Composers string notes of different pitch and duration together to create music. Similarly, cells join amino acids with different characteristics together to make proteins. Now, researchers have bridged these two seemingly disparate processes by translating protein sequences into musical compositions and then using artificial intelligence to convert the sounds into brand-new proteins. They report their results in ACS Nano. Watch how they did it here.
To make proteins, cellular structures called ribosomes add one of 20 different amino acids to a growing chain in combinations specified by the genetic blueprint. The properties of the amino acids and the complex shapes into which the resulting proteins fold determine how the molecule will work in the body. To better understand a protein’s architecture, and possibly design new ones with desired features, Markus Buehler and colleagues wanted to find a way to translate a protein’s amino acid sequence into music.
The researchers transposed the unique natural vibrational frequencies of each amino acid into sound frequencies that humans can hear. In this way, they generated a scale consisting of 20 unique tones.
Unlike musical notes, however, each amino acid tone consisted of the overlay of many different frequencies –– similar to a chord. Buehler and colleagues then translated several proteins into audio compositions, with the duration of each tone specified by the different 3D structures that make up the molecule. Finally, the researchers used artificial intelligence to recognize specific musical patterns that corresponded to certain protein architectures.
The computer then generated scores and translated them into new-to-nature proteins. In addition to being a tool for protein design and for investigating disease mutations, the method could be helpful for explaining protein structure to broad audiences, the researchers say. They even developed an Android app to allow people to create their own bio-based musical compositions.