a few months ago Google DeepMind revealed a new AI model called AlphaFold. This model is able to generate designer proteins, the very building blocks of life. Now, I wanted to cover this subject, but since I don't have a background in biology, this was a little

bit out of my wheelhouse. However, there was another channel in the AI space where an actual engineer with a biotech background had covered this very AI model and did it phenomenally well. Natalia from the Tech Rogue channel did an excellent job covering this subject.

Since that channel is new, it didn't get the views that I think it deserved. So, with her permission, I'm reposting it here. Let me know what you think. I'm considering trying to highlight some smaller channels in the AI space that are doing a great job explaining this

stuff, but maybe not getting the algorithmic recognition that they deserve. Keep in mind, since this video, the Google DeepMind team did win the Nobel Peace Prize for their work with proteins. This is including AlphaFold and probably AlphaFold as well. Take a

look. Google DeepMind just released AlphaFold. AlphaFold is the next step in the evolution of AlphaFold. AlphaFold shows great promise in what is called protein design. Most of the things that we think of as life and living creatures are governed and made possible by

proteins, like Lego pieces. Proteins are the building blocks of life, and now it seems like we're getting close to custom designing those building blocks that we need to, well, build life. Demis Hassabis, the founder of DeepMind, he's a big part of

why we're seeing this technology today He also has some thoughts on what he thinks that this technology is going to be bringing in the next couple decades We're showing the way with things like AlphaFold and isomorphic AI I think we could, you know, cure most diseases within

the next decade or two And keep in mind that this is not a person that is prone to hyperbole He's just not that hype guy So let's unpack why this is such an important breakthrough If you enjoy these videos please hit the thumbs

thumbs up button so that I can get this out to more people So first and foremost let's get some terms out of the way Proteins are the building blocks of life and help do everything that we think of as what a living organism does Life is

about to form on this planet for the very first time A group of amino acids are about to combine to form the first protein the building blocks of what you call life Protein binding refers to how proteins in your body interact with different molecules So this could be things like drugs

hormones or even some nutrients Each protein has a specific shape and only certain molecules can fit into it Think of it like a puzzle piece and a specific puzzle piece can connect to that specific protein and when they connect this is called binding So as an analogy

imagine that your body is a huge park filled with playgrounds and the proteins are the functional things like swings slides water fountains etc and the kids are the molecules that are flowing wildly around So if a kid sits down on a swing that's like a molecule binding to

A protein, if the kid can sit on the seat, it can use the swing. And if the seat is broken or deformed, the swing can't be used. Also, an elephant, for example, would not be able to use that swing; it's the wrong shape, it's the wrong size. I think

I nailed that analogy. But what does this have to do with Alpha Proteo? Take a look at these images. The big, light yellow blob is the target protein. This is the protein that we want to interact with, that we want to have an effect on. In this case, some of the target proteins

are ones that are involved with cancer, autoimmune disease, and viral infections. The blue ribbon is the binding protein that attaches itself to the target protein, and this blue ribbon is what Alpha Proteo designs. So now we're able to create these custom binding proteins that are able to attach to existing

proteins and have useful biological function. It allows us to influence the target protein's activity or function. We can do things like activate it or inhibit it. We can stabilize it, help it maintain its proper structure, and prevent misfolding. We can also mark it

for another process like degradation. If you've ever seen those trees that are marked with a spray-painted X, those are marked for being chopped down. In a similar way, we can tag certain proteins to be chopped up for parts. It's actually a really cool process where once a

protein gets tagged for degradation, this proteasome comes along and, well, takes care of it. Do you know a guy named Coco? He came up out of nowhere and just started saying all this weird stuff. Kind of weird stuff. He came over and pulled some crap. It's all right, I'll talk to

somebody peppers and cheese G all right we wow wa sit down sit down easy [Music] the other one a proteasome is a relatively large molecular machine that's made up of multiple proteins that together form a functional unit so this thing takes in

the tagged protein chops it up for parts or amino acids and then recycles it so why is all of this important why does it change anything well it seems to indicate that we might be entering entering into a whole new era of Life Sciences here's Jensen Wong the founder

of Nvidia on how he thinks that AI will permanently change scientific discovery take a listen I would realize uh one thing that one of the most complex fields of science is the understanding of biology human biology we call this field Life Sciences and we call drug Discovery

Discovery as if you wander around the universe and all of a sudden Hey look what I discovered nobody in computer science nobody in computers and nobody in the traditional industries that are very large today nobody says car Discovery we don't say computer Discovery we don't say software

Discovery we don't go home and say hey honey look what I found today this piece of software we call it engineering and every single year our science our computer science our software becomes better and better than the than the year before every single year our chips get better every single

year our infrastructure gets better however Life Sciences is sporadic if I were to do it over again right now I would realize that the technology to turn life engineering life science to life engineering is upon us and that digital biology will be a field of engineering not a field of science it

will continue to have science of course, but not a field just of science in the future. And so I I hope that that this is going to start a whole generation of people who enjoy working with proteins and chemicals and and enzymes and um materials and and they're engineering

these amazing things that are more energy efficient, that are lighter weight, that are stronger, that are more sustainable. All of these inventions in the future are going to be part of engineering, not scientific discovery. For most of human history, life sciences was about discovery. In 1674 we discovered

microorganisms and that must have been weird, realizing that you are entirely covered by small crawling bugs that you can't see. And then imagine trying to convince other people that this exists and no no you're not crazy. And then in 1928 we discover antibiotics, that's over

