next level DNA sequencing

so here is my idea for DNA: lets make an AI that knows human DNA, knows spider DNA and whatnot and understands that when all the letters in a 4 gigabyte DNA code are in correct order, it prints out a human

likewise it does for a spider

AI has access to incredible amount of data

AI can be given a piece of DNA which function is not known and it then "dreams" a picture of an organism that should come out of that DNA according to what AI believes is what makes a human, spider, a lizard, hydra vulgaris and whatever

we have huge amount of sequenced DNA from all these organisms but no AI that is being fed with this info so that it can work some AI magic upon it

Thalidomide Vintage Ad Shirt $22.14

Nothing Ever Happens Shirt $21.68

Thalidomide Vintage Ad Shirt $22.14

  1. 1 month ago
    Anonymous

    >when all the letters in a 4 gigabyte DNA code are in correct order, it prints out a human
    You don't need AI for that. If you have a full and correct/non damaged dna sequence you can just compare it to other sequences in a database and quickly find out who (or what) it belongs too

    • 1 month ago
      Anonymous

      >If you have a full and correct/non damaged dna sequence you can just compare it to other sequences in a database and quickly find out who (or what) it belongs too

      I know that.
      But AI can do something a human can't.

      If you give it a 4 gigabyte long DNA which doesnt match exactly anything on database but is somewhere in between a human and a gila lizard, the computer can "dream" the creature what could it look like when it had that DNA

      The image generation ability has already been demonstrated, it just that no DNA has ever been added to it.

      • 1 month ago
        Anonymous

        on what DNA database would you train it?

        • 1 month ago
          Anonymous

          dunno at the moment but commercial DNA databases are at least huge, just not free

  2. 1 month ago
    Cult of Passion

    >AI has access to incredible amount of data
    Anon.....every covid test was is also a DNA sample.

    You...dont know the gravity of the rope you tug at...

    • 1 month ago
      Anonymous

      wouldn't you need 3D body scans associated to each DNA? so AI can figure out the relationships between the DNA code and your physiology? Just having access to valid DNA code is practically useless for the purposes of figuring out how someone would look based on DNA sequence.

      • 1 month ago
        Cult of Passion

        >wouldn't you need 3D body scans associated to each DNA?
        Youre looking at a camera right now, kiddo...voice, retina, fingerprints...what makes you...tic on the inside.

        >so AI can figure out the relationships between the DNA code and your physiology?
        Study his body of work as he coveres thay exact topic many times from multiple angles.

        >Just having access to valid DNA code is practically useless for the purposes of figuring out how someone would look based on DNA sequence.
        He makes this points several times.

        My personal research was very much in line with his but from a far more BioPhysics approach than Molecular.

        • 1 month ago
          Anonymous

          that seems like a huge undertaking even for a governmental entity. even with access to the 23&me dump you'd still need to scrap social media accounts for photos/video but you'd still have a non-optimal database. but who knows, might work. even if only to some extent.

          • 1 month ago
            Cult of Passion

            >that seems like a huge undertaking even for a governmental entity
            Sure.

            >Estimates about the storage capacity of the Utah Data Center vary, but it has been reported that the facility was designed to handle yottabytes (1 yottabyte = 1 trillion terabytes)
            Your whole shit hyperlinked to any agency or organization containing further details...like a folder on Windows.

          • 1 month ago
            Anonymous

            >Your whole shit hyperlinked to any agency or organization containing further details...like a folder on Windows.
            ngl that is arousing, for science

          • 1 month ago
            Anonymous

            surely everyone can see that what has been proposed in OP is not impossible from viewpoint of tehcnology itself?

            it all comes to availability of necessary information that will be fed to the computer

          • 1 month ago
            Anonymous

            >it all comes to availability of necessary information that will be fed to the computer
            if you make it open you might get more quality info from people ok to help. huge pool across the planet. open means less compute resources but might mean more quality data. I have no idea if in this case less people but more quality and diverse data would be better than scrapped data without consent but way higher pool + clearly extra compute resources. what would be "better" in this case? for fastest way of "cracking" DNA?

          • 1 month ago
            Anonymous

            Even if you have the entire human genome sequenced you still need to know what every individual gene does and what effect changing a single base pair (of which there are more than 3 billion) is going to have on human body.

          • 1 month ago
            Anonymous

            changing only one bas pair in the entirety of DNA often doesnt do anything

  3. 1 month ago
    Anonymous

    plus full body MRIs, at various life stages. as in that would be even better for other stuff such as organs/bones etc. tons of useful stuff to get out of it.

