AI Solves Protein folding problem – For about half a century, our researchers have had a tough time predicting how proteins come up with their 3D structure. It is a tough job. Few researchers have also claimed that it would take longer than the existence of the universe to predict all the possible sets of molecular arrangements.
AI Solves Protein Folding Problem
Contents
In such a case, even a small development can do wonders in the field of science. It would help make advancements in drugs, pharmacy, making new inventions for disease treatments, etc.
In a recent announcement, this protein folding problem has been solved by AI giant DeepMind. Google’s child company DeepMind claims that it has been able to solve the 5-decade long problem with an accuracy of 92%. To be more specific, DeepMind is an Artificial Intelligence giant, and its AI system AlphFold has been able to solve the riddle.
You must be wondering what protein folding is? Well, protein folding is the science of learning as to how the protein’s amino acid turns its 3D structure. See, a protein made up of hundreds of amino acids. These acids are none other than the building blocks. They called building blocks for a reason. They form the basic entity of the protein. The same protein is present inside every living thing in the universe.
Solution for Protein Folding Problem In Few Hours
The shape of a protein determined by how its amino acid transforms its shape. This shape formed by thousands of such possibilities of transformation of the acids. And determining the accurate probability of the 3D form of one protein can turn out to be hard work for numerous years.
For this reason, our scientists and researchers all around the world going through a hard time solving the problem. And for the same purpose, an experiment called Critical Assessment of Protein Structure Prediction(CASP) was initiated in the 1990s. This was aimed to help scientists develop a system that could easily develop a system that helps in predicting AI Solves Protein folding problem.
After DeepMind’s AlphaFold has been able to bring a solution for years-long problems, now CASP’s efforts seem to be turning fruitful. While testing for the solution, DeepMind used a new architecture that capable of calculating the spatial graph of proteins. This in turn was helpful to predict the 3D structure of proteins.
All this wonder happened in CASP 14. CASP 14 is nothing but this year’s CASP problem-solving challenge. DeepMind’s AlphaFold system built up of a data pool of 170,000 protein structures. The amazing results of CASP 14 were able to leave every researcher present there, in an awestruck situation. AlphaFold was able to gain a median score of 92.4 GDT. These results are above the 90 GDT threshold.
AI Solves Health-Related Problems
Anything around 90 GDT said to competitive in this research arena. GDT is nothing but the short for Global Distance Test. AlphaFold was able to secure an average score of 92.4 out of 100 in deciphering the protein structure for the regular class of proteins. And for the complex protein structure, the AI system secured an average of 87.
The research results are a huge leap in the field of science. It used to treat people with various diseases. For instance, in diseases like cancer, COVID-19 this protein folding solution can be of great help. It can help the researchers to pave a method for treatments and discovering drugs for curing these diseases.
However, it noted that the research is still neither peer viewed yet nor is published in any science journal. On this, DeepMind has said that it is on its way to make it to a journal.