4 ms·

On the theme: The phrase "junk" DNA always irritated me. I'm glad it is being replaced with "non-coding". Anybody who has looked at a 4kb demo can intuit that

On the theme: The phrase "junk" DNA always irritated me. I'm glad it is being replaced with "non-coding". Anybody who has looked at a 4kb demo can intuit that "junk" code likely has a function, even if it isn't immediately obvious machine code for the host CPU. I'm no geneticist, and I understand cells aren't CPUs, but I've read enough to know there's at least a tenuous analogy to non-coding DNA and the kind of "junk" you might find reversing a 4kb demo that procedurally generates its output.

TeMPOraL 2y ago
Yup, DNA turned out to not merely be a sequence of triplets telling a dumb matter printer which hard-coded proteins to make - at least according to what little I understand of evolutionary developmental biology[0], DNA is much more like procedural generation in gamedev or demoscene. That is, there's plenty of recipes for various structures and body parts, and then there's lots of DNA that's responsible for conditionally enabling or disabling or modulating those recipes, depending on more DNA that controls when and where and how much to enable them, and then more - a complex network of logic. -- [0] - Didn't get much further than this four-minute intro to the field, but it is a good intro: https://youtu.be/ydqReeTV_vk https://youtu.be/ydqReeTV_vk. (EDIT: It's actually the second part of a trio, that starts with a four-minute bottom-up overview of organic chemistry[1], and ends on a three-minute intro to nanotechnology[2]. I recommend the series together for how well it frames humans in relation to other life and universe as a whole.) [1] - https://www.youtube.com/watch?v=f8FAJXPBdOg https://www.youtube.com/watch?v=f8FAJXPBdOg [2] - https://www.youtube.com/watch?v=ObvxPSQNMGc https://www.youtube.com/watch?v=ObvxPSQNMGc
- throwup238 2y ago
  There’s also epigenetics with mechanisms like histone modification and DNA methylation that can control expression without changing the DNA, but still being heritable.
  - mapt 2y ago
    There are something like (in the reference diagram I'm looking at) eight levels of organization in the structure 'chromosome': https://en.wikipedia.org/wiki/Chromatin#/media/File:Chromatin_Structures.png https://en.wikipedia.org/wiki/Chromatin#/media/File:Chromati... Only the first one is DNA, and only a small portion of 'coding' DNA was initially regarded as important. My question is how? Structural organization implies information. Who thought "Nah, evolution put that in for shits and giggles"? Was it just things we couldn't observe at all without modern scanning electron micrography?
    - adgjlsfhk1 2y ago
      so the thing that makes this complicated is that having non-coding DNA that does nothing for an individual can still be helpful for a population. they essentially serve the same purpose as commented out code that is really common when devs aren't using version control. the comments won't affect the program as it currently exists at all, but they make it so small changes to the code can more easily change the functionality.
      - TeMPOraL 2y ago
        Maybe not as much commented as unused. Turns out, there's plenty of that in DNA, and you can force turning it on. See e.g. the bit about snakes growing legs in the video I linked - they still carry the blueprint for legs in their genome, but have it suppressed.
    - TeMPOraL 2y ago
      > Was it just things we couldn't observe at all without modern scanning electron micrography? Perhaps also that we couldn't imagine, much less accept, that a dumb semi-random process tweaking bits could, over time, organize those bits into higher level abstract structures. We still mostly talk about evolution mutating genes, where perhaps if you zoom out a little, it's actually working at a higher level of abstraction. Incidentally, this is the same outlook as some people have today wrt. LLMs - they can't accept the idea that backprop running on a blob of weights representing a simple (if large) graph can start encoding increasingly high-level, abstract organizational structures.
725686 2y ago
You might enjoy the book Junk DNA by Nessa Carey. Very interesting how complex and interrelated our DNA is. Pretty much spaghetti code.

TeMPOraL 2y ago

Yup, DNA turned out to not merely be a sequence of triplets telling a dumb matter printer which hard-coded proteins to make - at least according to what little I understand of evolutionary developmental biology[0], DNA is much more like procedural generation in gamedev or demoscene. That is, there's plenty of recipes for various structures and body parts, and then there's lots of DNA that's responsible for conditionally enabling or disabling or modulating those recipes, depending on more DNA that controls when and where and how much to enable them, and then more - a complex network of logic. -- [0] - Didn't get much further than this four-minute intro to the field, but it is a good intro: https://youtu.be/ydqReeTV_vk https://youtu.be/ydqReeTV_vk. (EDIT: It's actually the second part of a trio, that starts with a four-minute bottom-up overview of organic chemistry[1], and ends on a three-minute intro to nanotechnology[2]. I recommend the series together for how well it frames humans in relation to other life and universe as a whole.) [1] - https://www.youtube.com/watch?v=f8FAJXPBdOg https://www.youtube.com/watch?v=f8FAJXPBdOg [2] - https://www.youtube.com/watch?v=ObvxPSQNMGc https://www.youtube.com/watch?v=ObvxPSQNMGc

throwup238 2y ago
There’s also epigenetics with mechanisms like histone modification and DNA methylation that can control expression without changing the DNA, but still being heritable.
- mapt 2y ago
  There are something like (in the reference diagram I'm looking at) eight levels of organization in the structure 'chromosome': https://en.wikipedia.org/wiki/Chromatin#/media/File:Chromatin_Structures.png https://en.wikipedia.org/wiki/Chromatin#/media/File:Chromati... Only the first one is DNA, and only a small portion of 'coding' DNA was initially regarded as important. My question is how? Structural organization implies information. Who thought "Nah, evolution put that in for shits and giggles"? Was it just things we couldn't observe at all without modern scanning electron micrography?
  - adgjlsfhk1 2y ago
    so the thing that makes this complicated is that having non-coding DNA that does nothing for an individual can still be helpful for a population. they essentially serve the same purpose as commented out code that is really common when devs aren't using version control. the comments won't affect the program as it currently exists at all, but they make it so small changes to the code can more easily change the functionality.
    - TeMPOraL 2y ago
      Maybe not as much commented as unused. Turns out, there's plenty of that in DNA, and you can force turning it on. See e.g. the bit about snakes growing legs in the video I linked - they still carry the blueprint for legs in their genome, but have it suppressed.
  - TeMPOraL 2y ago
    > Was it just things we couldn't observe at all without modern scanning electron micrography? Perhaps also that we couldn't imagine, much less accept, that a dumb semi-random process tweaking bits could, over time, organize those bits into higher level abstract structures. We still mostly talk about evolution mutating genes, where perhaps if you zoom out a little, it's actually working at a higher level of abstraction. Incidentally, this is the same outlook as some people have today wrt. LLMs - they can't accept the idea that backprop running on a blob of weights representing a simple (if large) graph can start encoding increasingly high-level, abstract organizational structures.

725686 2y ago

You might enjoy the book Junk DNA by Nessa Carey. Very interesting how complex and interrelated our DNA is. Pretty much spaghetti code.