  4. 1 month ago
    Anonymous

    you have no idea of genome structure. Cells are nondeterministic, genome is not stable through life, extrachromosomal randomly acumulating sequences exist and influence the phenotype, inactivation of one reduntant sex chromosome in women is random and leads to mosaic/phenotype in the body, gene expression on single cell scale is closer to quantum mechanics orbital calculations rather than computer circuits. Often apparently simple relations are closer to directed stochastic fluctuations rather than conected feedback loops. Viewing genome as source code that is interpreted deterministically with some variation is a clear sign of inexperience (along with calling the genome "DNA")

    • 1 month ago
      Anonymous

      it's what happens when you take the "computer" and "encoding" metaphors too seriously
      even gets the best of senior researchers unfortunately

    • 1 month ago
      Anonymous

      it's what happens when you take the "computer" and "encoding" metaphors too seriously
      even gets the best of senior researchers unfortunately

      I never claimed I know something about biology
      I have masters degree in computer science

      That's why I suggested my AI should be used as a tool of biologists

      • 1 month ago
        Anonymous

        you want your AI to do something that is impractical and probably impossible. Do you know how few organisms have whole genomes sequenced? As stated, the relationship between genome and phenotype is fuzzy and non-deterministic, so such AI would not work. There is no need to "dream" a hypothetical organism based on its genome, what would you use it for? Generate gorrilion genomes, most of which can't work, and choose ones that generated nice looking photos? Even it was fast and accurate you can't just plug new synthetic genome into a blank slate cell as such cells don't exist. As for identification, we already have alignment algorithms (like BLAST) which can be used to identify species based on comparatively short sequences (like quickly evolving 16S rDNA in prokayotic cells and 18S rDNA in eukaryotes). Most labs don't have enough money, to waste it on whole genome sequencing, contig and consesus sequence computation, not including the fact that 100% sequencing coverage of short sequences is hard to get let alone a whole genome with repetitive sequences which length is hard to approximate — how do you find the exact length of a sequence of repeating CG, that is longer than 1000 characters, if your reads are at most 300 long?

      • 1 month ago
        Cult of Passion

        >That's why I suggested my AI should be used as a tool of biologists
        Michael Levin does.
        >[wiki]...AI tools for discovery and testing of algorithmic models linking molecular-genetic data to morphogenesis; using techniques from AI, computational neuroscience...
        More attendy lectures and less posty-posty, mk?

        • 1 month ago
          Anonymous

          none of that was about how genes produce the outwards appearance of an organism

          • 1 month ago
            Cult of Passion

            You need to get your degree before you LARP youre a professor, pip.

            This place is moronic....

          • 1 month ago
            Anonymous

            Aren't you the moron with a PhD in psychology? Why are you weighing in on how the genome works when you clearly know as much as op?

          • 1 month ago
            Cult of Passion

            Evolutionary Biologist and BioPhysicist, actually.

            Psychology is downstream from Cognition, which is downstream from just 'Physics-Physics'.

            This has EVERYTHING to do with "appearence".

            THE SOLUTION TO THE ISSUE OF OP IS IN MICHAEL LEVIN'S WORK.

            ANATOMY=PHENOTYPE=APPEARENCE

            t.avid BPS Nova enjoyer

          • 1 month ago
            Anonymous

            someone who has as big popular persona as Levin does, is often overblown. Whether his research is really that groundbreaking, for people to write in the comments under videos with him, that he's incredible and #nobel, or just really good at promoting himself and projecting his self-importance is hard to tell. I firmly stand with Polanyi and Lemaitre that judging the value of scientific contributions is not objective and heavily influenced by the personality of a scientist presenting the work. If I remember correctly, the fact that cells are non-deterministic and can achieve same result from different initial conditions was one of the arguments for vitalism since it didn't fit with the contemporary thermodynamics, that is, the basis of his work has been laid long ago. The big impression he makes on many people appears to stem from their ignorance of modern conception of the cell — like some organic chemistry MS having his mind "blown" by the fact that genome is not source code. The answer to OPs question is not M. Levin's work, it's the work of 21st century molecular biologists as a whole, of which Levin's work is just one part, but the public discourse is always hype-centric

          • 1 month ago
            Cult of Passion

            >The answer to OPs question is not M. Levin's work,
            I wasnt asking. Im lecturing you.
            >someone who has as big popular persona as Levin does, is often overblown. Whether his research is really that groundbreaking, for people to write in the comments under videos with him, that he's incredible and #nobel, or just really good at promoting himself and projecting his self-importance is hard to tell.
            All of this means youre very far behind the cutting edge, you seriously need to wake up or stop grasping at crowns.

