AlphaFold2 Explained | Google's DeepMind Solves Protein Folding
Вставка
- Опубліковано 29 тра 2024
- 🔥 Get my A.I. + Business Newsletter (free):
natural20.com/
#google
#deepmind
#ai
00:00 Advancements in AI
00:18 What is Protein Folding?
01:10 DeepMind's AlphaFold
01:32 Why is this a problem?
03:32 How will this advance science?
We need more videos like this so i can show friends that AI is not all matrix and terminator scenarios.
No shit. "Curing all diseases" and "living to 200" has a lot more appeal than "robots killing us all/mass unemployment/disinformation/blah blah blah", the same shit we've been hearing since Frankenstein.
You need new friends 😂
😂
Thanks! Great content! 🙏🏼
Now to figure out what I can do with this.
better crystal meth
I love it
AI just tried to draw a hand, we miss took it for a protein folding....
Meta ESM-2 Fold - AI did 600 million
does that mean we can create better Crystal Meth thanks to Googles Alphafolding?
We could create extreme drugs beyond our understanding.
maybe drugs that make us smarter, or to aid genetic engineering that creates a new race of super-humans.
tell us the results!
This explained nothing about alphafold just protein folding in general
Misinformation. It's not "200 million proteins' but two million. Other than that, for a UA-cam video where you can get the same information in a few seconds Bing'ing it rather than spending five minutes of your life viewing it, "nice work'. bye.
Internet:
It is estimated that the human body may contain over two million proteins, coded for by only 20,000 - 25,000 genes. The total number found in terran biological organisms is likely to exceed ten million, but nobody knows for sure. Data is available on just over a million of them, taken mainly from information found in the over 100 genomes that have been fully sequenced.1 For a budding yeast cell of 40 m3, the two estimates give a range of 90-140 million proteins per cell. Extrapolating these protein densities to mammalian cells, a value of about 1010 proteins per cell is predicted for characteristic cell lines that have average volumes of 2000-4000 m3. DeepMind's AlphaFold tool has determined the structures of around 200 million proteins.
According to DeepMind's website:
"Currently, there are over 200 million known proteins, with many more found every year. Each one has a unique 3D shape that determines how it works and what it does."
They also link to:
www.uniprot.org/
"UniProt is a freely accessible database of protein sequence and functional information, many entries being derived from genome sequencing projects. It contains a large amount of information about the biological function of proteins derived from the research literature"
Looks like they have over 250 million proteins in the database.
@@WesRoth Shows you can't believe everything you read on the 'net. What they mean is that they have 200 solved proteins, out of the universe of solvable proteins that in theory are more than the atoms of the universe, but in fact humans have 1% of 200M or 2M proteins like Wikipedia says. PR for gullible VC types playing with OPM. Maybe Kathy Wood would be interested, but she might be more savvy than that.