The use of audio for Myrtle Rust DNA sequence analyses

DNA carries genetic information for the development and function of organism such as humans, plants and Fungi. A DNA sequence is a long, continuous chain made up of only four chemical bases referred to as G, A, T, or C. They repeat in various defined patterns to make up a gene.

Many genes are identical from person to person, plant to plant or from fungus to fungus. But sometimes one of the chemical bases in the sequence is different from the usual pattern – could be a mutation. These changes could create advantages so that the fungus can infect other plant species.

The process of gene expression from the Myrtle Rust genome has been simulated by a computer algorithm. The output of this analysis is both a visual and audio display as shown below. The use of audio for DNA sequence analysis is a process called Sonification.

Can the science audio be musical?

In my more recent sonification work, I made the effort to make the audio more tuneful and harmonious so that it could be listened /analyses for longer periods of time without fatigue. On reflection, I was quite taken with the musicality of the audio and decided to have a go at writing some music to accompany it.

This has taking me down a whole new research path to think about the science of music. How does music work, how do some things stand out and how do others blend into the overall sound? I thought about this whilst I was writing computer code to convert the DNA sequences into audio. It turns out that musical thinking improves the outcomes of the science research, and this allowed me to blend multiple layers of audio (and therefore more layers of data) into a single audio track.

In this video I show the first step towards making music from sonified audio. I find that adding drums to the audio gives a clear guide for the other musicians. Once we have something that was resembles the barebones of a song structure we can begin to add other instrumentation and create more diverse music.

As you can hear in this example the initial attempts as playing to the sequence are a little chaotic since there are competing accents within the audio data which can be disorienting for the musician.

The science audio can also be mapped to a variety of other instrument sounds. Here is an another example of a finished section of science audio and drums that is starting to sound much more musical. Everything you here except the drums is audio derived directly from the Myrtle Rust DNA sequence.

musification and Science outreach

The outcome of this drift into musification of the genomic data has been a series of ensemble performances of tracks from the CoronaCode Music album and new tracks based on sonification of the Myrtle Rust genome.

Media quotes

“this is just amazing” NPR
“beautiful and ethereal music” New Scientist
“turned the coronavirus genome into a musical masterpiece” Science Line
“as cool as it sounds” 3MBS Fine Music
“surprisingly lovely music” The Conversation
“a unique composition” Limelight magazine
“surprisingly chirpy, certainly melodic” The Wire (2SER)
“I friggin love the music” Jules ex Phantom Records

When I started generating sounds from biological sequences, I was very clear to point out that sonification was producing “audio” for data analysis rather than “music” as quirky artistic expression.

As a scientist I’ve tried to keep these activities separate since I want the sonification (the science audio) to be taken seriously by the scientific community. I think I’ve managed to achieve this since I have published a couple of peer reviewed journal articles describing the technical characteristics of the audio and its use for DNA sequence analyses.

In this current project I have used the raw computer generated science audio as if it were music made by a guest musician in a rehearsal room. The challenge for us live musicians is to play along to this strange computer generated audio.

The science audio itself has a shifting sense of rhythm, abstract and non-repeating melodies and it lacks traditional song structure. As musicians our goal was to make it more musical. We added compositional elements such as introductions, verse like sections, bridges and various musical dynamics to match the beginning and end of genes as well other sequence motifs in the data.

Surprisingly, this inspired us to play in unusual key signatures, at tempos we were not familiar with, and to play unusual song structures. I don’t want people to listen only because it contains science audio,

I want these track to sound good as pieces of music in their own right.

If you were to hear these tracks in a playlist without knowing anything about them, would they sound like music?