250 years between discovering microorganisms and the discovery of antibiotics. Interestingly, that might be the longest time in human history from the time that we've actually discovered something and learning how to kill it. So anyway, we also discovered viruses, DNA, stem cells, etc. None of those things were

invented, they weren't designed, they weren't engineered. We found them in nature and then tried to build on top of it or to improve it. But today we have this intersection between machine learning and life sciences, and this is where we're beginning to see that we

can actually engineer life. We can build custom proteins that have never evolved in nature. In a previous video, we talked about evolutionary scale and their AI model called ESM3. It was tasked with creating a new, never-before-seen GFP, or green fluorescent protein. It was able to

create that new protein structure and it actually worked as intended. Something like this evolving naturally would have taken over 400 million years, and now we can create it in a lab. This means that protein design might completely change multiple fields. One of the most promising areas where being able to

create functional custom proteins is first and foremost drug design. As you saw, AlphaFold was able to effectively create binding proteins for cancer, autoimmune disease, and viral infections. It could help with vaccine development, enzyme engineering for use in pharmaceuticals and biofuels. We might be

able to engineer enzymes that can break down pollutants and toxic chemicals. And this might also help with gene editing. For example, Jennifer Doudna, one of the creators of CRISPR, mentioned recently of a challenge that they met in treating disease. CRISPR is a powerful

gene editing technology. It relies on a specific group of proteins to cut and edit DNA. In a recent talk, Jennifer Doudna specifically mentioned how large the Cas9 protein is. This poses a lot of problems and makes some treatments not feasible. Being able to design proteins

that meet certain criteria, like having the function that we want while being significantly smaller, well, that would unlock a lot of different possibilities for treatments. Technology like AlphaFold could potentially give us the ability to cure and treat a lot more diseases. In fact, here's what Demis Hassabis

thinks that we can achieve. "We're showing the way with things like AlphaFold and Isomorphic AI. I think we could, you know, cure most diseases within the next decade or two if AI drug design works. And then there could be personalized medicines where it

minimizes the side effects on the individual because it's it's mapped to the person's individual illness and their individual metabolism and so on. So these are kind of amazing things, um, you know, clean energy, renewable energy sources, you know, fusion or better solar

power all of these types of things. I think they're all within reach, and then that would sort out water access because you could do desalination everywhere. So I just feel like, um, this enormous good is going to come from uh these technologies, um, but we

have to mitigate the risk too. So that's it for me. Please hit like, subscribe. My name is Natalia and this is the Tech Rogue Channel. See you soon.

AlphaProteo - Google DeepMind's Breakthrough AI for "Protein Design"

AI News by Wes Roth.

Bringing you the latest AI news, emerging tech and artificial intelligence breakthroughs.

We cover:
AI
Robots
Virtual reality
Space
Self-driving cars
Brain computer interfaces
Video Games
Elon Musk
Sam Altman
OpenAI
DeepMind

...and anything related.



Wes Roth

단백질을 설계할 수 있는 AI 모델인 AlphaFold를 소개합니다.

테크 로그의 나탈리아가 알파폴드에 대한 훌륭한 보도를 언급했습니다.

소규모 AI 채널을 강조하기 위해 허가를 받아 동영상을 다시 게시합니다.

알파폴드 등 단백질 연구로 노벨 평화상을 수상한 구글 딥마인드의 노벨상 수상에 대한 언급.

단백질 설계에서 알파폴드의 잠재력과 생명체 이해에 미치는 영향.

데미스 카사비스의 비전과 알파폴드의 잠재적 영향력

알파폴드 개발에서 데미스 하사비스의 역할과 그의 낙관적인 전망.

향후 10~20년 내에 대부분의 질병을 치료할 수 있다는 것이 카사비스의 예측입니다.

알파폴드 혁신의 중요성을 강조합니다.

주요 생물학적 개념에 대한 설명

생명의 구성 요소인 단백질의 정의.

단백질 결합과 생물학적 과정에서의 역할에 대해 설명합니다.

신체를 공원에 비유하면 단백질은 기능적 요소로, 분자는 어린이로 비유할 수 있습니다.

알파 프로테오의 기능 및 애플리케이션

알파 프로테오의 맞춤형 결합 단백질 설계 기능에 대한 설명입니다.

암, 자가 면역 질환, 바이러스 감염에 관여하는 단백질을 표적으로 삼는 데 활용됩니다.

알파 프로테오의 표적 단백질 활성에 영향을 미치는 기능(활성화, 억제, 안정화, 분해 표시).

프로테아좀과 관련된 단백질 분해 과정에 대한 설명입니다.

생명 과학에 대한 AI의 혁신적 영향력

생명과학을 발견에서 공학으로 전환하는 데 있어 AI의 역할에 대한 젠슨 웡의 관점.

생명과학과 컴퓨터 과학 및 소프트웨어 공학 등의 다른 분야를 비교합니다.

디지털 생물학이 단순한 과학이 아닌 공학의 한 분야가 될 것이라는 예측.

생명과학에 대한 역사적 관점, 발견과 공학의 대비.

단백질 설계의 응용 및 향후 시사점

생명을 공학적으로 설계하고 자연계에 존재하지 않는 맞춤형 단백질을 만들 수 있습니다.

속도와 효율성을 강조하는 새로운 GFP 단백질을 생성하는 ESM3의 예시입니다.

약물 설계, 백신 개발, 효소 공학 및 유전자 편집에 잠재적으로 응용될 수 있습니다.

단백질 설계를 이용한 크리스퍼 유전자 편집의 난제 해결.

데미스 카사비스의 비전은 대부분의 질병을 치료하고 개인 맞춤형 의약품을 개발하는 것입니다.