AI Used To Design a Multi-Step Enzyme That Can Digest Some Plastics 32
Leveraging AI tools like RFDiffusion and PLACER, researchers were able to design a novel enzyme capable of breaking down plastic by targeting ester bonds, a key component in polyester. Ars Technica reports: The researchers started out by using the standard tools they developed to handle protein design, including an AI tool named RFDiffusion, which uses a random seed to generate a variety of protein backgrounds. In this case, the researchers asked RFDiffusion to match the average positions of the amino acids in a family of ester-breaking enzymes. The results were fed to another neural network, which chose the amino acids such that they'd form a pocket that would hold an ester that breaks down into a fluorescent molecule so they could follow the enzyme's activity using its glow.
Of the 129 proteins designed by this software, only two of them resulted in any fluorescence. So the team decided they needed yet another AI. Called PLACER, the software was trained by taking all the known structures of proteins latched on to small molecules and randomizing some of their structure, forcing the AI to learn how to shift things back into a functional state (making it a generative AI). The hope was that PLACER would be trained to capture some of the structural details that allow enzymes to adopt more than one specific configuration over the course of the reaction they were catalyzing. And it worked. Repeating the same process with an added PLACER screening step boosted the number of enzymes with catalytic activity by over three-fold.
Unfortunately, all of these enzymes stalled after a single reaction. It turns out they were much better at cleaving the ester, but they left one part of it chemically bonded to the enzyme. In other words, the enzymes acted like part of the reaction, not a catalyst. So the researchers started using PLACER to screen for structures that could adopt a key intermediate state of the reaction. This produced a much higher rate of reactive enzymes (18 percent of them cleaved the ester bond), and two -- named "super" and "win" -- could actually cycle through multiple rounds of reactions. The team had finally made an enzyme.
By adding additional rounds alternating between structure suggestions using RFDiffusion and screening using PLACER, the team saw the frequency of functional enzymes increase and eventually designed one that had an activity similar to some produced by actual living things. They also showed they could use the same process to design an esterase capable of digesting the bonds in PET, a common plastic. The research has been published in the journal Science.
Of the 129 proteins designed by this software, only two of them resulted in any fluorescence. So the team decided they needed yet another AI. Called PLACER, the software was trained by taking all the known structures of proteins latched on to small molecules and randomizing some of their structure, forcing the AI to learn how to shift things back into a functional state (making it a generative AI). The hope was that PLACER would be trained to capture some of the structural details that allow enzymes to adopt more than one specific configuration over the course of the reaction they were catalyzing. And it worked. Repeating the same process with an added PLACER screening step boosted the number of enzymes with catalytic activity by over three-fold.
Unfortunately, all of these enzymes stalled after a single reaction. It turns out they were much better at cleaving the ester, but they left one part of it chemically bonded to the enzyme. In other words, the enzymes acted like part of the reaction, not a catalyst. So the researchers started using PLACER to screen for structures that could adopt a key intermediate state of the reaction. This produced a much higher rate of reactive enzymes (18 percent of them cleaved the ester bond), and two -- named "super" and "win" -- could actually cycle through multiple rounds of reactions. The team had finally made an enzyme.
By adding additional rounds alternating between structure suggestions using RFDiffusion and screening using PLACER, the team saw the frequency of functional enzymes increase and eventually designed one that had an activity similar to some produced by actual living things. They also showed they could use the same process to design an esterase capable of digesting the bonds in PET, a common plastic. The research has been published in the journal Science.
Re: (Score:3)
Have you looked into anger management courses? And you might get them to add a bit of anti-homophobia serum to your meds.
Re: (Score:2)
anti-homophobia serum
MAGA told me this is called soy milk.
Re: (Score:1)
Isn't that called anthrax?
Grey goo (Score:2)
Research first, then... (Score:2)
Looks like we're headed for a period of time where you can follow the formula "Something people do..." + "...with AI" and get a patent.
Sartre was right. He said, "People are like dice. We throw ourselves in the direction of our own choosing." The Dice are the patents.
Re: (Score:3)
Looks like we're headed for a period of time where you can follow the formula "Something people do..." + "...with AI" and get a patent.
Designing "multi-step enzymes" was not "something people do..." though was it ?!
