This Discovery Just Won the Nobel Prize in Chemistry

These scientists just won a shared Nobel Prize for solving “one of the most important issues of modern science.” Here's the issue they were solving and why it's such a big deal...

If you like optimistic science and tech stories, subscribe for our show Huge If True!

#shorts #animation #science #nobelprize #alphafold #chemistry

[Mr.Rostaski]

It's cool to see how machine learning is helping us a lot. Those scientists deserved it.

[Kittieseva]

The 3rd guy who won the shared prize figured out how to design new proteins for specific functions, which is arguably cooler in my opinion. It's another problem we thought we'd never solve

[eantropix]

Finally we can have the argument: "Oh so you're a microbiologist? Name every protein"

[notreallydavid]

Biochem degree here - I was surprised that more fuss wasn't made of this when the news originally came out. 3D structure prediction was still looking dauntingly hard not so many years ago.

Still need to validate predicted structures using X-ray crustallography - I think we'll always have to check proposed structures experimentally (or until prediction necomes consistently spotless), and the program can't handle multi-protein complexes yet.

[ulyssesyee]

As someone who studied genomics and proteomics; This is a very big deal! We have literally deciphered nature’s algorithm.

[FlubberGamer]

This is the type of stuff that AI or machine learning systems are not only great at, but what we should be using them for. Processing large amounts of data and information and finding connections or patterns is something we can do, but not nearly as well and as fast as computers can. I’m excited to see more stuff like this that could potentially make a genuinely positive impact in the world.

[anomalocaris]

I'm literary building a mutation database using alphafold, when in my undergrad they told us it was an impossible problem, and now, it's running in my laptop.

[ChaserX]

Big shoutout to all those in the X-ray crystallography field and their contributing research that made this possible.

[DucksArCut]

It’s so cool how the things we learned in the classroom as hypotheticals are becoming a reality!!! Genomics and proteomics is so fascinating and this is definitely helping revolutionize biotechnology/bioinformatics

[urvalpatel]

You are correct, I am biochemist very happy that they have won Nobel Prize in chemistry . Although one thing as biochemist, I have learned that there two major classes of proteins, ordered and disordered .Alphafold is very good for predicting ordered (hemoglobin)structure , when it comes to disordered structure such as (transcription factors )it’s not very best at it . There is still a lot to learn about protein particularly dark proteome which is -90 % and not known to our current knowledge..

[AutkastKain]

I remember also seeing a community project where they made this into a puzzle game to help the learning model learn how to put things together.

[JamesMiller-lbsk]

She brings so energy to her content and it’s great

[figo]

Imagine priests from the medieval times being informed by a time traveler about this, they would be like: *this is a new breed of heresy*

[lovingquesadillas]

Omg this is so awesome!!

About 10 years ago when I was in tenth grade I did an internship at a lab which was doing macromolecular crystallography to determine the structures of proteins. An incredibly time consuming work process, from isolating the protein, growing the crytsals and analyzing the x-ray results!

I can only image how excited the team was when they heard the news!!

[shayatz]

Amazing, this has huge potential! Props to everyone involved in the research.

[parthparekh]

This is genuinely the most craziest and amazing application of machine learning and ai. This is the kind of breakthroughs we ought to endeavour more for. Honestly its hard to overstate the gravity of what has been achieved here !

[thetoxbloxer]

this is what AI should be for, not replacing us in creative fields

[alikhan]

I studied Chemical Engineering with an emphasis on Bioorganic Compounds remember finding out the 3D structure of hemoglobin (I think it was the first one they did). It was mind blowing because of the way the protein folds using non-covenant bonds. The number of calculations that were required were so unbelievably vast that I just couldn’t wrap my head around the fact that a computer did it. I loved what I went to school for with passion but I couldn’t stand big pharma and now I own a construction business lol. Just remember, you go to school to learn how to learn; so learn how to learn and you’ll be successful in life.

[envylambird]

This was the biggest issue that was drilled into me in my Bioengineering classes. Not being able to predict protein structures based off amino acids meant we had to brute force test them based on how they bonded to other proteins, or the x rays mentioned. AlphaFold came out the moment my classes ended, but garunteed my professor is giddy asf every day now. She was obsessed with this topic haha

[nonotreallythx]

It didn't just go from X-rays to AlphaFold. We've had software modeling for protein folding for decades. The issue is that, until these A.I. breakthroughs, finding a protein structure that would fold the way you want has been an NP-hard problem.