WordPress database error: [Duplicate entry '0' for key 'wp_postmeta.PRIMARY']
INSERT INTO `wp_postmeta` (`post_id`, `meta_key`, `meta_value`) VALUES (62, '_oembed_5abce3d76b8eeeadd969a54867849088', '{{unknown}}')

AI-Powered Design of Functional Proteins: Unlocking the Potential of Protein Engineering – AI Ohool

AI-Powered Design of Functional Proteins: Unlocking the Potential of Protein Engineering

Now AI can be used to design new proteins

Above: (c/o ISTOCK.COM and CHRISTOPH BURGSTEDT

Artificial intelligence algorithms are having a dramatic impact on the structure of proteins. For example, DeepMind AlphaFold2 was able to predict the structures for 200 million different proteins. David Baker and his University of Washington biochemists have now taken AI-based protein folding a step forward. In a Nature article from February 22, the team described how they used AI for designing functional, tailor-made proteins they could synthesize in living cells. This opened up new possibilities for protein engineering. Ali Madani is the founder and CEO of Profluent. The company uses AI technology to create proteins. He says that this study has \”gone the distance\” and that we are now witnessing a \”burgeoning new field.\”

The protein is made of different amino acid combinations linked in folded chains. This produces a limitless variety of shapes. Humans cannot predict a protein’s 3D shape based solely on its sequence. This is due to the many factors that influence protein folding. These include the length and sequence of amino acids in the biomolecule, its interactions with other molecules and the sugars attached to its surface. Scientists have been determining protein structure using experimental techniques like X-ray Crystallography for decades. This technique can resolve folds of proteins in atomic details by diffracting the X rays through crystallized proteins. These methods can be expensive, time consuming and require skillful execution. Scientists have used these methods to solve thousands of protein structure problems, creating data that can be used to train AI algorithms for determining the structures of other molecules. DeepMind demonstrated with AlphaFold that machine learning can predict a protein structure from its amino acids sequence. They then improved their accuracy by training AlphaFold2 to 170,000 protein structures.

Source:
https://www.the-scientist.com/news-opinion/now-ai-can-be-used-to-design-new-proteins-70997


Comments

Leave a Reply

Your email address will not be published. Required fields are marked *