AI Turns Its Artistry to Creating New Human Proteins (nytimes.com) 27
Using many of the same techniques that underpin DALL-E and other art generators, these scientists are generating blueprints for new proteins -- tiny biological mechanisms that can change the way of our bodies behave. From a report: Our bodies naturally produce about 20,000 proteins, which handle everything from digesting food to moving oxygen through the bloodstream. Now, researchers are working to create proteins that are not found in nature, hoping to improve our ability to fight disease and do things that our bodies cannot on their own.
David Baker, the director of the Institute for Protein Design at the University of Washington, has been working to build artisanal proteins for more than 30 years. By 2017, he and his team had shown this was possible. But they did not anticipate how the rise of new A.I. technologies would suddenly accelerate this work, shrinking the time needed to generate new blueprints from years down to weeks. "What we need are new proteins that can solve modern-day problems, like cancer and viral pandemics," Dr. Baker said. "We can't wait for evolution." He added, "Now, we can design these proteins much faster, and with much higher success rates, and create much more sophisticated molecules that can help solve these problems."
Last year, Dr. Baker and his fellow researchers published a pair of papers in the journal Science describing how various A.I. techniques could accelerate protein design. But these papers have already been eclipsed by a newer one that draws on the techniques that drive tools like DALL-E, showing how new proteins can be generated from scratch much like digital photos. "One of the most powerful things about this technology is that, like DALL-E, it does what you tell it to do," said Nate Bennett, one of the researchers working in the University of Washington lab. "From a single prompt, it can generate an endless number of designs."
David Baker, the director of the Institute for Protein Design at the University of Washington, has been working to build artisanal proteins for more than 30 years. By 2017, he and his team had shown this was possible. But they did not anticipate how the rise of new A.I. technologies would suddenly accelerate this work, shrinking the time needed to generate new blueprints from years down to weeks. "What we need are new proteins that can solve modern-day problems, like cancer and viral pandemics," Dr. Baker said. "We can't wait for evolution." He added, "Now, we can design these proteins much faster, and with much higher success rates, and create much more sophisticated molecules that can help solve these problems."
Last year, Dr. Baker and his fellow researchers published a pair of papers in the journal Science describing how various A.I. techniques could accelerate protein design. But these papers have already been eclipsed by a newer one that draws on the techniques that drive tools like DALL-E, showing how new proteins can be generated from scratch much like digital photos. "One of the most powerful things about this technology is that, like DALL-E, it does what you tell it to do," said Nate Bennett, one of the researchers working in the University of Washington lab. "From a single prompt, it can generate an endless number of designs."
Fingers (Score:3)
Just like chatgpt? (Score:1)
If these are as incorrect as chatgpt's stackoverflow answers, we're in for quite a ride.
Re: Just like chatgpt? (Score:2)
It does what you tell you to it to do... with endless possibilities... RTFS
Re: (Score:2)
Creating them or combating them?
Re: (Score:2)
First one, then the other.
Zombies (Score:3)
Do you want zombies? Because that's how you get zombies.
...create proteins that are not found in nature (Score:2, Funny)
Re: (Score:1)
I won't inject these new proteins until they test them on at least 8 mice. That's the threshold.
Re: (Score:2)
A whole bunch of things can go wrong, but the more interesting question is how they will use them, and how will they avoid immune responses to what are basically foreign proteins. These won't be integrated into anyone's genome, they will be used as "biologics" to treat or cure diseases. So if you inject these into people, let's say to correct some protein deficiency, or to remove plaque from artery walls, you need to make sure the immune system doesn't just attack it and make antibodies. One way is to PEGyl
Re: (Score:1)
Re: (Score:2)
Look up enzyme replacement therapy (ERT). When you inject foreign proteins they elicit an immune response. This is why researchers PEGylate proteins that they use for ERT.
Re: (Score:1)
Re: (Score:2)
Yes, some amino acid sequences are more immunogenic than others, this is sorted out by antigen presenting cells who take up and digest exterior proteins, and then present specific amino acid sequences from them on their surface, in conjunction with other surface proteins. So the peptidase systems in these antigen presenting cells can determine which proteins are responded to, and that may be an interesting dataset that could be fed into the AI system to not just focus on protein function, but also on limiti
Re: (Score:2)
Just remember that the spike protein is just one feature on the surface of the virus. The spike protein is not just roaming around by itself (unless you are "vaccinated.") It is attached to the entire protein shell of the virus, a much larger and more complex molecular structure.
This makes a difference.
Re: (Score:2)
I am having flashbacks to prions. Just folding one of the normal proteins wrong will destroy your brain after a chain reaction lasting decades.
Based on recent developments, I doubt there will be adequate testing.
Re: (Score:1)
Re: (Score:2)
The law of unintended consequences combined with complex systems.
This will make prions look like snack cakes.
You Know... (Score:2)
...I think that when STEM students start on their Bachelor's Degree, they should get strapped down ala A Clockwork Orange and spend 48 hours being forced to watch every single movie on the Gone Horribly Wrong [tvtropes.org] section of TV Tropes.
Re: You Know... (Score:2)
Problem with TV is it hardly really reflects reality. Basically any movie of generally therapy or superhuman transformations is bullshit. it would take generations and that's not even what the article is suggesting.
i won't say the idea is stupid but it's riddled with problems. none of which immediately entice anything of significant worry for mankind...Go back to bed at forget the idea a scientist wasted 30 years of his life on this.
Likely Dangerous Prions (Score:1)