Turning Proteins Into Music

IRA FLATOW: This is Science Friday. I’m Ira Flatow. Later in the hour, the fascinating history of paternity science, but first, proteins. They are the building blocks of life. They make up cells, enzymes, skin, bones, hair, spider, [INAUDIBLE], conch shells– anything that is alive. But it’s notoriously difficult to understand the complex shapes and structures that give proteins their unique identities. So at MIT, researchers are unraveling the mysteries of proteins using a more intuitive language– the language of music. 

[MUSIC PLAYING] 

They’re translating proteins into music like this, composing orchestras of amino acids, concerts of enzymes, in hopes of better understanding proteins and making new ones. Markus Buehler’s research in creating this music appeared in the journal ACS Nano this week. He’s a materials scientist and a professor of Civil and Environmental Engineering at MIT. Welcome to Science Friday. 

MARKUS BUEHLER: Hi. Hello, Ira. Thanks for having me. 

IRA FLATOW: What made you look at a protein and say, you know what? I see music. 

MARKUS BUEHLER: Well, proteins, as you mentioned, form these amazingly different kinds of materials. And what’s common in proteins and all the different materials you’ve mentioned is that they form hierarchical structures that go from the molecular all the way to the macro scale. So they form little sheets, or helices, which assemble into bundles, which assemble into fibers, and so on and so on, until we go to the macro scale. 

And in music, we see something very similar. If you think about the sound of an instrument, the timbre of an instrument that’s really created by an overlay of many different sine waves, which are modulated in time, and then you can play with an instrument. You can play a melody. Usually, you can play multiple melodies together. You can play chords, and you can play longer sequences. And you can sort of build up a musical piece to understand music from a similar perspective and that is hierarchical, built up from small building blocks, and forms complex large scale systems, just like proteins. 

IRA FLATOW: So OK, say I want to translate a protein into music. Give me the steps that are involved. What’s the first step? 

MARKUS BUEHLER: Sure. So first, we realized to do this, it’s not as straightforward as one might imagine, if we want to do this physically sound or chemically sound. So we realized in this process. So in addition to conceptually realizing the similarity between proteins and music, and there might be some connection, the key to understanding this translation or doing this translation was really that we realized that matter at the nanoscale is always moving and vibrating. 

So if you take a microscope– first, well, actually, if you open your chemistry textbook, you’re going to find a picture of a molecule in there, and it’s static, right? It doesn’t move. But if you were to take a real microscope, and you look at a molecule, you see the chemical bonds are always moving and vibrating. And so what we did is we computed these vibrations using quantum mechanics, which are the physical laws or chemical laws that govern mechanics and behavior of systems at that scale. 

And we then could compute a spectrum of vibrations for each molecule. And we detected that each of the building blocks of proteins, which are the amino acids, has a unique frequency spectrum, which we could then make audible using a concept of transposition. And that way, we can begin to hear how these proteins sound like. 

IRA FLATOW: Well, let’s listen to them. Let’s play one of these protein songs. This one is called “Concert of Silk and Amyloid.” 

[MUSIC – “CONCERT OF SILK AND AMYLOID”] 

Wow. I’m hearing it be– let me play it under while we’re talking. Where does that rhythm come from? 

MARKUS BUEHLER: Sure. Yeah, so what we did this in this piece is so every protein that you might find– and there are millions, and millions, and millions of proteins out there in nature– has a particular sound if you do this translation in that way. And what we did in this piece, we actually combined the soundings and rhythms we detected from a silk protein and amyloid protein and combined them in a piece of music where we basically have the human brain composing out of the spaces of sounds that are created by nature through these natural vibrations of these individual amino acids in the proteins as a whole. 

And what you heard is really a collection of sounds that solely come from these very, very basic quantum mechanical vibrations. 

IRA FLATOW: All right, this is fascinating. Let’s listen to “Simple Amino Acid Beat and Melody.” 

[MUSIC – “SIMPLE AMINO ACID BEAT AND MELODY”] 

All right, what are we hearing? Tell us. 

MARKUS BUEHLER: What we’re hearing here is a little different. So what we did in the previous piece, we had the actual real protein as a basis and created songs or compositions from that. In this piece that we just heard, we only took the basic soundings of amino acids as a basis and actually made up our own proteins, if you wish, from these sounds and composed within this the tones available. 

And because there are 20 unique amino acids in nature and biology, we have 20 unique sounds that we can play with. So the scale we were playing with here isn’t the C major or C minor scale or anything like this. It’s actually a 20 note amino acid scale that’s unique to these materials. And that’s what we played with in creating this composition that we just heard. 

IRA FLATOW: Now I understand that after you translate the proteins into music, you let computers, meaning artificial intelligence systems, listen to the music. What’s the point there? 

MARKUS BUEHLER: Right, so as we translated proteins– we translated hundreds and thousands of these different proteins– we listened to them. I created our own compositions. And one of the really difficult things is to understand how do we actually relate what the sound [? flagged ?] with the function of a protein, or how the protein will fold, or whether a protein is a disease protein or a healthy protein, or how it assembles. 

And even though the human ear is very good and actually is very good at detecting patterns, we’re not very good at detecting and understanding this new type of a sound and this new type of functionality that’s expressed through music in these sonifications. So we decided to feed all these sonifications, these sounds, to a computer, to an AI, an artificial neural network. And instead of using the human brain and detecting what this protein sound really means, having a computer learn from all the data what these proteins look like and how they function. 

