Pondering: how to represent genes as "constellations" or "coastlines": distinct, visually recognizable shapes. One option: line that changes angle reliably at each splice junction, start and stop codon. Info in any gff file. E.g. here are five genes; F1 says it looks like a swan. 
In this case:
Exon-to-Intron: +0.2 radians
Intron-to-Exon: -0.3 radians
Gene start/end: +0.5 radians (in coding seq direction)
Obviously these are arbitrary. But fun to play with.
Exon-to-Intron: +0.2 radians
Intron-to-Exon: -0.3 radians
Gene start/end: +0.5 radians (in coding seq direction)
Obviously these are arbitrary. But fun to play with.
I mean pi*radians in those
