DNA Portraits from personal genomics data

In order to reduce the complexity and size of the datasets, I created a coarse summary of nucleotide pairs, reducing a million points of difference down to thirteen aggregates per chromosome. Would this still preserve the individuality of each genome? As it turns out, the answer is yes.

The starting point were raw data files from the genetic service 23andMe, extracted from my own DNA and that of a couple of friends. Each file contains almost a million two-letter genomic position readouts. For each chromosome, counts of the observed letter combinations are translated to overlapping shapes of corresponding size, orientation, and color. The resulting artwork is different for each person.