> Researchers grew specimens of the widely distributed weed Arabidopsis thaliana in a sheltered lab environment, where all plants, including ones with harmful mutations, could reproduce.
Some mutations are so bad that will kill the new plant almost instantly. It's very difficult to get all the mutations without any filtering.
> “We always thought of mutations appearing solely by chance across the genome,” says Grey Monroe, an assistant professor in the UC Davis Department of Plant Sciences and first author of the paper. “It now turns out that the pattern of mutation is not only very non-random, but also that it’s non-random in a way that benefits the plant.”
We already know that the mutation rate is not uniform in all the genome, for example microsatellites have a bigger mutation rate https://en.wikipedia.org/wiki/Microsatellite and that increased mutation rate is useful. From Wikipedia
> Length changes of microsatellites within promoters and other cis-regulatory regions can change gene expression quickly, between generations. The human genome contains many (>16,000) short sequence repeats in regulatory regions, which provide ‘tuning knobs’ on the expression of many genes.
I think there are entire different categories of genes that lead to nonviable offspring, where some genes are "truly essential for all life" (say, coding for a single ribosomal protein), some genes are "not essential in some rare conditions" (genes for surviving extreme desiccation during wet times), and other genes exist in groups that complex relations where a different subsets could be removed (various synthetic pathways, nothing is truly "necessary", but nothing is truly "sufficient"). These differences have to be addressed through different techniques.
I had a wonderful moment during my postdoc where I discovered a large fractions of genes falsely marked as "essential to life" happened to overlapping genes that were essential to life (IE, parts of both got deleted in the experiment). The underlying paper reporting this was never retracfted but later experiments show if you control for this, many genes aren't "essential" but happen to be overlapping ones that are (thus, you can delete the nonoverlapping part of the nonessential gene and the organism is still viable).
And yes, you're 100% right we already knew that mutation rates are varible. IIUC this paper is arguing for some special process that "protects" functionally important regions from mutation, but I haven't read it in detail and I'm too long gone from the field to make reasonable criticisms.
That's a great summary. If you want to go further, I think it's only really addressed in the literature and not much in textbooks.
The one interesting bit that I studied/researched and attempted to publish (no luck) was the observation that organisms that have RNAi, a process by which cells detect double-stranded RNA inside and eliminate it (likely a virus), tend to have fewer overlapping genes. When two different overlapping genes are transcribed to mRNA, they can form duplexes that kinda sorta look like viral RNA and get taken out by the RNAi surveillance system, even though they're legit transcripts.
8 comments
[ 5.4 ms ] story [ 35.8 ms ] threadSome mutations are so bad that will kill the new plant almost instantly. It's very difficult to get all the mutations without any filtering.
> “We always thought of mutations appearing solely by chance across the genome,” says Grey Monroe, an assistant professor in the UC Davis Department of Plant Sciences and first author of the paper. “It now turns out that the pattern of mutation is not only very non-random, but also that it’s non-random in a way that benefits the plant.”
We already know that the mutation rate is not uniform in all the genome, for example microsatellites have a bigger mutation rate https://en.wikipedia.org/wiki/Microsatellite and that increased mutation rate is useful. From Wikipedia
> Length changes of microsatellites within promoters and other cis-regulatory regions can change gene expression quickly, between generations. The human genome contains many (>16,000) short sequence repeats in regulatory regions, which provide ‘tuning knobs’ on the expression of many genes.
I had a wonderful moment during my postdoc where I discovered a large fractions of genes falsely marked as "essential to life" happened to overlapping genes that were essential to life (IE, parts of both got deleted in the experiment). The underlying paper reporting this was never retracfted but later experiments show if you control for this, many genes aren't "essential" but happen to be overlapping ones that are (thus, you can delete the nonoverlapping part of the nonessential gene and the organism is still viable).
And yes, you're 100% right we already knew that mutation rates are varible. IIUC this paper is arguing for some special process that "protects" functionally important regions from mutation, but I haven't read it in detail and I'm too long gone from the field to make reasonable criticisms.
I remember about overlapping genes in bacteriophages, but I though is was a weird trick because a big part of the bacteriophages is the DNA.
The one interesting bit that I studied/researched and attempted to publish (no luck) was the observation that organisms that have RNAi, a process by which cells detect double-stranded RNA inside and eliminate it (likely a virus), tend to have fewer overlapping genes. When two different overlapping genes are transcribed to mRNA, they can form duplexes that kinda sorta look like viral RNA and get taken out by the RNAi surveillance system, even though they're legit transcripts.
https://www.nature.com/articles/s41586-021-04269-6
[1] Wolfe, Kenneth H.; Nature volume 337, pages 283–285 (1989) ; "Mutation rates differ among regions of the mammalian genome" : https://www.nature.com/articles/337283a0