            >molecular biologists
            >some organic chemistry MS
            >molecular
            >molecule
            >chemistry
            Cat.
            Meow.

            Youre approaching this from a very "Chemist" perspective. Yeah, everyone has been listening to that for the last 30 years. Take yer gene therapy, yer CRISPER, yer tanshumangobbledeasiatic and SCRAM!

          • 1 month ago
            Anonymous

            >I wasnt asking. Im lecturing you.
            And I'm disagreaing, while you're projecting.
            I'm flatered by your insults. Expected nothing less from a namegay.

          • 1 month ago
            Cult of Passion

            [...]
            [...]
            Why are you pretending to be me?

            Youre at best attempting to master Molecular Biology (Chemistry approach to Biology), let alone BioPhysics and Meta-Genetics (Using my AI like brain to meta-analyse all faces I see traveling tbe world to get a sense of bloodlines and migrations around the world, also sub-bloodlines (developmental stunting/asymmetry).

            (Environmental factors missing from OP or anyone else meams youre all moronic, fricking Chemists...)

            Literally what OP is deluding on about but he (or you) wouldnt accept the answer because he has no clue why the answer is related to the topic (See; thread).

            KNOW WHERE YOUR IGNORANCE BEGINS, HUMAN-MORTALS.

          • 1 month ago
            Anonymous

            ok, DOI or you are a larper

          • 1 month ago
            Cult of Passion

            >DOI
            I dont memorize acolyte vernacular, what is what?

            Second, youre not a professor, youre not "judging credentials in a field youre not even an expert on". Thats moronic....hyper anti-scientific and the very reason "The Science" is garbage these days.

            I bet you didnt even attend his talks, you just LARP youre "b-b-basically a PhD while being some rando midwit student with terminology and give and win to be won".

          • 1 month ago
            Cult of Passion

            >terminology to* give
            Bong rip time...not like I'll get competition losing brain cells or anything....

          • 1 month ago
            Anonymous

            I want the Digital Object Identifier of the papers you wrote regarding "using your AI like brain to meta-analyse all faces you see travelling the world to get a sense of bloodlines and migrations around the world", but since you don't know what DOI is you didn't write any, or read any.

          • 1 month ago
            Cult of Passion

            Peer review is for those that are not certain.

            The fact you dodge the science to LARP as some aiditor is self evident youre untrustworthy woth scientific rigor...I would bar you from the field if I could...but the vast majority of bonafide doctors would be as well.

            IS THERE A LEGITIMATE RESEARCHER IN THE THREAD OR JUST ME?

          • 1 month ago
            Cult of Passion

            https://i.imgur.com/j9bx4We.jpg

            >know as much as op
            And Jesus Christ, you people read the moronic shit OP's post?!

            As soon as it was clear he has no fricking clue about anything its a "shitpost'n'flex" thread. I flexed.

            I can flex again, kid.
            [pops collar]
            ...as in 'This is why Psychology is tangent to OP's discussion'. Widdling Cognition down to the molecular level.

            https://youtu.be/I-mm8Y9D5hU
            Damnit.

            Why are you pretending to be me?

          • 1 month ago
            Cult of Passion

            >know as much as op
            And Jesus Christ, you people read the moronic shit OP's post?!

            As soon as it was clear he has no fricking clue about anything its a "shitpost'n'flex" thread. I flexed.

            I can flex again, kid.
            [pops collar]
            ...as in 'This is why Psychology is tangent to OP's discussion'. Widdling Cognition down to the molecular level.

          • 1 month ago
            Cult of Passion

            Damnit.

  5. 1 month ago
    Anonymous

    >4 gigabyte
    Very subtle bait. I remember that thread from last year.

    • 1 month ago
      Anonymous

      https://biology.stackexchange.com/questions/110458/dna-as-a-data-storage-medium-how-many-gb-can-a-human-genome-store
      seems like human DNA could store about 750MB of data. but I think most of what is already sequenced is not full but partial.

  6. 1 month ago
    Anonymous

    >It's another "AI" shitpost
    I can't recall any more misunderstood yet heavily referenced phenomenon in my lifetime. What even comes close?

    • 1 month ago
      Anonymous

      the entirety of quantum physics

Your email address will not be published. Required fields are marked *