Page 1 of 1
'Non-random mutations'?
Posted: Sun Jan 16, 2022 3:29 pm
by Allo V Psycho
https://www.msn.com/en-gb/health/mindan ... entnewsntp
"I was totally surprised by the non-random mutations we discovered," lead author Grey Monroe, a plant scientist at the University of California, Davis, told Live Science. "Ever since high-school biology, I have been told that mutations are random."
Random mutations are an important part of the theory of evolution by natural selection, in which mutations give rise to adaptations that are passed on to offspring and alter their chances of survival. Scientists have assumed that these mutations were random and that the first step in evolution by natural selection was, therefore, also random. But this may not be entirely true, the new study suggests.
Full disclosure: haven't read the original article. But it seems to me that what is being described is not in fact that mutations are non-random, but that active coding parts of the genome are protected by histones. Either these parts are less likely to be damaged, or repair mechanisms are more active, meaning that they have a lower chance of 'fixing' a mutation. But the mutations themselves are still as random as they ever were. Or am I missing something?
Re: 'Non-random mutations'?
Posted: Sun Jan 16, 2022 3:54 pm
by lpm
If a random thing is instantly discriminated for/against then for all intents and purposes it's not random. Like if you roll a million dice that instantly explode to dust the moment they land on 1, 2 or 3.
I think this is what it's saying? The repair kits in the gene regions instantly explode certain random mutations to dust while other mutations persist, so you only see 4, 5 and 6s. While in the non-gene regions this doesn't happen and you see the full range of mutations, 1 to 6. So the gene regions appear less random than the non-gene regions.
Re: 'Non-random mutations'?
Posted: Sun Jan 16, 2022 4:18 pm
by Bird on a Fire
It seems to be saying that the distribution of mutations across a genome is non-uniform.
If that's the case, non-random seems a bit misleading. There's plenty of non-uniform random distributions.
Re: 'Non-random mutations'?
Posted: Sun Jan 16, 2022 4:29 pm
by shpalman
Surviving organisms are the ones which don't have mutations in the important parts of their genomes shocker.
It didn't even occur to me that "random" in this context would be randomly distributed throughout the genome. I assumed it meant that the mutations weren't "directed" i.e. somehow "chosen" in order to go in a particularly evolutionary direction. But rather the evolutionary pressure is on the new version, whatever it is, as compared to the wild type. There wouldn't be much evolutionary pressure to select either way on mutations which are in parts of the genome which don't code at all, whereas anything which messes up the coding parts is probably most likely to be detrimental and (a) selected against and (b) have mechanisms which correct that sort of thing selected for.
So didn't they know this already or what's new here?
Re: 'Non-random mutations'?
Posted: Sun Jan 16, 2022 4:40 pm
by Bird on a Fire
They're not talking about the results of natural selection. They're talking about where mutations occur before they're exposed to selection.
From the paper:
The greatest barrier to investigating gene-level mutation variability has been a lack of data characterizing new mutations before they experience natural selection. We addressed this limitation by compiling large sets of de novo mutations in A. thaliana (hereafter referred to as Arabidopsis), for which there is rich information on sequence and epigenomic features plausibly linked to mutation rates. We first reanalysed existing Arabidopsis mutation accumulation lines12, combining putative germline and somatic mutations (Fig. 1a, Extended Data Figs. 1, 2, Supplementary Data 1; Methods). A filtering pipeline to eliminate false positives and based on mapping quality, depth and variant frequency retained less than 10% of called variants in a final high-confidence set of mutations. We found no evidence of selection on these mutations. The germline mutations had accumulated in randomly chosen single-seed descendants, so very few mutations, only those causing inviability or sterility, should have been removed by selection12. Somatic mutations experience even less selection21,22. Therefore, as expected, non-synonymous changes and premature stop codons accounted for a greater share of variants than in natural populations, and their frequencies were indistinguishable from a null model of random mutation.
So they've tried hard to remove the impact of selection by focusing on single generations and non-germline mutations, and still find an a priori non-uniform distribution.
As for novelty, I'm sure I was taught 10 years ago that mutation rates can differ between coding and non-coding regions of the genome because of different error-checking procedures. But they find even more distinction, eg between essential and merely nice-to-have genes, for instance.
Re: 'Non-random mutations'?
Posted: Sun Jan 16, 2022 4:43 pm
by Bird on a Fire
There's also evidence that some organisms can increase mutation rates in response to environmental stress, for instance. Which would make more sense if you know your new mutations won't give you a fubar metabolism or something.
Re: 'Non-random mutations'?
Posted: Sun Jan 16, 2022 7:11 pm
by jimbob
Bird on a Fire wrote: ↑Sun Jan 16, 2022 4:43 pm
There's also evidence that some organisms can increase mutation rates in response to environmental stress, for instance. Which would make more sense if you know your new mutations won't give you a fubar metabolism or something.
This was what I was expecting it to be a riff on. And when you think about it, it's not that surprising how that would occur.