Training mRNA Language Models Across 25 Species for $165

We built an end-to-end protein AI pipeline covering structure prediction, sequence design, and codon optimization. After comparing multiple transformer architectures for codon-level language modeling, CodonRoBERTa-large-v2 emerged as the clear winner with a perplexity of 4.10 and a Spearman CAI correlation of 0.40, significantly outperforming ModernBERT. We then scaled to 25 species, trained 4 production models in 55 GPU-hours, and built a species-conditioned system that no other open-source project offers. Complete results, architectural decisions, and runnable code below.

73 points | by maziyar 2 days ago

10 comments

  • seamossfet 49 minutes ago
    The problem with models like this is they're built on very little actual training data we can trace back to verifiable protein data. The protein data back, and other sources of training data for stuff like this, has a lot of broken structures in them and "creative liberties" taken to infer a structure from instrument data. It's a very complex process that leaves a lot for interpretation.

    On top of that, we don't have a clear understanding on how certain positions (conformations) of a structure affect underlying biological mechanisms.

    Yes, these models can predict surprisingly accurate structures and sequences. Do we know if these outputs are biologically useful? Not quite.

    This technology is amazing, don't get me wrong, but to the average person they might see this and wonder why we can't go full futurism and solve every pathology with models like these.

    We've come a long way, but there's still a very very long way to go.

  • rubicon33 3 hours ago
    Can someone explain what one might use this model for? As a developer with a casual interest in biology it would be fun to play with but honestly not sure what I would do
    • colechristensen 3 hours ago
      You can get your feet wet with genetic engineering for surprisingly little money.

      This guy shows a lot of how it's done: https://www.youtube.com/@thethoughtemporium

      Basically you can design/edit/inject custom genes into things and see real results spending on the scale of $100-$1000.

      • _zoltan_ 26 minutes ago
        My main concern is using fungi. If it ends up in my lungs I'm most likely screwed, right?
      • someuser54541 2 hours ago
        Is there something like this in text/readable format?
  • maziyar 2 days ago
    • xyz100 4 hours ago
      What makes this dataset or problem worth solving compared to other health datasets? Would the results on this task be broadly useful to health?
      • CyberDildonics 2 hours ago
        What other "datasets" are you talking about? How do you "solve a dataset" ?
  • khalic 4 hours ago
    > In Progress: CodonJEPA

    JEPA is going to break the whole industry :D

    • digdugdirk 3 hours ago
      Can you explain this? I haven't heard of JEPA, and from a quick search it seems to be vision/robotics based?
      • khalic 2 hours ago
        It’s a self supervised learning architecture, and it’s pretty much universal. The loss function runs on embeddings, and some other smart architectural choices allover. Worth diving into for a few hours, Yann LeCun gives some interesting talks about it
      • lukeinator42 3 hours ago
  • simianwords 4 hours ago
    What makes these Domain specific models work when we don’t have good domain models for health care, chemistry, economics and so on
    • colechristensen 3 hours ago
      >we don’t have good domain models for health care, chemistry, economics and so on

      Who says we don't?

  • yieldcrv 3 hours ago
    Distributing the load on this will probably be infinitely more useful than “folding at home”
  • HocusLocus 4 hours ago
    gray goo of the future
  • skyskys 1 hour ago
    hmmmm seems like some fake hype.