Jonathan Hus Week 4 - Favorable Folding

 

    I recently read a fascinating article about the folding of proteins and the Artificial Intelligence applications being used to predict these protein complexes. Today, I would like to share with you what I learned about this newfound technology. Proteins, a sequence of amino acids bound together by peptide bonds, are among the most complex structures observed throughout nature. These microscopic structures can be used and are seen in all aspects of life from vaccines, to antibodies, and even blood ph regulation. They can be used to cure diseases or even clean up toxic waste if engineered correctly. This is the newfound problem being tackled by scientists and researchers around the world.

    The folding of a protein is what allows it to function or become biologically active and therefore serve its intended purpose. Every day, researchers are trying to engineer new proteins to accomplish different tasks but the main issue is figuring out the correct shape of the protein to actually be functional in nature. Today, AI and Deep Learning tools are being developed and used to predict, create, and test novel proteins. Alpha Fold, a revolutionary protein folding technology uses artificial intelligence and deep learning to predict the most likely structure of a protein in its biological form. This application has greatly advanced the world of protein fabrication with a success rate of 1 of every 5 protein rendering resembling the predicted shape. Researchers recently have used Alpha Fold in succession with a similar program called RoseTTAFold to create protein chains similar to that in the Covid vaccine. However, when tested, none of the 150 designs actually worked. This led researchers to begin tackling another problem known as inverse folding. Inverse Folding is determining the sequence of a protein from its shape rather than figuring out the shape from the sequence. A new program called ProteinMPNN does exactly this. It can act as a sort of "spell checker" for protein manufacturing. When used with AlphaFold, test results have shown a 9 out of 10 success rate in protein structure prediction. 

    While this data is very promising, there is still a great area for improvement and perfection. I believe that this bioengineering of proteins will help many people in our generation and I look forward to seeing the next steps in this protein research. Let me know what you guys think about this upcoming technology and how you think it could benefit our lives.