say you wouldn’t need to engage in such shenanigans.
3 Likes
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.