Custom software turns genes ?incomprehensibly long strings of As, Cs, Ts and Gs ?into luminous pictograms that resemble Chinese or Sanskrit calligraphy. Based on currently available biophysical information, the pictograms are scientifically accurate representations of proteins encoded for by the genes.
In the SIGGRAPH installation, the representations are rendered in a 40-foot wide and 12-foot tall space by five video projectors, with the figures for human genes/proteins shown along a vertical axis and for the rice along a horizontal.
A whole-body computer vision interface tracks the movements of visitors and allows them to interact with the installation. By moving their bodies slowly within the space, visitors can draw shimmering light-filled traces. When a trace sufficiently matches a pictogram in the human dataset, it triggers a real-time bioinformatics comparison: BLAST begins to run, searching through the rice data for a homologue ?conducting in a novel (and visible) way the same sequence analysis done by scientists. Results are presented as two superimposed pictograms.
"This high-dimensional visualization reduces the complexity of sequence codes to the sorts of shapes or patterns that a human being can make sense of," West said. "It is an artistic approach to extracting what's important. And it is also an exploration of what art might have to offer for discovery in the sciences."
"Ecce Homology" premiered in 2003 at the UCLA Fowler Museum of Cultural History.
At SIGGRAPH 2005, "Ecce Homology" is being showcased as part of the international conference's Art Gallery and its Emerging Technologies program. It will also be featured in the August 2005 issue of Leonardo, an art, science and technology journal from MIT Press.
###
The project is supported by Intel Corp., NEC Solutions America/Visual Systems Division and groups at UCSD, UCLA and USC.
To learn more: http://www.insilicov1.org/
<