IRA FLATOW: And I understand that there’s a practical application that you are thinking about it. You’re doing some research on communicating with spiders with these sounds. 

MARKUS BUEHLER: Yeah, right. 

IRA FLATOW: Tell us about that. 

MARKUS BUEHLER: Yeah, so in general, what the artificial neural network, the AI, can do very well is detecting patterns and learn languages. And the proteins, we have, through these sonifications, learned the language of what determines how a protein will fold, and whether it’s a disease protein or a healthy protein, and so on. And we’re applying a similar way of learning from sounds to other systems. This is work we’re doing with collaborator [? Tomas ?] [INAUDIBLE], an artist in Berlin who’s a visiting resident artist at CAST at MIT. 

And [? Tomas ?] and I have started working on this idea of thinking about spiders as sound creators. A spider is actually a very obvious choice for thinking about the connection of sound material and living organisms, because they use sound in detecting how they’re building a web. They’re detecting other spiders in the web. They’re detecting prey in the web. And we’re trying to do there is to really take these different sounds that the spider hears and creates by actually vibrating the web. The spider uses the web as an instrument, if you wish, for communication. 

And we don’t understand this language. So we are doing the same similar idea that we’re doing for the proteins here at the macro scale in the case of the spider and trying to learn what is the spider trying to tell us with these different sounds. And for us humans to understand [INAUDIBLE] actually then feeding these sounds back to the spider web and trying to– this is the vision for this work that we’re trying to achieve, is to come up with a way of communicating with a spider. 

All right, so one of the interesting things– one of the things I’m very interested in, in general, is how do we communicate across scales? How do we make quantum mechanics audible, like in the case of proteins? How do we communicate across species? How can we, as humans, communicate with spiders? And how can we translate insights from different fields, like music, into material science, and so forth. 

So in the spider case, the real direct application is that we’re trying to be able to communicate with a spider and see how the spider responds. And if this works, it would be very exciting. It would be the first time we have a Google Translate, if you wish, app for communication between different species. 

IRA FLATOW: Do you have any reason to believe that the spider knows how to listen to what you would be sending it? 

MARKUS BUEHLER: Yes. Well, the spider relies very heavily on vibrational signals in its communication because it’s essentially blind. So when the spider builds the web, or detects prey, or detects other spiders or any kind of activity in the web of communicating, mating, and so on, the spider will actually use vibrational signals. So it’s very sensitive to vibrations. So we think that when we start to speak that language by actuating a web and pretending to be a spider, but actually using the right kind of frequencies, the right kind of spectra, the kind of language, we might be able to do that. 

IRA FLATOW: I’m worried– I’m not worried– I’m wondering. I’m wondering about your passion with music. And you use as an engineer, and in physics, and in designing these materials. Where does it come from? 

MARKUS BUEHLER: Well, yeah, I think I’m very interested in the translation of insights between different domains. And we have realized, as we said in the beginning, these systems across different manifestations have very similar construction principles, basically. And music has a similar construction– 

IRA FLATOW: No, no, no. I know I asked you the question, but I don’t think– you might not have heard me. You might not have answered it. When you discovered that proteins can be made to play music, you must have some sort of musical background, right? That you understand music, and you must have had– 

MARKUS BUEHLER: Oh, yes, sure. 

IRA FLATOW: Were you ever a rock musician or something in your childhood? 

MARKUS BUEHLER: [LAUGHS] No. No, no, no. But I’ve always been very interested in music, and sound, and compositions. And so it’s actually, it was a natural connection between the different things I’ve been doing in my professional career. 

IRA FLATOW: Mhm. And this will be just an extension of it. 

MARKUS BUEHLER: Yes, correct. 

IRA FLATOW: So when we talk about putting arts in STEM, this really would be your STEAM project. 

MARKUS BUEHLER: Right, correct. Well, yeah, I mean, a lot of these different areas are quite connected. And I think you mentioned STEM or STEAM outreach. I think it’s a terrific example of how we can teach about or make connections about something everybody [? knew about ?] which is sound, and music, and piano, and explaining how the piano that nature uses to create complex proteins and function has only 20 keys. And with these very simple 20 keys, these amino acids, nature builds amazing machinery, amazing materials. 

And that’s something very similar, [? as ?] humans have actually created complex systems out of universal building blocks– music, right? 

IRA FLATOW: Yeah. 

MARKUS BUEHLER: For thousands of years. Ever since humans have been around, they’ve been creating the systems. 

IRA FLATOW: Well, you’ve explained it very well, and we want to follow you on your communications with the spiders. You stay in touch, OK, Mark? 

MARKUS BUEHLER: Thanks so much. 

IRA FLATOW: You’re welcome. 

MARKUS BUEHLER: Thanks so much, Ira. 

IRA FLATOW: Good luck to you. Markus Buehler is a materials scientist and professor of Civil and Environmental Engineering at MIT.

Copyright © 2019 Science Friday Initiative. All rights reserved. Science Friday transcripts are produced on a tight deadline by 3Play Media. Fidelity to the original aired/published audio or video file might vary, and text might be updated or amended in the future. For the authoritative record of Science Friday’s programming, please visit the original aired/published recording. For terms of use and more information, visit our policies pages at http://www.sciencefriday.com/about/policies/