is evolving rapidly. AI has now mapped nearly every known protein of every living object and google has released it as a public data base. I'm not smart enough to understand all of the downrange applications of this knowledge but I have to assume it is immense. DeepMind found the structure of nearly every protein known to science (msn.com) DeepMind is releasing a free expanded database with its predictions of the structure of nearly every protein known to science, the company, a subsidiary of Google parent Alphabet, announced today. DeepMind transformed science in 2020 with its AlphaFold AI software, which produces highly accurate predictions of the structures of proteins — information that can help scientists understand how they work, which can help treat diseases and develop medications. It first started publicly releasing AlphaFold’s predictions last summer through a database built in collaboration with the European Molecular Biology Laboratory (EMBL). That initial set included 98 percent of all human proteins. Now, the database is expanding to over 200 million structures, “covering almost every organism on Earth that has had its genome sequenced,” DeepMind said in a statement. ............................. Alphabet continues to build on the success of AlphaFold. Alphabet launched a company called Isomorphic Labs that will aim AI tools at drug discovery, and while it’s separate from DeepMind, the two companies will collaborate. DeepMind also set up a lab at The Francis Crick Institute, where researchers can run experiments testing the information from the AI system. Having easy access to predicted protein structures gives scientists a boost in research efforts across the scientific landscape — like those trying to understand how complex processes work in the body or which molecules can be used to target things like pollution. “With this new addition of structures illuminating nearly the entire protein universe, we can expect more biological mysteries to be solved each day,” said Eric Topol, founder and director of the Scripps Research Translational Institute, in a statement.
This is perhaps the biggest success of machine learning to date. The protein folding problem had been sitting there for a while without anyone being able to solve it. The downstream utility will be rather unpredictable. The discovery of green fluorescent protein has produced myriad technologies that no one saw coming. Meanwhile, the human genome project still hasn’t even scratched the surface of what we predicted it would do.
This stuff is all amazing, but the need for more real time fixes is a missed opportunity in science at times. It seems that infallible quick testing of masks for instance should've been available 2 years ago. Yet we still have scientific debate about how well different masks/double masking etc works It's so crazy how we can now get pictures a gazillion light years away and map every protein, but are still unsure if a cloth mask does much. Lol. (Not a political point btw) Similarly, I dont get how we cant explain ball lightning or many other things. It's so wild and amazing and frustrating and beautiful all at once.
Related topic from just two days ago: some scientists are worried that the methods of machine learning are going to result in many false positive results. Could machine learning fuel a reproducibility crisis in science?
Anything which significantly prolongs life is a mixed bag. If one stays healthy and active, one will want to continue living, but stop working at an age to allow successive generations to enter the work force. That to me in where the bag is mixed. Is a long lifespan just for the wealthy?
To me it goes well beyond this amazing accomplishment. AI and advances in quantum computing capacity combined with rapid advances in physics and genetic tools are prime to grow exponentially all feeding off of each other over the next 10 - 20 years. Energy, medicine, materials etc should all look substantially different 20 years out if this plays out the way I hope it will. Open sourcing all this data by google is worthy of praise unless there is something Im missing
I hope I am gone long before any of this becomes real. I couldn’t imagine going through “life” without the limitations of a finite lifetime. Carpe Diem versus a meaningless life of drug induced euphoria.
