2. A Case Study of Proteomic Data Sonification
Working with biomedical scientists Philip Poronnik (University of Sydney) and Derren Saunders (University of New South Wales), a variety of sonifications of proteomic datasets associated with Amyotrophic Lateral Sclerosis (ALS) were designed to support sonic exploratory data analysis (EDA). In an effort to provide objective verification of the effectiveness of the design decisions made in the development of those sonifications, an experimental research study (Martens, Poronik and Saunders 2016) was executed. In that study, three distinct sonifications of three datasets were presented to a group of naïve listeners who were instructed to listen for differences between the three alternative sonifications of proteomic data distributions associated with three different types of cells. The motivation for the study was to determine whether the distinct sonifications could reliably indicate differences due to the neuropathology associated with ALS via auditory display (through exploration of the differentiation of induced pluripotent stem-cell-derived neurons). Auditory display was thought to provide potential benefits over pure visual displays of proteomic data, which for biomedical scientists often result in visual overload, resulting in the possibility that detailed or subtle differences in data pertinent to ALS neurodegradation are missed.
This project was launched with a focus on the selection of sound synthesis technologies that were thought to be the best candidates for transforming data structure and properties into auditory displays that could support more intuitive interpretations of the data. The preliminary results of this auditory data exploration were presented at the 22nd International Conference on Auditory Display (ICAD 2016) that was held at the Australian National University (Martens et al. 2016). Although the results of that study revealed only the perceptual distinctiveness of the sonifications, these preliminary results provided a basis for further exploration of the potential value of these sonifications, with the benefit that their perceptual distinctiveness had been established. Further work will be required to address important issues in sonification system usability for this application and to determine whether the system will provide a real benefit to domain experts (i.e., the biomedical scientists with interest in finding differences in proteomic datasets due to ALS neuropathology). Thus, this paper reports these preliminary evaluation results as an indication of progress on just one component of the larger medical research project, focusing solely upon the effectiveness of the display technology used to aid domain experts in interpreting and understanding medical data (particularly proteomic data). Nonetheless, it is expected that such preliminary examination of the perceptual distinctiveness of sonification system outputs will provide insight into the design and development of novel and potentially useful sonifications.