[ad_1]
Proteins, like those presented in 3D computer models, are extremely complex biological molecules.
Mohammed AlQuraishi / Harvard Medical School
Protein folding has been a difficult computer problem for decades: how do you determine the exact structure of these massive molecules defined by our DNA? The artificial intelligence now brings us to the answer much faster.
Mohammed AlQuraishi, a biologist at the Harvard Medical School's School of Medicine, used the latest machine-learning technology to detect well-understood structural protein patterns and then apply it to other proteins.
The results, even if they are not accurate enough for protein folding applications, such as the discovery of new drugs, are at least a million times faster than conventional computer techniques. And this is just a preview of a technology that can be improved and combined with other modeling techniques.
It's an illustration that AI, although sowed with fears about effects like encouraging police states or suppressing human jobs, can potentially improve medicine, among others.
Mohammed AlQuraishi of Harvard Medical School has developed an artificial intelligence technique to predict the formation of crucial biological molecules called proteins. As its model improves, the colored prediction progressively approaches the actual protein structure shown in gray.
Mohammed AlQuraishi / Harvard Medical School
"We now have a new perspective to explore protein folding," AlQuraishi said in a statement released on Wednesday. "We are just starting to scratch the surface."
Today, AI is more often referred to neural network technology based on the human brain. It revolutionizes everything from voice commands to facial recognition, to debugging software and wiper activation. Artificial intelligence models learn models from actual training data, which means that humans do not have to give specific instructions, like trying to define what this looks like when someone says "Alexa, what's the weather like today?"
In humans and in all other forms of life on Earth, DNA strands contain instructions on how to assemble amino acids into long chains that become proteins. The laws of physics determine exactly how these chains shrink into compact bundles, with the resulting surface structures essential for protein interactions within cells.
But modeling exactly how it will happen inside a computer quickly becomes difficult for larger proteins. This means that it's hard to understand what's going on with proteins. AlQuraishi, however, believes that the artificial intelligence technique could not only help with this understanding, but could also be used to design new proteins that perform specific work.
The results of AlQuraishi were published Wednesday in the journal Cell Systems.
Look at this:
AI in health care: the future of medicine
5:21
[ad_2]
Source link