Yes sorry, I’ve been locked away from humanity in a hotel with COVID for several days. Machine learning / AI is a cool method of finding patterns that uses a somewhat brute force approach that is simply unavailable to humans. It can find things correlated in ways people just would never look for. Meanwhile, the problem of protein folding has proved just such a challenge. It can be viewed as a kind “code” where each sequence of protein chain can be translated into a folded shape of the protein. The proper genetic code, from nucleotide sequence to protein sequence, took a long time to figure out but ended up being quite simple. The next step, protein sequence to folded shape, just wouldn’t solve. Then this damn deepmind came along and very quickly found hidden patterns that basically predict the folded shape of any protein sequence imaginable. It’s incredible. What’s the full value of this? It will definitely enable some forms of research, but I can’t say much more. To give an example of how hard it is to predict the future utility of a biological discovery, I gave two contradictory examples. One is a single protein that fluoresces green, imaginatively called green fluorescent protein. It was originally found in a jellyfish. A few scientists later, it was genetically inserted into a worm to make different cells glow. This was just to visualize which parts of the worm had certain genes active. It ended up winning the Nobel prize because it turned out to help all kinds of scientists solve problems, biological and other. Just a simple single protein. Meanwhile, we were all promised in 2000 that we’ve just unlocked all the molecular secrets of humans by sequencing the human genome. Soon we’d eradicate diseases and enjoy an era of personalized medicine. Instead, we’re still just staring at these giant sequences quite unclear about what almost all of these “genes” do. Not that sequencing the genome wasn’t a good idea. It’s just that the complexity of the situation was orders of magnitude beyond our puny predictions. The head of the genome project thought autism might be caused by 15 genes. Thousands is probably a better guess…for basically all traits, height, IQ, maybe even eye color.
more stuff I don't understand but seem to be the building blocks needed to build something much better. Physicists Create New Phase of Matter With "Extra" Time Dimension (futurism.com) Physicists say they've created a new phase of matter with an "extra" time dimension, according to a recently published study in the journal Nature. The researchers didn't set out with the goal of opening a "portal to an extra time dimension," as Scientific American described the experiment. They simply sought to create a new topological phase of quantum matter, which is essentially a form of quantum matter defined by the dynamic motions of its constituents, rather than their static arrangements. ........ In order to successfully create the topological phase, and thus the "extra" dimension, the scientists targeted a quantum computer's quantum bits — or qubits — with a quasi-periodic laser pulse based on the Fibonacci sequence. Think quasicrystal. "The Fibonacci sequence is a non-repeating but also not totally random sequence," study co-author Andrew Potter, a quantum physicist at the University of British Columbia, told Vice. "Which effectively lets us realize two independent time-dimensions in the system." The effect only lasted for a few moments, but the researchers say it's still a significant breakthrough. Fascinatingly, they believe that their findings could be a major step towards developing a complex information storage system for quantum computing with limited error — although, as the scientists themselves admit, that's still a long way off. "We have this direct, tantalizing application, but we need to find a way to hook it into the calculations," Dumitrescu said in a press release. "That's an open problem we're working on."
and an intriguing study about cancer prevention. got to double down on my green bananas Study Suggests That Pill Can Cut Hereditary Cancer Risk by 60 Percent (futurism.com) A years-long trial appears to have shown that resistant starch — a digestion-resistant molecule found in a range of everyday foods including slightly underripe bananas, oats, peas and beans, rice, pasta and more — has an astonishing ability to help prevent a range of hereditary cancers. Published in the journal Cancer Prevention Research, the double-blind longitudinal study followed nearly 1,000 patients with Lynch Syndrome, an inherited genetic condition that increases the individual risk of several cancers, for almost 20 years. According to a press release, the researchers at the universities of Newcastle and Leeds placed participants in two groups back in 1999. There was a short treatment period thereafter, which lasted until 2005. During the treatment period, each individual received a pill. One group took a placebo, while the other took a daily dose of resistant starch. There was no notable difference in the individuals' health at the end of that short treatment period. But the study was designed for follow-up, and in extremely intriguing results, those who took the daily starch pill — which contained the approximate amount of the digestion-defying molecule that one would find in an average-sized underripe banana — were far more likely to be cancer-free nearly 20 years later. "We found that resistant starch reduces a range of cancers by over 60 percent. The effect was most obvious in the upper part of the gut," John Mathers, professor of Human Nutrition at Newcastle University, said in the press release. "This is important as cancers of the upper GI tract are difficult to diagnose and often are not caught early on."
I’ve done ok in life, but sometimes I read some of these both threads and topics and amazed at the intellect and capabilities far beyond my own.
