hckrnws
Tiny changes in gene expression make big difference in eye size of fruit flies
by PaulHoule
Yep, those are the only words I know for that melody.
... Which is also true of some songs parodied by Weird Al, come to think of it.
Given how evolution seems to work in practice - building beings out of discrete organs and they scale up/down to an optimal size depending on conditions - I'd expect that this will hold for most things.
I'm not nature, but the low-entropy way of building these systems would be to have biological equivalents of variables that control the size (and probably shape) of different features. That would be an evolutionary advantage, because species with genomes organised that way would be amenable to evolving into local optimums.
To rephrase; evolution works best when there is a small area of the genome where tiny changes have large impacts on the size of an organ without altering its function.
> variables that control the size (and probably shape) of different features
Well, in humans, eye sight inherits quite aggressively. If one of your (grand)parents wears glasses, then you have a high probability of wearing them too. This could be related to that.
I'm surprised a tiny change in genes doesn't completely destroy the viability of the organism. You can change one character in code and it no longer compiles or the entire thing crashes.
1) Genetic code translates groups of 3 DNA bases (codons) to 1 aminoacid. But more than one group encodes the same protein. Example: TCT, TCC, TCA, TCG, AGT and AGC = all of them encode Serine. So there are some mutations that lead to the exact same protein. You may think of it as a form of data encoding with some basic data recovery properties.
2) DNA has 2 matching strands. Basically, 2 copies of the same information. DNA constantly degrades and is constantly repaired with information from the other pair. You may think of it as RAID1.
3) We have 2 copies for almost all the genes. Some more than 2. Hemoglobin-alpha has 4 copies. For keratin we have 100+ functional genes (and more non-functional). [1] Think of it as data backups and versioning.
4) For some organism functions, more than one protein do the same job. For some metabolic pathways there are alternatives. It's like we have both GTK and QT working at the same time for any program, and some ancient TK and SDL remains form our ancerstors that could be used if all else fails.
5) The "program" can still crash. It's a miscarriage / spontaneous abortion. Try again with different versions of "libraries" and it will work.
[1] https://en.wikipedia.org/wiki/Keratin-associated_protein
They do "knockout" experiments where people disable various genes all the time and find that some are essential and some aren't. See https://phys.org/news/2025-02-proteins-yeast-colonies-reveal... for one that my RSS reader found for me.
This provides some explanations: https://en.wikipedia.org/wiki/Silent_mutation
in cases where tiny changes destroy the viability of an organism, most likely mutation would cause the destruction of said organism, and the only versions of said organism that remain are those that are not destroyed by tiny changes.
Lookup Walter Gehring's fruit fly experiments if you want something similar and also want to be grossed out.
Dna is the last frontier of programming. It's gods programming language, so to speak.
There may be a day where we can grow machines. Self healing space ships etc
A little bit surprised we are still so early in our understanding of it.
We can grow machines right now. We do it all the time.
yeah, it’s humbling to visit the bio folk from the ivory tower of compsci :)
Crafted by Rajat
Source Code