Researchers at MIT have developed a model to convert proteins into sound. After, by making variations into the music and converting it back to proteins, new proteins are made that have never been seen before. Researchers were able to convert the proteins into music by sequencing a protein's amino acid sequence and translating it into a musical score. Specifically, the model converts the twenty types of amino acid into a 20-tone scale. To translate the music back into proteins, the researchers developed an AI system that could introduce variations into the musical sequence and convert it back to a new protein. For example, researchers were able to translate silk protein into music, and then the AI changed the music and presented back a new silk protein never seen before. One major drawback of the system is that it does not allow for direct modifications. The AI does its own thing, so the properties of the new proteins are completely random to the researchers.
What I find most fascinating about the researchers' methods are that the musical tones are based off of the vibrational frequencies of the amino acid sequence. We are used to looking at a textbook and seeing proteins as motionless and flat, but we have to remind ourselves that proteins are vibrant and moving. Another aspect I found interesting in the article is that the researchers were able to hear the melodies and could differentiate what the sequences were. For instance, the researcher could hear a sequence and definitively tell that he was hearing an alpha helix or a beta sheet. The method of translating proteins into music is creative, and I believe it will have major applications in the years to come.