The largest commercial classes of multi-domain therapeutic proteins include the crispr (and similar) that drive gene therapies, and the chimeric antigen receptors (and similar) that drive cell therapies.
But lead optimization there look different than this page’s efforts.
I mean, the most cutting edge of iPhones, iPads and MacBook Pros _today_ are quite capable of running in realtime today’s high-end local LLMs.
If you project out that hardware just a couple of years, and the trained models out a couple of years, you end up in a place where it makes so much more sense to run them locally, for all sorts of latency, privacy, efficacy, and domain-specific reasons.
Not all that different from the old terminal & mainframe->pc shifts.
Finally - hardware has seemingly gotten out ahead of software that most folks use - watching YouTube, listening to music, playing a game or two. There was a time when playing an mp3 or watching a 4k video really taxed all but the nicest systems. Hardware fixed that problem, like it very well could this one.
A representative idea: if you are the one to kill a deer to eat, you get a little bit of each cut - rib, loin, filet, heart, etc. When was the last time most people ate a filet mignon?
And if you gather food like berries and fruits - you (at least sometimes) get to eat foods ripened that day in the field. How many today get that luxury?
And if one of the things that provides joy to humans is to prepare their family’s food - many folks today would be disqualified.
Yep - cells would die. ...UNTIL they accrue enough dna damage to accidentally either:
- create a new way to generate thymidine
- fall into a proliferation strategy that no longer needs thymidine
Human cells did the same thing and created checkpoints to preclude malignant growth. There are a number of checkpoints already employed in your own bodies. But if you accrue enough DNA damage, you can get around each of those checkpoints (and get cancer):
- IF too much DNA damage, then die
- IF divided too many times, then die
- IF committed to die, actually die
But if a cell collects enough damage, it can get around ALL of the checkpoints. And evolution has shown us that there is no perfect watcher of the watchmen. Still pretty cool to create a new checkpoint this way.
If the DNA is in the cell's nucleus, the DNA will be utilized to produce whatever gene is encoded in the DNA. But if the DNA is just floating around on its own it will not survive a cell division event and the material would need to be periodically reinjected to keep working. However, if the injected DNA is part of a full chromosome, it will be replicated when the cell divides - and will be permanent as long as the cell or its progeny survive.
Some viruses will just inject the DNA into cells, but will not become part of the cell's genome ("transient" transduction). Other viruses (like lentiviruses and these adeno-associated viruses [AAVs]) inject their DNA not just into the cells, but also have machinery that splices their payload DNA directly into the cell's chromosomes ("integrated"). The location in the genome of the splicing event is relatively random. Random is not necessarily great as it could interrupt other genes already in the chromosome. CRISPR is a now-famous tool that helps "integrate" DNA into a specific spot in the genome by being guided to a specific location with a small piece of a specific sequence.
Once the DNA is integrated, any cell, and any of the cell's progeny, will produce or "express" the gene on the delivered DNA. In this case, they delivered the 5991 characters of DNA associated with the OROF gene [1].
There’s one law of physics difference here: (with a few exceptions) helium, because it’s lighter than air, will never replenish on our planet. There is currently most of what there ever will be. Unlike (almost?) every other object of commerce, there is an almost completely inelastic total supply. It is the ultimate non-renewable resource.
Indianapolis was too fast and hot that one year, one of the (two) tire manufactures disqualified their tires (and teams). So not even half the field ran the race. The Indy 500 course has a half-mile straightaway which allowed the F1 cars to build up unprecedented speed - all in hot midwestern summer heat.
Seems the diligence on local weather is not so good with these folks...
Snowden also brought the goods (actual docs) so that the existing cases and inquiries could proceed without getting the Glomar response - “cannot answer the question of whether you have standing, as we can neither confirm nor deny the program either exists.”
And this was deliberate. Those in the know knew _of_ the program, but could do little actionable with that knowledge without Snowden’s public proof.
If you change one of those numbers, you'll need to rejigger the rest, and you'd need to reoptimize. And there are competing goals which at least include:
- maximize access to biophysical/chemical diversity
- minimize energy expenditure to produce each component, chemically
- minimize energy expenditure to both copy instructions & produce products
- maximize information fidelity
- minimize or at least degrade gracefully in the context of errors
In the context of a 3-base system, you very well could throw off those optimizations given the consequences for the other 2 parameters (#AA & nt/AA). 3^3 = 27, which is very close to the maximum of 20 amino acids. Which means you'd probably need a 4nt->AA translation layer to keep the same number of AAs, and that alone would add 30% more energy expenditure. If you kept the 3nt->AA system you'd BOTH need to reduce the number of accessible amino acids AND you'd lose some of the error correction mechanisms of having degenerate codons code for the same amino acid.
BUT, if you look at the codon table, precisely because it's base-4 and not base-3, many base flips are silent when coded.
By using base-4, there's enough space to permit lossiness of the coding itself - given the number of amino acids and the 3-NT encoding.
So you really aren't optimizing JUST for nucleotide encoding, but you're also optimizing in concert with 3-nt/AA, and 20AA codes.
So if you have to optimize for information density and fidelity, given X-nucleotides, Y nucleotides/AA, and Z AAs, and sample as much chemical and physical diversity in those AAs life has settled upon:
X=4, Y=3, Z=20.
If we went with X=3, you might need Y=4 to get the same kind of fidelity, but that cranks up your energy costs by 30% (from 3 to 4 NT per AA).
And in this case, tech savy individuals are willing to send $XXM in 6 clicks, 3min after getting a slack message. That's a pretty quick kind of run relative to old-school 'stand in line for your personal life savings' kind of run.