Bacteriophage in Granular Waves: AMICAD Performance

The field of Viral Genomics, and Genomics more generally, is experiencing a crisis as datasets are growing in size and complexity at an astounding rate. With the with NCBI's Sequence Read Archive (SRA) now sitting somewhere in the range of 80-90 petabytes of data as of 2025 it is becoming sonification is becoming an increasingly critical tool for exploring and representing this data.
This piece represents a form of play at the intersection of the fields of experimental music composition and viral genomics.
Bacteriophage in Granular Waves, is a data-driven musical performance that employs generative sonification techniques to map synthetic virology data to musical parameters. It is the result of initial steps towards the development of a more generalised system for sonifying viral genomics data for contexts in biomedical analytics and education. The data was created in-silico and modelled after the characteristics of viral bacteriophages we might expect to find in human tissue samples. The data of interest are base pair (bp) length and coverage values which are measures of physical length of the double helix for a given DNA strand and the completeness of the genomic sequencing process respectively. While coverage gives a sense of how much of a given bacteriophage is present in a sample, the bp length of a viral genome provides an indication of the size and, to some extent, the complexity of a given virus. The system maps the data to control the parameters of a generative system, specifically an agent based model simulation consisting of an ecosystem of autonomous agents where each agent represents a unique bacteriophage in the dataset. The agents are engaged in simple target seeking behaviours in a 2D environment, intermittently passing through a “hotspot” which causes their data associated to be passed to the sound synthesis routines. Their speeds and steering rates are determined by the data. This simulation is visualised and projected live during performances. Agents entering a hotspot send their midi notes to wavetable synthesis routines to produce a signal. The first round of sound synthesis mappings are focus on representing the data at the timbral level. Synthesis parameters are mapped to the data to ensure that audible clarity of the sound produced is driven by the coverage while the spectral complexity is mapped to bp length. These signals are sent to granular synthesis routines for further processing. At this layer data are mapped to represent the data at the textural layer with coverage defining the audible clarity of the textures while bp length defines the textural complexity of the sound. The result is a rich sonorous soundscape that operates as an aesthetic experience of the data as rendered to sound by a data-driven generative music system.