Exchange format like GFF and databases like Chado allows for arbitrary configurations of feature graphs. The GFF spec provides recommendations, e.g. so called 3-level graphs (gene->transcript->exon) but this is encoded in free text to be implemented procedurally by validators.
One of the original ideas was to use SO to constrain the shapes of these graphs but of course OWL is open world, and SO is under-axiomatized so in practice SO just serves to document like associations.
It seems that faldo plus a shape language would be ideal here, and very useful for faldo in general
Exchange format like GFF and databases like Chado allows for arbitrary configurations of feature graphs. The GFF spec provides recommendations, e.g. so called 3-level graphs (gene->transcript->exon) but this is encoded in free text to be implemented procedurally by validators.
One of the original ideas was to use SO to constrain the shapes of these graphs but of course OWL is open world, and SO is under-axiomatized so in practice SO just serves to document like associations.
It seems that faldo plus a shape language would be ideal here, and very useful for faldo in general