Great art isn't necessarily about popularity / wide appeal, though. In fact, the art that isn't of wide, general appeal, is what stands to benefit the most from this kind of benefit.
Both are broad spectrum antivirals, but completely different mechanism.
DRACO "is a chimeric protein with one domain that binds to viral double stranded RNA (dsRNA) and a second domain that induces apoptosis when two or more DRACOs crosslink on the same dsRNA." (Ridder et al 2011). This article is about packaging mRNA for a set of 10 interferon-stimulated genes that express multiple proteins that target various stages of viral replication.
When you get a viral infection, immune cells make a signalling protein called a IFN-1 (Type I Interferon) cytokine, and this flips a boolean flag to True on a bunch of genes (interferon-stimulated genes or ISGs) that produce a bunch of proteins (hundreds) that control the infection. ISG15 is one of them and its role appears to be to downregulate and to limit the inflammation.
The paper title refers to a ISG15 deficiency, meaning if you are dificient in ISG15 that inflamation limitation goes away. But the paper is actually about how in people that naturally have a ISG15 deficiency, there is an always-on low level expression of some of these pro-inflamation genes. So they take that as a safe level.
The did RNA sequencing on experimental ISG15 deficient cells and from heatlhy individuals, identified the mutations, narrowed down to 10 genes (antiviral ones not inhibitors) that in combination significantly inhibited viral replication. They stuck the RNA for such genes in lipid nanoparticles such that they enter host cells, whose ribosomes happily read the RNA like a turing head reads a tape in base 20 and produce proteins encoded by these genes, similar to how the mRNA vaccine works.
So why not dose with the IFN-I directly? Three referenced papers show its "poorly tolerated with significant side effects" and all those downregulators that get expressed limit the inflammation response.
Disclaimer: IANAB (not a biologist) corrections might be due..
This is typical of "why not just one drug/treatment" for something big like viruses, cancer, etc.
I think we'll never have this "one shot," but continue to find tailored treatments for individual conditions. There's no way out of this complexity with "one simple trick," which seems really appealing to the people who determine what gets popular in social media and seemingly politics now. Its just going to be boring and grueling academia and medical trials that are hard for the layperson to understand, hence the important of funding these programs. The recent right-wing election wins and thus a right-wing government cutting all manner of medical grants is supported by the "one weird trick" crowd. Hopefully, the USA will have better leadership in the future to get us back to actual science and to find actual new treatments.
Already, even on HN, the top comments are conspiracy-culture coded, "but, but this one company bought the patent and disappeared with it!" Sigh.
Cool. Could we train a "potential oncoprotein" classifier on Orthrus embeddings? IMO self serve diagnosis and detection is a far larger market than synthesis.
This is a really interesting direction. There is this big field of Cell Free (cfRNA) cancer detection. We talked to a few people in the field and think that embedding sequences for this direction could be really valuable. One challenge here is that it's hard to set up evaluation tasks since the public data is scarce
Maybe we can crowd source data. My platform, currently in beta, has ai assistants for compute infrastructure and biology and will soon let people to do self serve research on their own omics data using models like yours. So there could be a monetization path too if enough people start looking their own cell data (which they might once they fully understand the risks of engineered pathogens, and certainly will when the risks materialize and start hitting home). Email in bio if you want to brainstorm.
That would be really cool. Navigating SRA and mining out reasonable $ relevant tasks is a huge bottleneck.
I find it takes a large amount of effort to parse what the authors are doing, whether the data is high quality, and how to pre-process it in a way that makes sense for the task at hand.
Would love to chat more about how you're thinking of evaluating quality of these agents.
