Google DeepMind’s AI AlphaFold protein model for better medicines
Human beings can potentially receive more targeted treatment with Artificial Intelligence based protein folding prediction 3D modelling enabled by AlphaFold AI by DeepMind, a Google company. Biologically, proteins are the main components of our physical body along with fat and water. There are 20,000 to over 100,000 unique types of proteins within a typical human cell as per a blog on Harward University website. Each performs a specific task. Proteins folding is a phenomenon which determines the health of the body. If it folds properly the body remains healthy and if not a living being can even face life-threatening problems. Biologists have been trying to model the proteins in order to arrive at better medicines and therapies.
As per DeepMind, AlphaFold focusses “specifically on the problem of modelling target shapes from scratch, without using previously solved proteins as templates. We achieved a high degree of accuracy when predicting the physical properties of a protein structure, and then used two distinct methods to construct predictions of full protein structures.”
Prior to AI modelling, biologists studied the shapes of proteins in labs using experimental techniques like cryo-electron microscopy, nuclear magnetic resonance and X-ray crystallography. These methods depend upon a lot of trial and error, making the process expensive and time-consuming.
” The 3D models of proteins that AlphaFold generates are far more accurate than any that have come before – marking significant progress on one of the core challenges in biology,” claims the company literature available online. AlphaFold is deployed at CASP13 (13th Community Wide Experiment on the
Critical Assessment of Techniques for Protein Structure Prediction) after winning a competition against RaptorX-Deepmodeller by Toyota Technological Institute at Chicago, Quark by the Zhang group (University of Michigan) and Cheng’s human group MULTICOM. Critical Assessment of protein Structure Prediction (CASP) experiments focus on protein structure prediction. AlphaFold is available on Github. Mohammed AlQuraishi, a computational biologist at Harvard Medical School in Boston, has an AI program which is claimed as one million times faster at predicting structures than AlphaFold but may not as accurate in all situations.
The cells themselves have natural abilities called Chaperones and proteasome to cope up with faulty protein folding. AI can help in understanding why some cells are able to evade the natural process and aid biologists in targeted cures.
The Indian Space Research Organisation (ISRO) on 14th July announced the successful conclusion of its third long-duration hot test of the liquid-propellant Vikas engine...
Samsung Heavy Industries Co. (SHI) said on Wednesday that it has teamed up with the Korea Atomic Energy Research Institute (KAERI) to develop nuclear-powered...
Humans beings will retain their edge over artificial intelligence (AI) for some time yet, say scientists at Plekhanov Russian University of Economics (PRUE). The...