Tuesday 29 October 2013

Science gets weird: Microbial Bebop, creating music from complex dynamics in microbial ecology


'Here, we combine music and biology to generate musical compositions from microbial ecology data.'

A bizarre one, this. A team of American scientists has decided to tackle the issue of ineffective communication of scientific ideas to the general public by taking the only logical course of action, the conversion of complex biological data sets into jazz music. 
Microbial bebop makes use of the patterns seen in microbial ecological data sets and transforms them into patterns of notes and chords to be used in jazz bebop improvisation. This study generated four compositions from data collected by a marine monitoring system in the Western English Channel, each composition (apparently) represents the relationships between microbial community structure and a range of environmental parameters.

Inspired by the lack of scientific awareness displayed by the American public in recent surveys, the authors explain their obligation to explore more effective ways to convey ecological science.  Proposed as an alternative to the graph-and-chart style of data communication, their use of jazz music capitalises on repeated patterns and the complex interactions seen between microbial populations and their environment. It is not the first time music has been used as a possible way to communicate data, amino acid and genomic sequences have both been converted into musical scores, though the linearity of the data makes for less aesthetic compositions (honestly, I’m not making this up). One proposed advantage of musical compositions is the shear amount of biological data they can convey relative to a graph. A picture may tell a thousand words but the authors have calculated there are 6.88×10109 unique combinations possible using standard compositions, offering almost unlimited interpretations of the same data set using different environmental parameters.

For those of you more versed in music theory than I am, the complete methodology for transforming data to music is in the paper. In layman’s terms, a melody represents one element from the biological dataset, for example nitrate or chlorophyll A concentration, these are ‘plotted’ against the chords, which represent another column of data, such as monthly measures of temperature or salinity. In one of their compositions ‘Fifty Degrees North, Four Degrees West’, a chord progression of 12 chords used in the chorus represent the changes in salinity and temperature data over the 12 months of the year.


Figure 1: The general approach to Microbial music is summarised using a hypothetical data set.

Though this may all sound very farfetched, the authors genuinely propose Microbial Bebop as one approach to engage the non-scientific community in ecological science. Perhaps one serious point they make in their conclusion is that the use of human intuitive understanding to solve problems that computers find it hard to deal with has precedents, and that the innate human ability to detect patterns and subtle changes in music may make the handling and interpretation of complex biological data sets easier with this technology. 

They claim in their final note that the ‘possible permutations of data transformed into music are nearly infinite’. I personally do not see it becoming part of the mainstream in scientific literature, ‘Fig 1: an MP3 of our dataset’ is not likely to be the future of ecological research. However they do raise the point that more effective methods of engaging the public are needed. Just perhaps not in the form of bebop jazz improvisation.


All in all a surreal and yet interesting area of research to look in to, and if nothing else, it proves that you can receive funding to do just about anything (so long as the words ‘climate change’ are in there somewhere).


Larsen, P., & Gilbert, J. (2013). Microbial Bebop: Creating Music from Complex Dynamics in Microbial Ecology. PloS one8(3), e58119.

5 comments:

  1. I wonder if you can buy the records of it soon..I would definitely listen to them. ;-)
    However, this is actually very interesting, do they focus on the whole community or just some phylogenetic groups? I could imagine that it would sound quite confusing having to focus on many groups at the same time that might react differently to changes in the environment (e.g. halotolerant vs. halophiles/halophobes).

    ReplyDelete
  2. Well for example in of their compositions 'Bloom' there are many different melodies that each associate with a particular rare species or taxa, one each for: Cyanobacteria, Vibrionales, Opitulates, Pseudomondales, Rhizobiales, Bacillales, Oceanospirallales, and Sphingomonadales. Each melody comes to the foreground in the music as each associated taxa blooms across the year, with the chords representing the salinity and chlorophyll A concentrations.
    I feel its in no way an exact science, and is very much dependent on the musical skills and tastes of those creating it, more a novelty than anything, yet they claim it will have its uses.

    ReplyDelete
  3. That's a fascinating approach! The data come from the work that Jack Gilbert did when he was at PML. You can hear examples at this link https://soundcloud.com/plos-one-media/sets/microbial-bebop
    Is anybody good at choreography - maybe we could have a dance instead of or next seminar?

    ReplyDelete
  4. This is amazing! It’s really interesting that they choose jazz as their choice of style to base their study on, as it’s a much more “laid-back” genre of music that can involve a wide range of different techniques and interpretations, and microbial communities are so diverse and complicated it almost seems fitting!
    On the other hand, as you mentioned before about the different melodies being associated with specific taxa or groups, this could be an interesting way of demonstrating the phylogenetic relationships between them to people who cannot understand the biology and the genetics, but are able to relate to the music- i.e. the more similar two particular melodies are, the more closely related the groups?

    ReplyDelete