"AI" in the hands of qualified people made it feasible.
Re: (Score:1)
Designing "multi-step enzymes" was not "something people do..." though was it ?!
I don't know why you think that.
"The ignorant man is not free, because what confronts him is an alien world, something outside him and in the offing, on which he depends, without his having made this foreign world for himself and therefore without being at home in it by himself as in something his own. The impulse of curiosity, the pressure for knowledge, from the lowest level up to the highest rung of philosophical insight arises only from the struggle to cancel this situation of unfreedom and to make t
Re: (Score:2)
Ah, another brainless blurb of fuzziness from the our resident "philosopher" quoter. He doesn't understand shit so he substitutes that with philosophy that he also does not understand.
Re: (Score:2)
1. In 2007, combination patents were banned by the SCOTUS in KSR v. Teleflex [wikipedia.org].
2. This is not a combination patent. The innovative enzyme is being patented, not the fact that AI was used to design it.
Re: (Score:1)
2. This is not a combination patent. The innovative enzyme is being patented,
This is also not a story about a patent. Lay off the weed a bit.
Nothing could possibly go wrong (Score:1)
...go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, ...
Re: (Score:2)
In the early 1970s there was a series on TV in the UK called Doomwatch. The opening episode was called "The Plastic Eaters" and featured just this scenario.
The story line was expanded into a full length novel "Mutant 59 - the Plastic Eater" by the original screenwriter.
the TV episode is now on youtube
https://www.youtube.com/watch?... [youtube.com]
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Well....maybe. On the other hand there is a lot wrong with flooding our one and only Earth with plastics.
Re: (Score:2)
...go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, go wrong, ...
Da do wrong-wrong-wrong, da do wrong-wrong...
But what we really want to know is (Score:2)
Afterward, does it go "Mmm... plastic."?
Re: (Score:2)
And it's set in China. Wuhan to be precise.
Old AI (Score:4, Interesting)
Old and weak AI has so much potential it will take decades for humans to catch on and use all of it. I find it really interesting that we can now create atom-level models of parts of a cell and figure out exactly how they work and we can even design our own proteins that interact with them. This is pretty much what has been missing when researching the cure for aging and all other deceases and grow back missing limbs and tooth. This is the beginning of the age of immortality.
And once we have cure for everything, imagine how big of an impact it will be to the budget of countries that currently use at least quarter of their budgets to healthcare. And if retirement age is removed. Suddenly governments have a lot more money, more workers and at the same time many people are left unemployed from healthcare. Industrial revolution will look weak in comparison.
Re: (Score:2)
AFAIK, we don't yet understand aging as a process. That being the case, interventions are at the "try it and find out" level. You can make a lot of progress that way, but you can't really design interventions that will work...at least not work the way you intended.
P.S.: Many interventions that *might* reduce aging *do* produce cancers.
Re: (Score:2)
I find it really interesting that we can now create atom-level models of parts of a cell and figure out exactly how they work and we can even design our own proteins that interact with them.
It is interesting but there is still plenty we don't understand because biology is really what you get when you apply statistics to chaos. Understanding biology is still quite a ways off but we will continue to chip away at it in increasingly larger amounts.
This is the beginning of the age of immortality.
It would be quasi-immortality because you could still be killed by significant injury.
And once we have cure for everything...
It seems more likely that we will go the route of self-modification before that happens and that is a whole new can of worms. Also, please keep in mind that knowledge i
I hope they read Slashdot (Score:2)
I hope ... (Score:1)
This is the proper use of AI (Score:2)
..not using the tools to make so-called "art" or "music"
We already have artists who make art just fine
We need AI to solve previously intractable problems
Wasn't research involving females banned? (Score:2)
We use plastic--because it doesn't dissove! (Score:2)
Wuhaneda strain (Score:2)
In other news - high potency esterase escapes from Chinese lab, dissolving plastic parts around the globe.
Michael Crichton warned us (Andromeda Strain). We were fools to not listen.
Veritasium Video is Good (Score:2)
The recent Veritasium video on AlphaFold and RF Diffusion is a great backgrounder for those unfamiliar with the algorithms and applications.
I had a passing knowledge but learned quite a bit.